fix(plugins): resolve review findings and integrate main

fix(ghidra): fail doctor when the JVM os.arch selects an unvalidated decompiler directory (#2792 )
* fix(ghidra): fail doctor when the JVM os.arch selects an unvalidated decompiler directory probeJava now also reads the JVM os.arch property. When it maps to a different native decompiler directory than the one validated from the host architecture (for example an x86_64 JDK under Rosetta on Apple silicon selecting os/mac_x86_64 while only mac_arm_64 was built), the native_decompiler check fails with a remediation pointing at a matching-arch JDK instead of reporting a healthy installation that fails every analysis. Unknown os.arch values keep the previous behavior. * fix(ghidra): tighten JVM arch remediation and cover os.arch extraction Address review findings on the JVM os.arch decompiler-directory check: - The remediation no longer suggests building the decompiler for the JVM-selected directory, which could never clear the check; it only advises pointing JAVA_HOME at a matching-architecture JDK. - Add a test running the real Java probe against captured -XshowSettings:properties output, asserting os.arch reaches the installation diagnostics. - Document the JDK-architecture prerequisite in the Ghidra setup guide. * fix(ghidra): refuse unrecognized JVM native architectures * docs(ghidra): align doctor prerequisites with JVM architecture admission * fix(ghidra): reject unreported JVM architecture * test(ghidra): separate JVM and executable admission cases * fix(ghidra): defer decompiler arch validation when Java is absent * fix(ghidra): reject JVM native architecture override * fix(ghidra): validate JVM platform overrides * fix(ghidra): validate JVM os.name and helper-environment platform overrides * test(ci): isolate report fixtures and declare PowerPC Java bounds * test(ghidra): provide observed JVM platform in PowerPC fixture * fix(ghidra): parse quoted JVM environment options before admission --------- Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
2026-10-14 07:51:46 +08:00 · 2026-10-14 07:31:11 +08:00 · 2026-10-14 07:30:14 +08:00 · 2026-10-14 07:28:00 +08:00 · 2026-10-14 07:18:50 +08:00 · 2026-10-14 07:12:10 +08:00
2398 changed files with 589758 additions and 23674 deletions
@@ -0,0 +1,21 @@
+{
+  "name": "rea",
+  "interface": {
+    "displayName": "REA"
+  },
+  "plugins": [
+    {
+      "name": "rea",
+      "source": {
+        "source": "npm",
+        "package": "rea-agents",
+        "version": "6.4.0"
+      },
+      "policy": {
+        "installation": "AVAILABLE",
+        "authentication": "ON_INSTALL"
+      },
+      "category": "Developer Tools"
+    }
+  ]
+}
@@ -25,6 +25,7 @@ node scripts/release-notes.mjs inventory \
  --manifest /tmp/rea-VERSION.contributions.json
 node scripts/release-notes.mjs record \
  --manifest /tmp/rea-VERSION.contributions.json \
+  --summary \
  --output /tmp/rea-VERSION.record.md
 ```

@@ -56,25 +57,31 @@ version/date/link heading. Use single-line Markdown list items under:
  PR/commit references required by the existing checkpoint validator.

 Group related work by analyst workflow or user impact. Explain the resulting
-behavior; internal cleanup, tests, CI, and routine documentation generally
-belong only in the contribution record. Read ambiguous PRs/diffs before making
+behavior in short, concrete entries. Highlights summarize the release; avoid
+repeating their full descriptions in Changes. Internal cleanup, tests, CI, and
+routine documentation generally belong only in the provenance file. Read ambiguous PRs/diffs before making
 claims. Distinguish simulated-provider checks from real-provider verification.

-Each bullet covering a PR carries its link and `Thanks @...` for its verified
-human author, known co-authors, and confirmed closing-issue reporters. Grouping
-must preserve **all** those credits in the same bullet. Include the issue
-links. Direct-commit bullets carry the commit link and known human credit.
+Credit each represented contribution once in the public notes with its link
+and `Thanks @...` for its verified human author, known co-authors, and confirmed
+closing-issue reporters. Grouping preserves all those credits in one line.
+When Highlights, detailed entries or migrations reference the same work,
+put its thanks in one of those entries, not all of them. Include issue links.
+Direct-commit bullets carry the commit link and known human credit.
 Exclude GitHub bots and known agent accounts. REA's observed Anthropic and
 Cursor co-author trailers resolve to `@claude` and `@cursoragent` with GitHub
 type `User`; the credit policy excludes them while retaining their metadata.
-Do not copy OpenClaw's maintainer-specific exclusions.
+At the REA maintainer's request, omit `@morluto` self-thanks while preserving
+authorship in the inventory. Other human contributors remain eligible.
 Use explicit GitHub links for cross-repository references to avoid ambiguous
 `#NNN` numbers. Unknown handles stay unknown; they are not fabricated credit.

-Append the generated `### Complete contribution record` verbatim as the last
-section of this release. It accounts for every discovered in-range PR and
-direct contribution, including internal work omitted from the prose. Do not
-hand-edit generated credits to make validation pass. Refresh the inventory
+Append the generated `record --summary` provenance link verbatim at the end
+of this release. The linked inventory accounts for every discovered in-range
+PR and direct contribution, including internal work omitted from the prose.
+Do not paste the full `record` output into public notes: it duplicates thanks
+and obscures the user-facing changes. Full records remain available for review.
+Do not hand-edit generated credits to make validation pass. Refresh the inventory
 when product history changes; a prose-only correction can reuse the snapshot.

 ## Finalize with Release Please
@@ -0,0 +1,112 @@
+---
+name: rea-ci-tests-audit
+description: Audit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.
+---
+
+# REA CI and tests audit
+
+Establish which checks still catch real regressions and which costs can be
+removed without losing that protection. The outcome may be a verified fix,
+an evidence-backed recommendation, or an explicit decision to stop.
+
+Follow the checkout's `AGENTS.md`, [testing guide](../../../docs/testing.md),
+and [contribution rules](../../../CONTRIBUTING.md). Use their current commands
+and pinned toolchain, rather than commands remembered from an earlier audit.
+Preserve the requested scope: investigation alone does not authorize edits;
+this skill does not grant permission to publish issues, push, or merge. Existing
+user authorization still applies. Keep audit notes and raw results outside the
+tracked tree, normally in a task-specific `/tmp` directory.
+
+## Establish the executed baseline
+
+Fetch when the user asks for latest main. Pin the baseline SHA and inspected
+workflow revision; use an isolated worktree when changing or comparing code.
+Do not modify another task's checkout. Record dirty changes and the exact
+command, toolchain, host, selected tests, instrumentation, and prerequisites.
+An old run is evidence for its executed revision, even if main advances.
+
+Trace workflow triggers, changed-path selection, job dependencies, reusable
+actions, package scripts, test discovery, artifact producers, and report gates
+through to the commands that actually execute. Include scripts, generated data,
+runtime registration, and subprocess boundaries missing from import graphs.
+Check positive and negative selection paths: required files must be discovered
+once per intended test configuration; deliberate platform/configuration repeats
+are separate coverage. A missing explicit test or missing required report must
+not turn green. Optional lanes may skip; preserve their stated prerequisites.
+
+Use existing timing and failure artifacts to select suspicious families. Run
+relevant checks to establish a baseline; reserve a full local suite for broad
+CI diagnosis or uncertainty that focused checks cannot resolve. Do not rebuild
+or rerun an unchanged complete suite after every documentation edit.
+
+## Prove the failure or protection
+
+For each candidate, state the expected observable behavior and its owner,
+then construct the smallest reproduction against production behavior. Follow
+the producer and affected consumers before deciding where to fix it.
+
+- For drift, compare fixtures with current admission requirements, protocol
+  replies, named contracts, tool versions, and actual public entrypoints. A
+  plausible-looking mock can fail before the behavior its test claims to cover.
+  Search the whole affected fixture family, including other version strings.
+  Preserve intentionally malformed fixtures and test their intended rejection.
+- For a suspicious green test, perturb the relevant behavior in an isolated
+  copy or inject a controlled failure. Verify that the existing assertion or
+  gate detects it. A test may merely repeat its mock, import abandoned code,
+  or hide a subprocess/checker failure. Restore mutations after the experiment.
+- For shared setup or pruning, map distinct assertions to their retained
+  scenarios. Preserve malformed inputs, cancellation, cleanup failure,
+  permissions, capacity, identity, and meaningful concurrency interactions.
+  A faster run that skips a case or weakens its oracle is not an improvement.
+- For provider coverage, separate fake-producer, boundary, and actual-engine
+  evidence. Inspect named skips and prerequisites; fixtures do not establish
+  real Ghidra, LLDB, browser, or other provider behavior.
+
+Classify findings as reproduced defects, supported costs, hypotheses, or
+unavailable evidence. Record the baseline command, observed output/exit status,
+expected result, affected scope, and competing explanation. A failure on main
+may be pre-existing; that does not by itself make it irrelevant to a CI audit.
+Attribute PR failures using the executed revision and a clean baseline, not
+just their resemblance to a previous failure. Retry only to test a specific
+transient explanation; retain the original failure and stop when repetition
+adds no evidence. A passing retry does not establish a root cause or a fix.
+
+## Assess efficiency without losing the gate
+
+Read [measurement and CI checks](references/measurement-and-ci.md) when
+investigating slowness, caching, runner waits, shared artifacts, or log noise.
+Compare equivalent completed work before claiming a speedup. Distinguish
+setup/build time, test execution, coverage instrumentation, dependency waits,
+and runner scheduling; state which costs remain unmeasured.
+
+Prefer removing proven duplicate preparation, reusing immutable setup, and
+reducing repeated subprocess work before shrinking correctness coverage.
+Preserve independent sessions where lifecycle or concurrency is the oracle.
+Keep live tests fresh; a cached previous pass cannot validate current host,
+installed provider contents, permissions, or external state.
+
+## Fix, verify, and hand off within scope
+
+If implementation is authorized, make the smallest structural change at the
+owning boundary and add a regression that fails on the baseline for the right
+reason. Retest affected consumers and meaningful negative paths; verify cache
+invalidation or discovery when those mechanisms change. Read the installed
+Turbo documentation before changing its configuration or commands.
+
+Run the repository's applicable checks and inspect complete current-head CI
+results before an authorized merge. Preserve required coverage and discovery
+gates. Do not suppress a failure, add blanket retries, raise timeouts, or delete
+distinct assertions solely to make CI green. Changes to these controls need
+their own demonstrated reason and retained failure detection.
+
+Report each confirmed issue with exact revision, reproduction, consequence,
+proposed correction, and validation limits. Report performance with workload
+and measurements, not just a percentage. Use English issue/PR records when
+requested; follow the user's assignment, commit identity, review, and merge
+preferences without assuming them for other users. Open ready-for-review PRs
+only when that lifecycle is authorized. Resolve necessary review findings and
+check final main against the tested tree after merging.
+
+Stop when the authorized work is verified, remaining benefits are unproven,
+or further changes would trade correctness for speculative speed. List genuine
+uncertainties without turning them into additional optimization rounds.
@@ -0,0 +1,4 @@
+interface:
+  display_name: "REA CI and tests audit"
+  short_description: "Audit CI/test correctness and measured performance"
+  default_prompt: "Use $rea-ci-tests-audit to investigate CI and test correctness, drift, and performance before proposing changes."
@@ -0,0 +1,65 @@
+# Measurement and CI checks
+
+## Equivalent completed work
+
+Record the source/workflow SHA, command, test identities and counts, skips,
+assertions or output contract, exit status, coverage scope and thresholds,
+toolchain/provider versions, host resources, and prepared artifacts. Inspect
+both the child result and its report: a profiling wrapper may exit successfully
+after the measured command failed. Focused tests can pass while global coverage
+thresholds fail; describe both rather than calling the command green.
+
+Measure wall time, user/system CPU, and peak process-family memory for relevant
+completed tasks. Parent-only RSS misses workers and their children. On Linux,
+summed RSS can double-count shared pages; report RSS/PSS semantics, sampling
+interval, missing samples, and observed cleanup limits. Use equivalent host
+tools elsewhere. State unknown costs, including agent tokens; fewer log bytes
+alone prove neither fewer tokens nor faster tests.
+
+Separate cold execution, warm build/file caches, and warm runtime caches.
+Use repeated or alternating runs when variance affects the decision, with
+enough observations to distinguish the proposed benefit from noise. Keep
+workload, instrumentation, and host conditions comparable; record interference.
+A run with fewer tests, early rejection, lost coverage, or a different runner
+does not prove suite acceleration. CI-wide claims need end-to-end measurements,
+not an extrapolation from one improved family.
+
+## Workflow, caches, and shared artifacts
+
+Trace each reused result from producer to consumers and identify its lifetime.
+Check key/input coverage against the actual file inventory and result-affecting
+configuration, environment, generated data, tool versions, and host state.
+Exercise an affected input change to prove invalidation; exercise an irrelevant
+change to support reuse where useful. Cache deterministic preparation and
+static analysis only when their inputs fully describe the result.
+
+Verify downloaded artifacts match the expected revision, platform, runtime,
+and schema. Check restore misses, failed/partial producers, stale reports,
+removed output files, and consumers after producer failures. Shared writers
+need output ownership or locks; tests need isolated state, not an unnecessary
+global lock. Worker counts do not bound children spawned within each test.
+Preserve intentionally concurrent scenarios while bounding unrelated batches.
+
+Check matrix expansion and all job pages before saying CI passed. Inspect
+required gates after failed, cancelled, and skipped dependencies. `always()`
+alone does not preserve failures; inspect the aggregation logic. Test/report
+uploads should retain available diagnostics after failure, while absent or
+malformed required results must fail closed. Do not demand an optional report
+from a lane intentionally excluded by the selector.
+
+## Runner waits and console noise
+
+Use per-job and per-step timestamps where available. Workflow creation to job
+start includes dependencies and other delays; it is not necessarily runner
+queue time. Identify when the job became eligible before attributing waits to
+runner capacity. If eligibility/queue timestamps are unavailable, report the
+combined delay and unknown split. Compare equivalent lanes and runner types
+across enough runs; a single busy runner is not a workflow regression.
+
+For console reductions, retain checker exit codes, actionable errors, useful
+summary counts, and complete machine-readable artifacts. Verify authored code
+remains in the actual checker inventory when excluding generated outputs.
+Introduce an authored error as a control; it must still fail. An ignore pattern
+can hide real source as well as noise. Show that cached reports are restored,
+stale successful reports cannot mask current failures, and report parse errors
+remain failures. Measure readability or byte reduction separately from runtime.
@@ -2,7 +2,7 @@
 name: reverse-engineer-anything
 description: Reverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.
 metadata:
-  version: "37"
+  version: "41"
 ---

 # REA
@@ -17,6 +17,9 @@ readiness or provider commands.

 Installing this skill supplies instructions; it does not register the REA MCP
 server, install analysis engines, or add tools to an already-running agent.
+The exception is a Claude Code or Codex plugin install: the plugin bundles
+this skill and registers its own `rea` server, outside the state that
+`rea setup` and `rea doctor` manage.

 If REA tools are available and their registration is not known to be stale,
 proceed directly to the target. Do not run diagnostics or setup before every
@@ -31,6 +34,49 @@ Diagnose the observed blocker, continue useful supported work, and resume the
 original task after repair. Client approval, shell PATH, registration and engine
 availability have different remedies; avoid repeating setup for all of them.

+1. Diagnose without changing REA or client configuration. `npx` may download
+   the package and write to npm's cache. For Codex, run
+   `npx -y rea-agents@latest doctor --client codex --json`. Substitute the current
+   supported client: `claude_code`, `claude_desktop`, `codex`, `cursor`,
+   `gemini_cli`, `windsurf`, `devin`, `opencode`, `antigravity`, `copilot_cli`,
+   `commandcode`, `vscode`, `grok_build`, `omp`, or `grok_bot`. `grok_bot` has no local
+   file registration: doctor reports the Grok Bot chat step, and setup does not
+   write `mcp.json`. If the client is unknown, use `doctor --json` and
+   inspect its registration results before choosing a setup scope.
+2. Distinguish the reason. Missing, malformed, or stale registration needs a
+   scoped configuration repair. An aligned registration with no tools in the
+   active session needs a restart/reconnection; doctor cannot prove that the
+   current agent has connected. A missing provider affects only tasks requiring
+   that provider: static JavaScript inspection needs neither Hopper nor Ghidra,
+   and Android inspection has separate bring-your-own JADX/Java prerequisites.
+3. For a configuration repair, prepare the read-only plan:
+   `npx -y rea-agents@latest setup --client codex --dry-run --json`.
+   Use the current client's ID, show its exact proposed paths, backups, and
+   changes, and obtain approval before setup writes configuration or installs
+   Hopper. Setup normally installs the matching bundled skill too; include that
+   replacement in the reviewed plan. After approval, apply the same scope with
+   `npx -y rea-agents@latest setup --client codex --yes`. Add `--install-hopper`
+   only if that separate installation was needed and explicitly approved.
+4. Restart/reconnect the affected agent. Verify that REA tools actually appear
+   in the session, then resume the original investigation. If they remain
+   absent, inspect the client's MCP launch error rather than repeating setup.
+
+REA setup never installs or upgrades Node.js, npm, Homebrew, Java, Ghidra, IDA,
+JADX, Binwalk, or Unblob. Use existing prerequisites; do not install unrelated
+software to repair
+MCP registration. For an unsupported client, use manual stdio registration
+with a version-pinned `rea-agents` package or continue through the CLI.
+
+A concrete CLI fallback for an operator-supplied JavaScript tree or ASAR is:
+
+```bash
+npx -y rea-agents@latest analyze-javascript-application /absolute/path/to/app --json
+```
+
+No MCP registration or native engine is required. Read the returned Evidence,
+graph, limitations, and unknowns with the same care as an MCP result; the CLI
+returns the Evidence record directly. This fallback does not establish that
+MCP is configured. Native CLI tasks still require their selected engine.
 Prefer an existing `rea` CLI. When none is installed and REA package acquisition
 is within the authorized scope, use `npx -y rea-agents@latest` in its place.
 Configuration repair starts with a read-only setup plan; show paths, backups and
@@ -61,17 +107,84 @@ explicit path or endpoint and do not need it.
  consult the NativeAOT workflow in `references/native-and-artifacts.md` for
  optional metadata recovery and its supported host/layout boundary.
 - User-owned browser page already open: `list_browser_targets`.
- Retained HAR/native mitmproxy capture: `inspect_web_network_capture`. Select
+- Fiddler Everywhere traffic or a native `.saz` capture: read the
+  [Fiddler workflow](references/runtime-observation.md#fiddler-everywhere-traffic-debugging).
+  Inspect saved SAZ with `inspect_web_network_capture` and `format: "saz"`;
+  connect Everywhere's separate MCP server for live traffic.
+- Retained HAR/SAZ/native mitmproxy/aioquic qlog capture: `inspect_web_network_capture`. Select
  original record ordinals when useful; inspect producer fields and byte/number
  sidecars without fetching recorded URLs or claiming live attribution. See
  [historical captures](https://github.com/morluto/rea/blob/main/docs/web-network-captures.md) for the exact upstream
-  profile and credential exclusions.
+  profile and credential exclusions. Qlog support starts with aioquic 1.5.0
+  JSON 0.3; stream coverage and endpoint loss declarations do not recover bodies
+  or establish network loss. Prefer detail summary, then kind qlog trace/stream
+  pages; select one trace, stream ID and direction for contribution/range pages.
+  Facts retain all-recorded-events-before-selection scope. Other qlog layouts
+  and PCAP correlation are unsupported.
 - User-owned Electron runtime already open: `list_electron_targets`.
+- Packet-field discovery: `inspect_packet_fields` searches installed protocols
+  and field names/types before choosing fields for `inspect_packet_capture`.
+  Query/page the relevant records; registrations do not establish observed traffic.
+- Recorded packet captures: `inspect_packet_capture` with selected TShark fields,
+  display filters and decode-as rules. Requires Linux, util-linux prlimit and `REA_TSHARK_COMMAND`; see
+  [capture guide](https://github.com/morluto/rea/blob/main/docs/packet-captures.md).
+  Packet limits count input packets before filtering. Output defaults to 50 matching
+  frames; use `match_offset` / `match_limit` for later pages, or explicit null
+  limit for complete matching output. Preserve the source digest across pages.
+  This does not perform live capture.
+  For address/DNS/SNI/protocol summaries use `inspect_pcap_capture` with MCP
+  detail summary, then retained kind pcap pages. Requires Wireshark 4.x tshark
+  and capinfos, with Linux prlimit or Windows Job Objects; see
+  [packet summaries](https://github.com/morluto/rea/blob/main/docs/pcap-analysis.md).
+  Select an exact DNS question for address pages; page exhaustion does not prove
+  network completeness. This reads recordings and does not acquire traffic.
+- Decoder discovery: `inspect_signal_decoders` searches installed decoder IDs
+  or selected decoder help, including channels and options. Query and page the
+  relevant details before choosing settings; this does not scan hardware.
+- Recorded logic/signal traces: `decode_signal_recording` with explicit decoder,
+  channel and protocol settings. On Linux with util-linux prlimit, supply sigrok-cli 0.7.2 through
+  `REA_SIGROK_COMMAND`; see [recording guide](https://github.com/morluto/rea/blob/main/docs/signal-recordings.md).
+  Output defaults to 50 literal lines; select annotation classes and page with
+  `line_offset` / `line_limit`, or explicit null for all remaining output.
+  Window counts are display lines, not annotations or samples. Preserve source digests across pages.
+  This reads existing files; it does not acquire hardware signals.
+- Explicit local WebAssembly artifact: `inspect_wasm_artifact`. Supply WABT 1.0.42
+  through `REA_WABT_BIN_DIRECTORY`; retain exact byte/tool identity and decoded WAT.
+  Select JavaScript glue and candidate paths explicitly and preserve URL/path
+  ambiguity. No implicit fetch or target execution; WAT is not original source.
+  See the [WASM guide](https://github.com/morluto/rea/blob/main/docs/wasm-artifacts.md).
 - Explicit local EVM bytecode carrier: `inspect_evm_interface` with caller-selected
  `raw` or `hex` encoding. Preserve carrier/decoded digests and treat selectors,
  argument strings and mutability as inferred candidates; this performs no chain
-  lookup or target execution. See the
+  lookup or target execution. For the complete inferred interface without hex or
+  duplicate stdout, discover support and select `detail: "summary"`; complete
+  Evidence stays retained under its original identity. Use the returned
+  `parent_evidence_id` for exact function or `bytecode_hex`/stdout/stderr pages.
+  Candidate coverage does not establish interface completeness. See the
  [EVM bytecode guide](https://github.com/morluto/rea/blob/main/docs/evm-bytecode.md).
+- Embedded Go compiler, module or build-setting metadata: `inspect_go_binary`
+  with the explicit local file. This built-in target-free inspection requires no
+  Go installation or disassembler. Preserve raw module bytes and source ranges;
+  absent metadata does not prove a binary is not Go, and unparsed text records
+  remain explicit. See the
+  [Go binary guide](https://github.com/morluto/rea/blob/main/docs/go-binaries.md).
+- Embedded cargo-auditable Rust dependency metadata: `inspect_rust_binary`
+  with the explicit local ELF, PE, or thin Mach-O file. This built-in
+  target-free inspection requires no Rust installation or disassembler. Preserve
+  package order, dependency indices, and the section file range. A missing
+  `.dep-v0` section is null metadata, not proof the binary has no dependencies.
+  A rustc commit hash is an inference from a read-only `/rustc/<40-hex>/` path.
+  See the
+  [Rust binary guide](https://github.com/morluto/rea/blob/main/docs/rust-binaries.md).
+
+- Explicit local SQLite database snapshot: `inspect_sqlite_database`. Inspect
+  with Node.js 24.x >=24.15 or 26+ providing native SQLite limits. Read schema
+  first (compact inventory by default), or select an exact ordinary table
+  and `columns` when rows are needed. Defaults are selected-table schema and
+  20 rows; request `schema_scope: "database"` for the complete stored schema. Preserve both database and existing sibling WAL identity; row samples
+  report truncation and have unspecified order. Use a quiescent snapshot rather
+  than assuming a live database/WAL pair was copied atomically. See the
+  [SQLite guide](https://github.com/morluto/rea/blob/main/docs/sqlite-databases.md).
 - Explicit Linux ELF file for offline layout, symbols, relocations or static
  mitigation evidence: `inspect_binary_layout`. This target-free operation uses
  caller-supplied pwntools without opening a disassembler database. Preserve its
@@ -110,8 +223,14 @@ ask when matches are ambiguous. Never choose an example app on the user's behalf
 In a target-free session, use `open_binary` to bind any archive/package or
 native target whose analysis tool operates on the active target. Do not call an
 unadvertised tool; inspect `binary_session` with `{}` and its
-`result.tool_availability` for availability reasons and remediation. The tool
-list stays complete; availability depends on the operation, target, and host.
+`result.tool_availability` for availability reasons and remediation. Expand each
+group's `names` into entries, inheriting the inventory-level client requirements
+unless the group overrides them. In its closed effect and annotation vocabularies,
+listed flags are true and omitted known flags are false. The grouped name union
+is the complete tool inventory; use `tools/list` for canonical tool schemas.
+Provider candidate facts describe possible providers, while top-level session
+capabilities describe the current route and `analysis_provider_binding` names the
+selected provider. Do not treat those scopes as interchangeable.

 ## Work summary-first

@@ -135,6 +254,16 @@ session to a selected path. The analysis remains complete in the ledger; avoid
 repeating it merely to request the same oversized response. Complete CLI JSON
 output also streams without a single MCP frame.

+When the retained parent ID is unknown, use `get_evidence_bundle` with
+`detail: "summary"` and exact filters for operation, target SHA-256,
+analysis-profile digest or canonical procedure address. Select its returned
+Evidence ID for the next view instead of requesting the complete bundle.
+`retention: "complete-record"` means the whole record is stored; native
+value-flow truncation and unknown metadata remain separate. For an exported
+portable bundle, use `rea inspect-evidence-bundle` with JSON containing `path`,
+`detail: "summary"` and the same `filters`. Reading that file still requires
+validating its full contents within the local memory budget.
+
 A client can also truncate a successful tool result without a REA transport
 error. Treat that preview as incomplete; use the retained parent ID for smaller
 views or export when the full record is needed. Do not parse the preview as
@@ -11,14 +11,14 @@ main and an older release expose the same arguments.

 ## Choose the remedy from the observed failure

-| Observation                                                     | Meaning and next step                                                                                                                                                                                            |
-| --------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| No REA tools in the active session                              | Inspect registration with `rea doctor --client codex --json`. A skill installation alone does not register MCP.                                                                                                  |
-| Missing, malformed or stale registration                        | Prepare a scoped setup plan; apply only authorized changes with backups.                                                                                                                                         |
-| Aligned registration, but no connected tools                    | Restart/reconnect the client. Doctor checks registration; it cannot establish the current agent's connection. Inspect the client's launch error if reconnection fails.                                           |
-| `MCP tool call requires approval, but approval policy is never` | Codex rejected the call before REA executed it. Do not install an engine or repeat setup to solve this client-policy decision. See the approval section below.                                                   |
-| REA returns `provider_unavailable` or an unavailable tool entry | Read that operation's provider diagnostics or `binary_session.result.tool_availability`. Resolve only the prerequisite needed for the task; a complete tool catalog does not imply every operation is available. |
-| `rea` is missing in a child shell, or Node is incompatible      | Inspect the actual shell PATH and launcher/runtime. A prefix installation may be absent from a login shell's PATH; this does not establish a broken MCP registration.                                            |
+| Observation                                                     | Meaning and next step                                                                                                                                                                                                                                                                                                 |
+| --------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| No REA tools in the active session                              | If REA was installed as a client plugin, follow the plugin section below. Otherwise inspect registration with `rea doctor --client codex --json`. A skill installation alone does not register MCP.                                                                                                                   |
+| Missing, malformed or stale registration                        | Prepare a scoped setup plan; apply only authorized changes with backups.                                                                                                                                                                                                                                              |
+| Aligned registration, but no connected tools                    | Restart/reconnect the client. Doctor checks registration; it cannot establish the current agent's connection. Inspect the client's launch error if reconnection fails.                                                                                                                                                |
+| `MCP tool call requires approval, but approval policy is never` | Codex rejected the call before REA executed it. Do not install an engine or repeat setup to solve this client-policy decision. See the approval section below.                                                                                                                                                        |
+| REA returns `provider_unavailable` or an unavailable tool entry | Read that operation's provider diagnostics or `binary_session.result.tool_availability`. Expand each group's tool names and inherit the shared client requirements unless overridden. Use the operation's availability, reason, and remediation; a complete tool catalog does not imply every operation is available. |
+| `rea` is missing in a child shell, or Node is incompatible      | Inspect the actual shell PATH and launcher/runtime. A prefix installation may be absent from a login shell's PATH; this does not establish a broken MCP registration.                                                                                                                                                 |

 If the client cannot launch its shell/sandbox helper or read the installed skill,
 inspect its launch diagnostic. For example, Codex can refuse helper aliases when
@@ -34,6 +34,33 @@ answers part of the question, and name the unavailable evidence separately.
 Do not substitute another target. Label observations from ordinary shell tools
 as such; only actual REA output supplies REA Evidence.

+## REA installed as a client plugin
+
+A Claude Code or Codex plugin owns its `rea` server and this skill. Doctor's
+`--client` and `--skill` checks and `rea setup` read and write only
+setup-owned state, so for a plugin-only install they report `missing`, and
+applying setup would add a second `rea` server. If this skill was loaded from
+a plugin, or the client lists a `rea` plugin, do not prepare a setup plan for
+missing tools. Instead:
+
+- Check that the plugin is installed and enabled with the client's own
+  listing: `claude plugin list` and `/mcp` in Claude Code, `codex plugin list`
+  in Codex.
+- Restart or reconnect the client, then read its launch error for the `rea`
+  server if tools are still absent. The first start downloads the pinned
+  package through `npx` and can exceed the client's startup timeout; a second
+  start uses the cache.
+- Check analysis engines through the connected server's
+  `binary_session.result.tool_availability`. `rea doctor --provider <id> --json`
+  from a shell reports that shell's environment, which can differ from the
+  environment the client gave the plugin server. If you still run it through
+  the package runner, use the version pinned in the plugin manifest's server
+  entry (`rea-agents@<version>`), not `@latest`, so it matches the server the
+  plugin starts.
+
+Use the setup repair below only when the user wants a setup-owned registration
+instead of the plugin, and remove or disable the plugin first.
+
 ## Repair registration when needed

 For Codex:
@@ -45,8 +72,8 @@ rea setup --client codex --dry-run --json

 Use the current supported client's ID: `claude_code`, `claude_desktop`, `codex`,
 `cursor`, `gemini_cli`, `windsurf`, `devin`, `opencode`, `antigravity`,
-`copilot_cli`, `commandcode`, `qwen_code`, `vscode`, `grok_build`, `omp`, `pi`,
-`hermes`, or `grok_bot`. If unknown, inspect `doctor --json` before choosing a
+`copilot_cli`, `commandcode`, `qwen_code`, `vscode`, `trae`, `grok_build`, `omp`, `pi`,
+`hermes`, `qoder`, `droid`, `zcode`, `dsh`, or `grok_bot`. If unknown, inspect `doctor --json` before choosing a
 scope. `grok_bot` uses a chat registration step rather than local `mcp.json`.
 For unsupported clients, use version-pinned manual stdio registration or CLI.

@@ -11,6 +11,19 @@ returned inline. When complete `inspect_binary_layout` or
 `inspect_analysis_view` for a summary, one section/symbol/module, or a stable
 page instead of repeating the producer.

+Completed `inspect_artifact` inventory also supports a summary and
+`view: {kind: "inventory", collection: "occurrences", logical_paths: ["package.json", "main.js"]}`.
+Omit `logical_paths` to page all available occurrences; `[]` selects none.
+Other inventory collections are `nodes`, `edges` and
+`integrity_contradictions`. Keep `source_complete` separate from page exhaustion:
+a missing match in partial source Evidence does not establish absence. Use the
+returned parent Evidence ID for subsequent views and export before closing the
+session when the source must survive it.
+When `tools/list` advertises `inspect_artifact.detail`, request `"summary"` to
+retain its complete source and receive that view immediately. Without that
+advertised choice, keep the existing complete call. CLI inspection returns
+portable complete Evidence; save it before projecting a selected inline view.
+
 Use `record_unknown` to track unresolved questions in the user's investigation
 and name the authority or environment still required. Supply supporting and
 contradicting evidence IDs. It updates the session registry directly.
@@ -46,3 +59,13 @@ observed exit signal stays in `exit.signal`, `exit.code` is null for every
 deadline exit, and the code of a finalizer that exited on its own is in
 `exit.finalization.exit_code` (null on a signal exit or when no exit was
 observed).
+
+Use `inspect_analysis_view` with `view: {kind: "api-surface"}` on retained
+`inspect_web_network_capture` HAR Evidence to group observed request lines by
+method, origin and path template. Format matches are inferences, not handler
+identity or schemas. Group pages cite parent record ordinals; consult record
+coverage and exclusions before claiming completeness. No requests are replayed.
+When the connected schema advertises `include_query_parameter_names`, select
+it in the view for distinct observed names with original record ordinals. Query
+values are omitted; the names establish neither required parameters nor a
+complete server schema. Keep the original parent Evidence for exact URLs.
@@ -1,7 +1,30 @@
 # JavaScript and Electron application artifacts

+For captured or bundled JavaScript that needs readable source recovery, use
+`recover_javascript_sources` with an absolute `path` and absent
+`output_directory`. Supply related runtime/shared/lazy files in
+`additional_paths` when advertised, then analyze the returned `analysis_input`
+directory together. Each input retains its own snapshot, digest and byte-range
+provenance; inputs and recovered code are never executed. Recovery requires a
+caller-supplied Wakaru on Linux x64; see [the recovery guide](https://github.com/morluto/rea/blob/main/docs/javascript-recovery.md).
+
 Use `analyze_javascript_application` directly on the operator-supplied ASAR or
-extracted tree. When size is unknown or large and the connected schema advertises
+extracted tree. When the question concerns known source paths and the connected
+schema advertises `source_paths`, select exact inventoried paths such as
+`["package.json", "main.js", "preload.js"]`. Include supporting manifests or
+source maps explicitly; dependencies are not implicitly analyzed. REA retains
+the complete artifact inventory and identity but reads source text and analyzes
+only selected files. The result and summary record `source_selection`; files
+outside scope and their relationships remain unknown, so partial application
+coverage does not mean the selected source failed. Use repeated `--source-path`
+on either static application CLI route. Omit the selection when the question
+requires complete application analysis. `detail: "summary"` reduces delivery
+and does not by itself reduce analysis work.
+
+Exact `#alias` specifiers resolve through the importer's
+enclosing package `imports` map to package-local files; nested `node_modules`
+packages use their own map. Wildcard, recursive, and external-package import
+targets stay unresolved with that limitation. When size is unknown or large and the connected schema advertises
 `detail`, request `detail: "summary"`. This returns an `inspect_analysis_view`
 Evidence record with a distinct ID and a `normalized_result.parent_evidence_id`
 for the complete retained application analysis. Source-map contents are part of
@@ -12,7 +35,10 @@ the complete application Evidence inline when it fits.
 When the connected schema advertises `max_heap_mb` and `analysis_timeout_ms`,
 these select the isolated source analyzer's old-space heap and per-source
 deadline. Defaults are 1024 MiB and 300000 ms. They do not change the MCP
-process heap. Resource-limited analyses preserve completed static facts and
+process heap. A Bun-hosted REA does not enforce the heap choice; on Windows it
+instead limits the analyzer's committed memory to that choice plus a 512 MiB
+runtime reserve.
+Resource-limited analyses preserve completed static facts and
 report partial semantic coverage with specific unknown scopes; this does not
 mean the source is malformed. Inspect those limits before retrying a source.
 When expanded relationships exceed the parent's available memory, inspect the
@@ -1,5 +1,43 @@
 # Native, managed, and packaged artifacts

+## Go build metadata
+
+Use `inspect_go_binary` for an explicit local ELF, PE or thin Mach-O when the
+question concerns the embedded compiler version, module/dependency versions,
+local replacements or build settings. No `open_binary`, Go installation or
+target execution is required. The CLI equivalent is `inspect-go-binary PATH`.
+Preserve artifact SHA-256, original file ranges, framed module bytes and every
+unparsed module-text line. A missing record is an unknown, not proof of a
+non-Go binary; observed metadata does not authenticate its claims. Function,
+source and type recovery remain a separate native-analysis question.
+
+## Rust dependency metadata
+
+Use `inspect_rust_binary` for an explicit local ELF, PE, or thin Mach-O when
+the question concerns the cargo-auditable crate list embedded in `.dep-v0`
+(Mach-O: `__DATA,.dep-v0`). No `open_binary`, Rust installation, or target
+execution is required. The CLI equivalent is `inspect-rust-binary PATH`.
+Preserve artifact SHA-256, package order, dependency indices, and the section
+file range. A missing section is null metadata, not proof the binary has no
+dependencies. An omitted package kind stays null. A rustc commit hash is
+inferred from a read-only `/rustc/<40-hex>/` path and does not by itself
+identify the compiler. See [Rust binaries](https://github.com/morluto/rea/blob/main/docs/rust-binaries.md).
+
+## SQLite snapshots
+
+Use `inspect_sqlite_database` with an explicit local database snapshot for its
+schema and optional rows from one exact ordinary table. It requires Node.js
+24.x >=24.15 or 26+ with the native `DatabaseSync.limits` API. It also consumes the
+existing sibling `-wal`; retain both original digests when citing its result.
+Default output is a compact inventory; a selected table expands only its own
+schema and samples 20 rows. Select `columns` to avoid irrelevant or large cells,
+and request `schema_scope: "database"` only for complete schema details.
+Choose `row_limit` when a finite sample is useful and keep truncation and
+unspecified row order explicit. Virtual tables, shadow tables and views are
+schema facts, not row-read targets. REA inspects private copies and leaves
+the source files untouched; obtain a coherent quiescent database/WAL snapshot
+before inspection. See [SQLite snapshots](https://github.com/morluto/rea/blob/main/docs/sqlite-databases.md).
+
 ## Recorded Linux crashes

 Use `inspect_recorded_crash` with an explicit Linux x86-64 ELF core path to
@@ -25,6 +63,20 @@ retained, call `inspect_analysis_view` for a summary, mitigations or linkage
 facet, one section or symbol, or a stable page. Do not re-run the decoder to
 read a single object.

+## Offline WASM
+
+Use `inspect_wasm_artifact` with caller-supplied WABT for selected local bytes.
+For counts, validation and WAT identity, select `detail: "summary"` on the first
+MCP call. The complete original Evidence stays retained; use
+`normalized_result.parent_evidence_id` for exact exports or text pages. Omitted
+or complete detail keeps the existing complete result.
+For a completed result, use `inspect_analysis_view` with its reported retained
+reference for a summary or `view: {kind: "wasm", facet: "exports", offset: 0, limit: 32}`.
+Facets also include sections, imports, candidates, glue, WAT, headers, details
+and tool profile. Text offsets count UTF-16 code units; follow `next_offset`.
+The complete source remains retained after an oversized MCP reply. No target
+execution or network acquisition is implied, and WAT is decoded representation.
+
 ## Native targets

 After `open_binary`, use focused search, procedure, or function tools directly.
@@ -43,6 +95,14 @@ names, the native API record, and value-flow summaries. Pseudocode offsets and l
 code units; use the returned `next_offset` for continuation. Provider limitations
 and unavailable facts remain explicit; the complete parent Evidence is retained.

+`inspect_macho` on a large framework returns megabytes of imports and exports.
+Select `detail: "summary"` for slice identity, UUID and per-collection counts,
+then page rows with `view: {kind: "macho", facet: "exports", offset: 0, limit: 64}`
+and `normalized_result.parent_evidence_id`. Facets are entrypoints,
+architectures, build_metadata, load_commands, dependencies, imports, exports,
+segments and provenance. `source_exhaustive: false` means unreturned symbols
+may exist.
+
 To see which functions or Objective-C methods a Mach-O actually calls in one
 run, use `observe_native_calls` with explicit breakpoints and a bounded
 `duration_ms`/`max_events`. It launches a new process under LLDB, so the target
@@ -50,10 +110,19 @@ runs and may change files or use the network. Hardened-runtime targets need the
 `get-task-allow` entitlement.

 Use `binary_session` with no arguments to check the open target, selected
-provider, and alignment. Its default `result.tool_availability` includes the
-complete tool inventory with availability reasons and remediation. When choosing
-a tool, use its entry in that result; `tools/list` retains the complete catalog
-when the target or provider state changes.
+provider, and alignment. Its default `result.tool_availability` carries a
+complete inventory grouped by shared facts. Expand each group's `names` and
+inherit the inventory-level client requirements unless a group overrides them;
+availability, reasons, remediation, modes, and flags remain inline. The
+`effects` and `annotations` arrays name true flags from fixed vocabularies, so an
+omitted known flag is false. `tools/list` remains the canonical source for full
+tool schemas when target or provider state changes.
+
+Provider capability groups carry provider identity and ordered limitations.
+Prepend each set's `shared_limitations` to a group's `limitations` to recover the
+operation's full ordered list. Candidate capabilities describe possible
+providers. Top-level session capabilities describe the current route, and
+`analysis_provider_binding` identifies the selected provider.

 ## DOS MZ with Ghidra

@@ -62,6 +131,8 @@ loaded bytes, file mappings, relocations and the entry against the immutable
 snapshot. Keep `mismatch` and `unsupported` explicit. Verified import covers the
 reported static image, not runtime DOS or PC-98 hardware behavior.

+For Ghidra decompilation, `procedure_pseudo_code` also accepts `procedure: "file-offset:0x..."`; CLI `rea decompile <path> --file-offset=0x... --provider ghidra` resolves the source byte offset internally. An unresolved 0x-prefixed numeric procedure selector also falls back to a unique file mapping. Existing symbols and analysis addresses win; multiple mappings fail as ambiguous. Preserve the returned file-offset and address locations and raw selection observations. ELF virtual addresses are distinct from file offsets; never assume an arbitrary offset plus image base is an analysis address.
+
 Use `read_bytes` for initialized analysis memory and `address_to_file_offset` to
 anchor one observed address to original file bytes. Loader fixups can change a
 word; a source offset does not imply byte equality. Complete function body ranges
@@ -86,6 +157,18 @@ assembly. Keep managed/native boundaries and unavailable reconstruction facts
 explicit. A bring-your-own reconstruction oracle is separate from the canonical
 parser and must not become an implicit setup dependency.

+For identity/classification questions, `inspect_managed_artifact` can explicitly
+select `detail: "summary"` on the first MCP call. It retains complete metadata
+and returns a linked overview with full classification/evidence, nullable
+module/MVID, frameworks, metadata coverage/issues/limitations and returned-row
+counts. Default complete requests remain unchanged. Read original rows with
+`inspect_analysis_view`, source `normalized_result.parent_evidence_id`, and
+`view: {kind: "managed", collection: "attributes", offset: 0, limit: 16}`.
+Collections are `references`, `resources` and `attributes`. Paging does not
+repeat parsing. Returned-row exhaustion
+does not prove complete metadata or absence under partial/unavailable coverage;
+preserve build-local tokens and their artifact/MVID scope.
+
 ### .NET NativeAOT

 NativeAOT binaries contain native machine code, so ordinary CIL decompilers
@@ -101,15 +184,17 @@ them. For packaged apps, inventory/extract with REA's artifact tools and retain
 the exact contained image identity. PDB, ELF debug, or dSYM sidecars are
 additional evidence only when the provider accepts and matches them.

-For deeper type recovery, the optional third-party
-[Ghidra NativeAOT analyzer](https://github.com/Washi1337/ghidra-nativeaot)
-rehydrates ReadyToRun metadata and can recover method tables, type relationships,
-frozen objects, and strings. This is a Ghidra extension with its own install and
-interactive workflow; it is not currently part of REA's managed parser or
-headless Ghidra bridge. Treat its annotations as analysis evidence, retain the
-binary identity, and do not present recovered types as original source. Header
-discovery may require a symbol or analyst identification of the ReadyToRun
-header, depending on the binary and analyzer version.
+REA has a built-in read-only snapshot parser for x64 PE targets with .NET 8 RTR
+9.1 metadata. An optional headless Ghidra adapter adds database annotations and
+broader metadata recovery on Linux; it requires an explicitly configured
+NativeAOT JAR. These are separate routes. The built-in parser also runs on
+Windows Ghidra P0; read frozen literals from its report, because
+`search_strings` does not include them. See the
+[NativeAOT guide](https://github.com/morluto/rea/blob/main/docs/ghidra-nativeaot.md)
+for setup, supported layouts, and host restrictions. Preserve binary identity
+and distinguish derived metadata from Ghidra database observations. Recovered
+types and native pseudocode are not original C# source. Windows Ghidra P0 has
+no database mutation authority.

 ## Packages and extraction

@@ -133,8 +218,9 @@ every executable in the bundle. It also lists missing or weak loads and earlier
 `architecture`. System paths stay undetermined because the dyld shared cache
 provides them.

-Use `extract_artifact` when materialized files are needed. It takes no arguments
-and materializes all regular files into a fresh temporary directory chosen by
+Use `extract_artifact` when materialized files are needed. It extracts the active
+target and accepts optional `integrity_policy`; see the [MCP contracts](https://github.com/morluto/rea/blob/main/docs/mcp-contracts.md).
+It materializes all regular files into a fresh temporary directory chosen by
 REA. Symlinks and encrypted entries are inventory facts, not extractable files.
 The result reports the materialized directory; callers do not choose the
 destination. No separate permission grant is needed.
@@ -150,6 +236,16 @@ For headerless DOS COM, explicitly open with `format: "dos-com"` (CLI
 analysis: the entry is `0x10100`, file offset zero, with imposed real-mode segment
 context. PSP/stack/device state remains unmodeled. See [DOS guide](https://github.com/morluto/rea/blob/main/docs/ghidra-dos.md).

+For other headerless images (embedded firmware without ELF/Mach-O/PE/MZ
+headers), open with `format: "raw-binary"`, `architecture` (`x86`, `x86_64`,
+`arm`, `arm64`, or `6502`), and `entry_point` (`0x` hex). File offset 0 is mapped at
+that address. Hopper and IDA refuse this interpretation; Ghidra uses
+`BinaryLoader` with a little-endian generic language. Big-endian, Thumb-only,
+and Cortex-M variants are unsupported. `6502` selects NMOS 6502; the entry and
+complete file mapping must fit within `0x0000..0xffff`. 65C02, 65816, cartridge
+containers, bank switching, and mapper semantics are unsupported. Windows
+Ghidra P0 admits raw images only for `6502`. See [raw binary guide](https://github.com/morluto/rea/blob/main/docs/raw-binary.md).
+
 ### Function annotations in Ghidra

 Use `annotate_native_function` for one function name and/or entry comments, with at least one explicit change. Review its annotation readback and refreshed dossier inline. Changes are atomic and session-scoped; empty comments clear them, omitted fields preserve them. Later MCP calls observe edits until close. CLI `annotate-native-function` returns the updated analysis before discarding the session. Original executable bytes are unchanged; edits invalidate immutable snapshots. Windows P0 does not admit database mutations. For naming only, `set_address_name` / `set_addresses_names` also work with Ghidra: they name function entries or data addresses (primary labels) and return compact per-address success booleans instead of a dossier.
@@ -165,3 +261,9 @@ resolution is explicit; missing references and size mismatch remain visible.
 Other payloads are opaque, and this operation does not decode images or simulate
 Windows resource loading. Consult the budgets and supported tree profile in
 [offline binary diagnostics](https://github.com/morluto/rea/blob/main/docs/binary-diagnostics.md).
+
+For exact core-module import/export differences, use `compare_wasm_interfaces`
+with two complete inspection Evidence records (or current-session retained
+references) and explicit scope. Preserve unknown types/duplicate names; static
+interface equality is not runtime compatibility. CLI accepts complete inline
+Evidence in `rea compare-wasm-interfaces input.json --json`.
@@ -26,3 +26,18 @@ passive browser/Electron Evidence exist. Prefer captured-source digest identity;
 report path/digest disagreement as a mismatch and preserve ambiguity. Explicit
 path mappings are inference inputs and never broaden filesystem or CDP authority.
 A bundle observed at runtime does not prove every contained module executed.
+
+## Fiddler Everywhere traffic debugging
+
+For large saved captures, request `detail: "summary"`, then page retained Evidence
+with `inspect_analysis_view` and `view: {kind: "capture", detail: "preview", offset: 0, limit: 20}`.
+Select complete record pages only for needed headers, metadata or payloads; views
+read the historical parent without repeating capture decoding.
+
+Inspect native `.saz` snapshots with `inspect_web_network_capture` and
+`format: "saz"`; encrypted archives are unsupported. For live
+traffic analysis, connect Everywhere's separate MCP server and use its
+advertised session tools with an explicit `sessionsSource` (`LiveTraffic` or
+`AgentCalls`). Preserve existing sessions and filters when inspecting a
+recorded run. See the [capture and MCP guide](https://github.com/morluto/rea/blob/main/docs/web-network-captures.md)
+for the workflow and Telerik's public setup and traffic-debugging skills.
@@ -0,0 +1,21 @@
+{
+  "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
+  "name": "rea",
+  "description": "Reverse engineer anything with agents, from app behavior down to native binaries.",
+  "owner": {
+    "name": "morluto",
+    "url": "https://github.com/morluto"
+  },
+  "plugins": [
+    {
+      "name": "rea",
+      "description": "Reverse engineer anything from your terminal or agent with one CLI and MCP server.",
+      "source": {
+        "source": "npm",
+        "package": "rea-agents",
+        "version": "6.4.0"
+      },
+      "category": "development"
+    }
+  ]
+}
@@ -0,0 +1,28 @@
+{
+  "name": "rea",
+  "version": "6.4.0",
+  "description": "Reverse engineer anything from your terminal or agent with one CLI and MCP server.",
+  "author": {
+    "name": "morluto",
+    "url": "https://github.com/morluto"
+  },
+  "homepage": "https://rea.tools/",
+  "repository": "https://github.com/morluto/rea",
+  "license": "MIT",
+  "keywords": [
+    "reverse-engineering",
+    "binary-analysis",
+    "decompiler",
+    "mcp-server"
+  ],
+  "skills": "./skills",
+  "mcpServers": {
+    "rea": {
+      "command": "node",
+      "args": [
+        "${CLAUDE_PLUGIN_ROOT}/scripts/plugin-mcp-launcher.mjs",
+        "6.4.0"
+      ]
+    }
+  }
+}
@@ -0,0 +1,38 @@
+{
+  "name": "rea",
+  "version": "6.4.0",
+  "description": "Reverse engineer anything from your terminal or agent with one CLI and MCP server.",
+  "author": {
+    "name": "morluto",
+    "url": "https://github.com/morluto"
+  },
+  "homepage": "https://rea.tools/",
+  "repository": "https://github.com/morluto/rea",
+  "license": "MIT",
+  "keywords": [
+    "reverse-engineering",
+    "binary-analysis",
+    "decompiler",
+    "mcp-server"
+  ],
+  "skills": "./skills",
+  "mcpServers": "./.mcp.json",
+  "interface": {
+    "displayName": "REA",
+    "shortDescription": "Reverse engineer anything with agents",
+    "longDescription": "Use REA to explain shipped behavior from native binaries, managed assemblies, Electron and JavaScript bundles, packages, firmware, and browser targets, with evidence that separates observations from inferences.",
+    "developerName": "morluto",
+    "category": "Developer Tools",
+    "capabilities": [
+      "Interactive",
+      "Read",
+      "Write"
+    ],
+    "websiteURL": "https://rea.tools/",
+    "defaultPrompt": [
+      "Use REA to explain how this binary implements a feature.",
+      "Use REA to compare two versions of this application.",
+      "Use REA to recover the sources from this JavaScript bundle."
+    ]
+  }
+}
@@ -0,0 +1,8 @@
+# Keep text checkouts consistent with the repository formatter on every platform.
+* text=auto eol=lf
+
+# Windows command fixtures retain their native line endings.
+*.cmd text eol=crlf
+
+# Preserve whitespace in original upstream licensing texts.
+licenses/npm-runtime-notices.txt whitespace=-blank-at-eol
@@ -1,5 +1,8 @@
 blank_issues_enabled: false
 contact_links:
+  - name: Contribution scope and integration requirements
+    url: https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope
+    about: Check supported capabilities, excluded workflows, and the fail-closed integration rule before proposing a feature.
  - name: REA Documentation
    url: https://github.com/morluto/rea#readme
    about: Read the README and AGENTS.md for usage guides and configuration details.
@@ -1,5 +1,5 @@
 name: Design discussion
-description: Resolve an REA architecture, provider, or contract question before implementation
+description: Discuss an REA architecture, provider, contract, or scope decision
 title: "[Design] "
 labels:
  - enhancement
@@ -9,7 +9,12 @@ body:
    attributes:
      value: |
        Use this for a substantial provider, public/persisted contract, or
-        authority-boundary decision. Keep one decision per discussion.
+        authority-boundary decision. One question and a short explanation are
+        enough to start. Choose "Not sure yet" where needed; maintainers can help
+        with technical details, alternatives, and verification.
+
+        Read the [contribution scope](https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope).
+        Unresolved boundaries block integration, not opening a discussion.

  - type: checkboxes
    id: before-submitting
@@ -18,14 +23,16 @@ body:
      options:
        - label: I searched the open and closed issues and pull requests for a related discussion.
          required: true
-        - label: I removed credentials, tokens, private paths, and sensitive target data.
+        - label: I understand the [contribution scope](https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope) must be reviewed before integration; open questions are welcome here.
+          required: true
+        - label: This discussion does not include credentials, private samples, or sensitive target data.
          required: true

  - type: textarea
    id: decision
    attributes:
      label: Decision needed
-      description: State the exact question maintainers need to resolve.
+      description: What would you like maintainers to help decide? A short question is enough.
      placeholder: Should REA ... or ...?
    validations:
      required: true
@@ -34,23 +41,57 @@ body:
    id: context-evidence
    attributes:
      label: Context and evidence
-      description: Explain why the decision is needed now. Separate observed behavior, derived evidence, and inference.
+      description: Explain the problem or motivation. If you have evidence, distinguish observed behavior from ideas or guesses.
+    validations:
+      required: true
+
+  - type: dropdown
+    id: target-interaction
+    attributes:
+      label: Expected interaction with targets
+      description: >-
+        Select all known effects, including upstream tool behavior. Add "Not sure
+        yet" if anything is unclear. Use "No target interaction" only on its own.
+        These selections help discussion and do not establish scope approval.
+      multiple: true
+      options:
+        - Read existing files or captured data
+        - Launch or attach to a process
+        - Inject instrumentation or run scripts/code
+        - Change process memory or registers
+        - Write or change files/configuration
+        - Contact network services
+        - Recover protected code or work around protections
+        - No target interaction (documentation or usability only)
+        - Not sure yet
    validations:
      required: true

  - type: textarea
    id: boundary-invariants
    attributes:
-      label: Boundary and invariants
-      description: Identify the owning layer, provider/bridge boundary, artifact or evidence invariants, and any process or authorization constraints.
-    validations:
-      required: true
+      label: Boundary details (optional)
+      description: >-
+        If known, clarify target selection, selected effects, provider boundaries,
+        non-goals, or evidence invariants. Leave blank if you need help identifying
+        them. Do not include private samples or proof-of-authorization documents.

  - type: textarea
    id: options
    attributes:
      label: Options and tradeoffs
-      description: Compare a small number of viable approaches, including platform, compatibility, safety, evidence, and maintenance costs.
+      description: Optional approaches or tradeoffs. One idea is enough; maintainers can help compare alternatives.
+
+  - type: dropdown
+    id: verification-readiness
+    attributes:
+      label: Verification available so far
+      description: Pick the closest match. An idea without samples or test results is welcome.
+      options:
+        - Idea only; I need help defining checks
+        - I can suggest a sample or manual check
+        - I have a prototype or observed results
+        - Not sure yet
    validations:
      required: true

@@ -58,9 +99,10 @@ body:
    id: success-criteria
    attributes:
      label: Success criteria and open questions
-      description: State what outcome would settle the decision and list only unresolved questions that could change the choice.
-    validations:
-      required: true
+      description: >-
+        Optional expected output, sample, possible check, or unresolved question.
+        A complete test plan is not required to start; maintainers and implementers
+        will resolve material verification gaps before integration.

  - type: textarea
    id: references
@@ -8,9 +8,13 @@ body:
  - type: markdown
    attributes:
      value: |
-        Keep one primary capability or workflow outcome per request. For a
-        substantial provider, public contract, or architecture change, explain
-        the decision maintainers need to make before implementation.
+        Describe one problem and the outcome you would like. Short answers are
+        welcome. Choose "Not sure yet" wherever needed; you do not need code, a
+        finished design, or a complete test plan to propose an idea.
+
+        Read the [contribution scope](https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope).
+        Maintainers can help resolve scope and technical questions. Unresolved
+        boundaries block integration, not opening a request for discussion.

  - type: checkboxes
    id: before-submitting
@@ -19,7 +23,9 @@ body:
      options:
        - label: I searched the open and closed issues and pull requests for a related proposal.
          required: true
-        - label: I removed credentials, tokens, private paths, and sensitive target data.
+        - label: I understand the [contribution scope](https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope) must be reviewed before integration; open questions are welcome here.
+          required: true
+        - label: This request does not include credentials, private samples, or sensitive target data.
          required: true

  - type: dropdown
@@ -36,6 +42,7 @@ body:
        - MCP or agent integration
        - Documentation
        - Other
+        - Not sure yet
    validations:
      required: true

@@ -43,7 +50,7 @@ body:
    id: current-limitation
    attributes:
      label: Current limitation
-      description: What user or agent workflow is blocked, ambiguous, or unnecessarily difficult today?
+      description: What are you trying to understand or do, and what is difficult today? A short example is enough.
    validations:
      required: true

@@ -51,23 +58,69 @@ body:
    id: use-cases
    attributes:
      label: Use cases
-      description: Give concrete applications, targets, callers, or investigation workflows. Explain why current tools are insufficient.
-    validations:
-      required: true
+      description: Optional additional examples of applications, targets, or callers. Leave blank if covered above.

  - type: textarea
    id: desired-outcome
    attributes:
      label: Desired outcome and success criteria
-      description: Describe observable behavior, inputs, outputs, evidence, and compatibility criteria that would make this useful.
+      description: What would a useful result look like? Describe the output or behavior you want; implementation details are optional.
+      placeholder: For a selected APK, show whether it is packed and include the detection evidence.
+    validations:
+      required: true
+
+  - type: dropdown
+    id: target-interaction
+    attributes:
+      label: Expected interaction with targets
+      description: >-
+        Select all known effects, including upstream tool behavior. Add "Not sure
+        yet" if anything is unclear. Use "No target interaction" only on its own.
+        These selections help discussion and do not establish scope approval.
+      multiple: true
+      options:
+        - Read existing files or captured data
+        - Launch or attach to a process
+        - Inject instrumentation or run scripts/code
+        - Change process memory or registers
+        - Write or change files/configuration
+        - Contact network services
+        - Recover protected code or work around protections
+        - No target interaction (documentation or usability only)
+        - Not sure yet
    validations:
      required: true

  - type: textarea
    id: boundaries
    attributes:
-      label: Authority boundaries and non-goals
-      description: Call out process execution, target mutation, provider licensing, evidence provenance, supported platforms, and work that should remain out of scope.
+      label: Target and scope details (optional)
+      description: >-
+        If known, name the target/platform, clarify selected effects, or list a
+        non-goal or scope question. You can leave this blank; maintainers can help.
+        Do not include private samples or proof-of-authorization documents.
+
+  - type: dropdown
+    id: verification-readiness
+    attributes:
+      label: Verification available so far
+      description: Pick the closest match. An idea without samples or test results is welcome.
+      options:
+        - Idea only; I need help defining checks
+        - I can suggest a sample or manual check
+        - I have a prototype or observed results
+        - Not sure yet
+    validations:
+      required: true
+
+  - type: textarea
+    id: verification
+    attributes:
+      label: Verification details (optional)
+      description: >-
+        Share an expected output, a possible check, or existing results if you have
+        them. Self-owned or appropriately licensed samples can help, but are not
+        required to open a request. Do not attach private or sensitive data.

  - type: textarea
    id: alternatives
@@ -0,0 +1,42 @@
+name: Cache npm package downloads
+description: Restore npm's download cache and save it only from trusted default-branch events.
+outputs:
+  can-save:
+    description: Whether this run may update the shared npm download cache.
+    value: ${{ steps.context.outputs.can-save }}
+  path:
+    description: The npm download-cache directory.
+    value: ${{ steps.context.outputs.path }}
+runs:
+  using: composite
+  steps:
+    - id: context
+      shell: bash
+      env:
+        DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
+        LOCKFILES_HASH: ${{ hashFiles('npm-shrinkwrap.json', '.npmrc') }}
+      run: |
+        cache_path="$(npm config get cache)"
+        runtime_arch="$(node -p 'process.arch')"
+        can_save=false
+        if [[ -n "${DEFAULT_BRANCH}" && "${GITHUB_REF}" == "refs/heads/${DEFAULT_BRANCH}" && ( "${GITHUB_EVENT_NAME}" == "push" || "${GITHUB_EVENT_NAME}" == "schedule" || "${GITHUB_EVENT_NAME}" == "workflow_dispatch" ) ]]; then
+          can_save=true
+        fi
+        cache_key="npm-download-${RUNNER_OS}-${runtime_arch}-${LOCKFILES_HASH}"
+        {
+          printf 'can-save=%s\n' "${can_save}"
+          printf 'path=%s\n' "${cache_path}"
+          printf 'key=%s\n' "${cache_key}"
+        } >> "${GITHUB_OUTPUT}"
+    - name: Restore and save npm package downloads
+      if: steps.context.outputs.can-save == 'true'
+      uses: actions/cache@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
+      with:
+        path: ${{ steps.context.outputs.path }}
+        key: ${{ steps.context.outputs.key }}
+    - name: Restore npm package downloads without saving
+      if: steps.context.outputs.can-save != 'true'
+      uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
+      with:
+        path: ${{ steps.context.outputs.path }}
+        key: ${{ steps.context.outputs.key }}
@@ -0,0 +1,53 @@
+name: Portable build cache
+description: Restore portable Turbo build outputs from the current run or cross-run cache.
+inputs:
+  save:
+    description: Save newly built outputs after the job succeeds.
+    default: "false"
+  build_cache_artifact:
+    description: Reuse portable Turbo outputs uploaded by the current workflow run.
+    default: ""
+runs:
+  using: composite
+  steps:
+    - name: Download portable build outputs from this workflow run
+      if: inputs.build_cache_artifact != ''
+      uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
+      with:
+        name: ${{ inputs.build_cache_artifact }}
+        path: .turbo/cache/
+    - name: Resolve portable build task hashes
+      id: plan
+      if: inputs.build_cache_artifact == ''
+      shell: bash
+      env:
+        BUILD_CACHE_PREFIX: rea-portable-build-v2-${{ runner.os }}-${{ runner.arch }}-${{ hashFiles('.npmrc', '.nvmrc', 'npm-shrinkwrap.json') }}
+      run: |
+        npm run --silent test:prepare -- --dry=json |
+          node --input-type=module -e '
+            import { createHash } from "node:crypto";
+            import { readFileSync } from "node:fs";
+            const { tasks } = JSON.parse(readFileSync(0, "utf8"));
+            const hashes = tasks.map(({ taskId, hash }) => {
+              if (typeof hash !== "string")
+                throw new Error(`Build task ${taskId} must have an eager Turbo hash`);
+              return `${taskId}:${hash}`;
+            }).sort();
+            const hash = createHash("sha256").update(hashes.join("\n")).digest("hex");
+            console.log(`key=${process.env.BUILD_CACHE_PREFIX}-${hash}`);
+            console.log(`prefix=${process.env.BUILD_CACHE_PREFIX}-`);
+          ' >> "$GITHUB_OUTPUT"
+    - name: Restore portable build outputs
+      if: inputs.build_cache_artifact == '' && inputs.save != 'true'
+      uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
+      with:
+        path: .turbo/cache/
+        key: ${{ steps.plan.outputs.key }}
+        restore-keys: ${{ steps.plan.outputs.prefix }}
+    - name: Restore and save portable build outputs
+      if: inputs.build_cache_artifact == '' && inputs.save == 'true'
+      uses: actions/cache@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
+      with:
+        path: .turbo/cache/
+        key: ${{ steps.plan.outputs.key }}
+        restore-keys: ${{ steps.plan.outputs.prefix }}
@@ -19,17 +19,39 @@ describe the use case and observable outcome. -->
 <!-- Explain the chosen approach, why it fits REA's architecture, and what is
 intentionally not included. -->

+Follow the [contribution scope](https://github.com/morluto/rea/blob/main/CONTRIBUTING.md#contribution-scope).
+Unresolved scope, authority, side effects, or required verification block
+integration. Link the required maintainer scope decision for new providers or
+capabilities involving target execution, injection, mutation, or protection
+recovery; ordinary fixes within existing boundaries may state "not applicable".
+Draft PRs may leave the scope decision pending and the final review checklist
+unchecked. Maintainers can help resolve open questions.
+
 ## Contract and boundary impact

-<!-- Complete the applicable lines; use "none" or "not applicable" explicitly. -->
+<!-- Short answers are welcome. Complete applicable lines; mark unresolved facts as "unknown".
+Use "none" or "not applicable" only when established. -->

 - Semantic owner and earliest changed stage:
+- Contribution scope fit and linked maintainer decision:
 - CLI and MCP/tool-catalog contract:
 - Provider, bridge, target-format, or platform compatibility:
 - Evidence, artifact, provenance, or reconstruction contract:
- Process execution, authorization, cleanup, or containment impact:
+- Target selection, authority, upstream behavior, and cleanup details (if applicable):
 - Generated metadata (`docs/public/product-catalog.json`), package, or installation impact (CI generates and validates ignored catalogs and `skills/`; edit authored instructions in `.agents/skills/reverse-engineer-anything/` and never commit build outputs):

+Expected effects (check known effects; use documentation-only on its own).
+Unchecked items do not establish absence; leave uncertainties visible.
+
+- [ ] Read existing files or captured data.
+- [ ] Launch or attach to a process.
+- [ ] Inject instrumentation or run scripts/code.
+- [ ] Change process memory, registers, files, or configuration.
+- [ ] Contact network services.
+- [ ] Recover protected code or work around protections.
+- [ ] Documentation or usability only; no target interaction.
+- [ ] Not sure yet; investigation or maintainer guidance needed.
+
 ## Evidence and regression coverage

 <!-- State whether evidence is executed, source-derived, or proposed. For fixes,
@@ -68,9 +90,13 @@ limits, security/privacy/containment implications, and meaningful proof gaps. --
 - Package or release metadata impact:
 - Security, privacy, process, or containment review:

-## Review checklist
+## Before merge
+
+<!-- Draft PRs may leave pending items unchecked. Resolve material scope and
+authority questions before integration. -->

 - [ ] This PR addresses a concrete problem or an agreed enhancement.
+- [ ] Material scope and authority questions are resolved, and any required maintainer scope decision covers the final diff.
 - [ ] The PR has one focused outcome and the title follows `type(scope): outcome`.
 - [ ] Related issue is linked, or the reason for not linking one is stated above.
 - [ ] Tests cover changed observable behavior and meaningful failure paths.
@@ -0,0 +1,18 @@
+name: Full PR qualification
+
+on:
+  pull_request:
+    branches: [main, "release/*"]
+    types: [labeled]
+
+permissions:
+  contents: read
+  actions: read
+
+jobs:
+  qualification:
+    name: Qualify labeled PR revision
+    if: github.event.label.name == 'ci:full'
+    uses: ./.github/workflows/ci.yml
+    with:
+      source_sha: ${{ github.sha }}
@@ -0,0 +1,84 @@
+name: CI timings
+
+on:
+  schedule:
+    - cron: "17 * * * *"
+  workflow_dispatch:
+
+concurrency:
+  group: ci-timings-${{ github.repository }}
+  cancel-in-progress: false
+
+permissions:
+  contents: read
+  actions: read
+  pull-requests: read
+
+jobs:
+  report:
+    name: Collect recent completed workflow attempts
+    runs-on: ubuntu-latest
+    timeout-minutes: 10
+    steps:
+      # Consume only API data and execute batch code from the trusted default branch.
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+        with:
+          ref: ${{ github.event.repository.default_branch }}
+          persist-credentials: false
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        continue-on-error: true
+        uses: ./.github/actions/npm-cache
+      - name: Diagnose missing and startup-failed PR CI
+        continue-on-error: true
+        env:
+          GH_TOKEN: ${{ github.token }}
+          CI_REPOSITORY: ${{ github.repository }}
+          CI_REPOSITORY_ID: ${{ github.repository_id }}
+          CI_DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
+        run: node scripts/ci/pr-ci-health.mjs
+      - name: Record queue, execution, and matched step timing
+        continue-on-error: true
+        env:
+          GH_TOKEN: ${{ github.token }}
+          CI_REPOSITORY: ${{ github.repository }}
+        run: node scripts/ci/job-timings-batch.mjs
+      - name: Upload CI timing artifacts
+        if: always() && hashFiles('.cache/ci-job-timings/*') != ''
+        continue-on-error: true
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        with:
+          name: ci-job-timings-batch
+          path: .cache/ci-job-timings/
+          include-hidden-files: true
+          if-no-files-found: error
+          overwrite: true
+          retention-days: 14
+      - name: Install dependencies for static failure attribution
+        id: failure-deps
+        if: ${{ !cancelled() && hashFiles('.cache/ci-job-timings/batch-manifest.json') != '' }}
+        continue-on-error: true
+        run: npm ci
+      - name: Attribute supported static failures against main
+        if: ${{ !cancelled() && steps.failure-deps.outcome == 'success' && hashFiles('.cache/ci-job-timings/batch-manifest.json') != '' }}
+        continue-on-error: true
+        env:
+          GH_TOKEN: ${{ github.token }}
+          CI_REPOSITORY: ${{ github.repository }}
+          CI_DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
+        run: node scripts/ci/failure-report.mjs
+      - name: Upload CI health and failure diagnostics
+        if: ${{ !cancelled() }}
+        continue-on-error: true
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        with:
+          name: ci-health-diagnostics
+          path: |
+            .cache/ci-pr-health/
+            .cache/ci-failures/
+          include-hidden-files: true
+          if-no-files-found: warn
+          retention-days: 14
@@ -12,22 +12,23 @@ jobs:
    name: Build Windows x64 NativeAOT fixture
    # .NET 8.0.22's bundled linker discovery supports VS 2019/2022.
    runs-on: windows-2022
-    timeout-minutes: 15
+    # Two bounded restore attempts and one compile can each take five minutes.
+    timeout-minutes: 20
    env:
      DOTNET_CLI_TELEMETRY_OPTOUT: "1"
      DOTNET_PROCESSOR_COUNT: "1"
      COMPlus_GCHeapHardLimit: "30000000"
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-      - uses: actions/setup-dotnet@v4
+      - uses: actions/setup-dotnet@67a3573c9a986a3f9c594539f4ab511d57bb3ce9 # v4.3.1
        with:
          dotnet-version: "8.0.416"
      - name: Build benign fixture and independent byte oracle
        run: node scripts/generate-nativeaot-fixtures.mjs
-      - uses: actions/upload-artifact@v4
+      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
        with:
          name: nativeaot-win-x64-${{ github.event.pull_request.head.sha || github.sha }}
          path: |
@@ -19,10 +19,12 @@ jobs:
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
      - run: npm run docs:build
@@ -0,0 +1,29 @@
+name: PR title
+
+on:
+  pull_request:
+    branches: [main, "release/*"]
+    types: [opened, synchronize, reopened, edited]
+
+permissions: {}
+
+concurrency:
+  group: pr-title-${{ github.event.pull_request.number }}
+  cancel-in-progress: true
+
+jobs:
+  pr-title:
+    name: Conventional PR title
+    runs-on: ubuntu-24.04
+    timeout-minutes: 2
+    steps:
+      - name: Validate title for Release Please
+        env:
+          PR_TITLE: ${{ github.event.pull_request.title }}
+        run: >-
+          node -e 'const title = process.env.PR_TITLE ?? "";
+          const conventional = /^(?:feat|fix|docs|refactor|perf|test|build|ci|chore|revert)(?:\([a-z0-9._/-]+\))?!?: .+/u;
+          if (!conventional.test(title)) {
+            console.error("PR title must use Conventional Commits, for example: feat: add historical source import");
+            process.exit(1);
+          }'
@@ -8,14 +8,16 @@ on:
        type: boolean
        default: false
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""

 permissions:
  contents: read

-concurrency:
-  group: real-binary-layout-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
 env:
  NODE_OPTIONS: --max-old-space-size=768
  HUSKY: "0"
@@ -26,12 +28,17 @@ jobs:
    runs-on: ubuntu-24.04
    timeout-minutes: 15
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - name: Acquire exact upstream libraries in an isolated runner environment
        shell: bash
        run: |
@@ -2,12 +2,14 @@ name: Real browser verification

 on:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""
  workflow_dispatch:

-concurrency:
-  group: real-browser-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
-
 env:
  HUSKY: "0"

@@ -19,12 +21,17 @@ jobs:
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
-          node-version: 24
-          cache: npm
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - run: npm run verify:browser
        env:
          REA_BROWSER_EXECUTABLE: /usr/bin/google-chrome
@@ -0,0 +1,76 @@
+name: Real upstream Cutter verification
+
+on:
+  workflow_dispatch:
+  repository_dispatch:
+    types: [real-cutter-upstream]
+
+permissions:
+  contents: read
+
+jobs:
+  windows-native-artifact:
+    name: Build Windows bridge security boundary
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
+      - name: Install the cross compiler used by the Windows native lane
+        run: |
+          sudo apt-get update
+          sudo apt-get install --no-install-recommends -y g++-mingw-w64-x86-64
+      - run: npm run build:windows-native
+      - uses: actions/upload-artifact@v4
+        with:
+          name: cutter-windows-native
+          path: native/windows/build/
+          if-no-files-found: error
+
+  cutter-upstream:
+    name: Cutter upstream smoke (${{ matrix.platform }})
+    needs: windows-native-artifact
+    runs-on: ${{ matrix.runner }}
+    timeout-minutes: 20
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - platform: linux
+            runner: [self-hosted, cutter-upstream, linux]
+          - platform: macOS
+            runner: [self-hosted, cutter-upstream, macOS]
+          - platform: Windows
+            runner: [self-hosted, cutter-upstream, Windows]
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
+      - run: npm ci
+      - if: matrix.platform == 'Windows'
+        uses: actions/download-artifact@v4
+        with:
+          name: cutter-windows-native
+          path: native/windows/build/
+      - if: matrix.platform == 'Windows'
+        run: npm run verify:windows-native:artifact
+      - if: matrix.platform == 'Windows'
+        run: npm run build:cached
+      - if: matrix.platform == 'Windows'
+        run: npm run verify:windows-native
+      - name: Verify host and real Cutter bridge
+        env:
+          EXPECTED_PLATFORM: ${{ matrix.platform }}
+        run: npm run verify:cutter
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: cutter-upstream-${{ matrix.platform }}
+          path: cutter-smoke.json
+          if-no-files-found: ignore
+          retention-days: 30
@@ -3,14 +3,16 @@ name: Real offline EVM interface
 on:
  workflow_dispatch:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""

 permissions:
  contents: read

-concurrency:
-  group: real-evm-interface-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
 env:
  NODE_OPTIONS: --max-old-space-size=768
  HUSKY: "0"
@@ -20,12 +22,17 @@ jobs:
    runs-on: ubuntu-24.04
    timeout-minutes: 15
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - name: Acquire unchanged fixture compiler in isolated runner storage
        shell: bash
        run: |
@@ -38,3 +45,5 @@ jobs:
          echo "REA_VERIFY_SOLC_MODULE=${evm_fixture_root}/node_modules/solc/index.js" >> "${GITHUB_ENV}"
      - name: Verify source-owned compiled bytecode through public CLI and MCP
        run: npm run verify:evm:interface
+      - name: Verify content-hash-free and CBOR-free owned compiler artifacts
+        run: npm run verify:solidity:metadata
@@ -16,8 +16,8 @@ jobs:
    if: github.ref == 'refs/heads/main'
    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
      - name: Prepare cross compiler
@@ -31,11 +31,25 @@ jobs:
          sudo apt-get install --no-install-recommends -y g++-mingw-w64-x86-64
      - run: npm run build:windows-native
      - run: npm run verify:windows-native:artifact
-      - uses: actions/upload-artifact@v4
+      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
        with:
          name: real-ghidra-windows-native
          path: native/windows/build/
          if-no-files-found: error
+      # The boundary observations need a separate Windows executable, which the
+      # native lane skips when it is absent. Cross-build it here, where MinGW is
+      # already installed, so the proof run observes them too.
+      - name: Build Windows process boundary fixture
+        run: |
+          mkdir -p .cache/windows-process-boundary
+          x86_64-w64-mingw32-g++ -municode -std=c++17 -O2 -static -Wall -Wextra \
+            -o .cache/windows-process-boundary/process-boundary.exe \
+            tests/fixtures/windows/processBoundary.cc -ladvapi32
+      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        with:
+          name: real-ghidra-windows-process-boundary
+          path: .cache/windows-process-boundary/
+          if-no-files-found: error

  verify:
    name: Real Ghidra 12.1.4 Windows x64 PE proof
@@ -45,17 +59,23 @@ jobs:
    runs-on: [self-hosted, Windows, x64, ghidra-12-1-2]
    timeout-minutes: 20
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
      - run: npm ci
-      - uses: actions/download-artifact@v4
+      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
        with:
          name: real-ghidra-windows-native
          path: native/windows/build/
      - run: npm run verify:windows-native:artifact
-      - run: npm run verify:windows-native
+      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
+        with:
+          name: real-ghidra-windows-process-boundary
+          path: .cache/windows-process-boundary/
+      - run: >-
+          npm run verify:windows-native -- .
+          "${{ github.workspace }}\.cache\windows-process-boundary\process-boundary.exe"
      - name: Verify all admitted Windows Ghidra operations
        env:
          REA_ANALYSIS_PROVIDER: ghidra
@@ -88,7 +108,7 @@ jobs:
        env:
          REA_ANALYSIS_PROVIDER: ghidra
        run: npm run verify:ghidra:windows:package -- --x86 --dll | Tee-Object -FilePath windows-ghidra-x86-dll-package-proof.log
-      - uses: actions/upload-artifact@v4
+      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
        if: always()
        with:
          name: windows-ghidra-proof
@@ -0,0 +1,60 @@
+name: Real Go binary metadata
+
+on:
+  workflow_dispatch:
+  pull_request:
+    paths:
+      - ".github/workflows/real-go-binary.yml"
+      - "src/go/**"
+      - "src/domain/go/**"
+      - "src/domain/dosMz.ts"
+      - "src/application/go/**"
+      - "src/contracts/go/**"
+      - "src/composition/go.ts"
+      - "src/cli/goCommands.ts"
+      - "src/server/registerGoTools.ts"
+      - "src/server/createServer.ts"
+      - "src/cli.ts"
+      - "src/cliCommandNames.ts"
+      - "src/contracts/toolContracts.ts"
+      - "src/contracts/toolEffects.ts"
+      - "src/artifacts/readStableArtifact*.ts"
+      - "src/artifacts/ArtifactReader.ts"
+      - "src/process/PrivateRuntimeRoot.ts"
+      - "scripts/verify-go-binary.mjs"
+      - "scripts/lib/mcp-verifier-results.mjs"
+      - "scripts/rea.mjs"
+      - "package.json"
+      - "npm-shrinkwrap.json"
+      - ".nvmrc"
+
+permissions:
+  contents: read
+
+concurrency:
+  group: real-go-binary-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+env:
+  NODE_OPTIONS: --max-old-space-size=768
+  HUSKY: "0"
+
+jobs:
+  real-go-binary:
+    runs-on: ubuntu-24.04
+    timeout-minutes: 20
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
+      - uses: actions/setup-go@40f1582b2485089dde7abd97c1529aa768e1baff # v5.6.0
+        with:
+          go-version: "1.26.1"
+          cache: false
+      - run: npm ci
+      - name: Verify compiler-produced ELF, PE and Mach-O through public CLI and MCP
+        run: npm run verify:go:binary
@@ -15,11 +15,13 @@ jobs:
    runs-on: ubuntu-24.04
    timeout-minutes: 25
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
      - name: Pack CLI for setup verification
        shell: bash
@@ -36,11 +38,13 @@ jobs:
    runs-on: [self-hosted, Linux, hopper]
    timeout-minutes: 20
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
      - name: Install packed CLI in isolation
        shell: bash
@@ -15,11 +15,13 @@ jobs:
    runs-on: [self-hosted, macOS, hopper]
    timeout-minutes: 20
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
      - name: Install packed CLI in isolation
        shell: bash
@@ -2,15 +2,17 @@ name: Real JavaScript recovery

 on:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""
  workflow_dispatch:

 permissions:
  contents: read

-concurrency:
-  group: javascript-recovery-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
-
 env:
  HUSKY: "0"
  NODE_OPTIONS: "--max-old-space-size=768"
@@ -21,12 +23,17 @@ jobs:
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
-          node-version: 24
-          cache: npm
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - name: Fetch pinned Wakaru release into the job workspace
        shell: bash
        run: |
@@ -0,0 +1,60 @@
+name: Real packet capture
+on:
+  workflow_dispatch:
+  pull_request:
+    paths:
+      - "tests/boundary/captures/**"
+      - "tests/boundary/cli/captureCleanup.test.ts"
+      - "tests/boundary/mcp/packetCaptureMcp.test.ts"
+      - "src/captures/**"
+      - "src/pcap/**"
+      - "src/domain/pcap/**"
+      - "src/domain/analysisView/pcapCaptureView.ts"
+      - "src/application/pcap/**"
+      - "src/contracts/pcap/**"
+      - "src/server/registerPcapTools.ts"
+      - "tests/boundary/pcap/**"
+      - "src/domain/captures/**"
+      - "src/application/captures/**"
+      - "src/contracts/captures/**"
+      - "src/composition/captures.ts"
+      - "src/cli/captureCommands.ts"
+      - "src/server/registerCaptureTools.ts"
+      - "src/server/createServer.ts"
+      - "src/server/sessionAvailabilityPolicy.ts"
+      - "src/application/CapabilityInventory.ts"
+      - "src/contracts/toolContracts.ts"
+      - "src/contracts/toolEffects.ts"
+      - "src/cli.ts"
+      - "src/cliCommandNames.ts"
+      - "src/process/**"
+      - "src/artifacts/readStableArtifact.ts"
+      - "scripts/verify-packet-capture.mjs"
+      - ".github/workflows/real-packet-capture.yml"
+      - "package*.json"
+      - ".nvmrc"
+permissions:
+  contents: read
+concurrency:
+  group: real-packet-capture-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+jobs:
+  tshark:
+    runs-on: ubuntu-24.04
+    timeout-minutes: 15
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
+      - run: sudo apt-get update && sudo DEBIAN_FRONTEND=noninteractive apt-get install -y tshark util-linux python3
+      - run: npm ci
+      - run: npm run build:cached
+      - run: npm run test:focused -- tests/boundary/captures/fieldDiscovery.test.ts src/process/LinuxResourceLimits.test.ts tests/boundary/captures/captureProvider.test.ts tests/boundary/captures/captureDescriptorCleanup.test.ts tests/boundary/captures/resourceLimits.test.ts tests/boundary/captures/packetOutput.test.ts tests/boundary/mcp/packetCaptureMcp.test.ts tests/boundary/cli/captureCleanup.test.ts
+      - run: node scripts/verify-packet-capture.mjs
+        env:
+          REA_TSHARK_COMMAND: /usr/bin/tshark
+          REA_CAPINFOS_COMMAND: /usr/bin/capinfos
@@ -3,14 +3,16 @@ name: Real recorded crash evidence
 on:
  workflow_dispatch:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""

 permissions:
  contents: read

-concurrency:
-  group: real-recorded-crash-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
 env:
  NODE_OPTIONS: --max-old-space-size=768
  HUSKY: "0"
@@ -23,12 +25,17 @@ jobs:
    runs-on: ubuntu-24.04
    timeout-minutes: 15
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - name: Install GDB on the disposable CI runner
        shell: bash
        run: |
@@ -0,0 +1,52 @@
+name: Real signal recording
+on:
+  workflow_dispatch:
+  pull_request:
+    paths:
+      - "tests/boundary/signals/**"
+      - "tests/boundary/cli/signalCleanup.test.ts"
+      - "tests/boundary/mcp/signalRecordingMcp.test.ts"
+      - "src/signals/**"
+      - "src/domain/signals/**"
+      - "src/application/signals/**"
+      - "src/contracts/signals/**"
+      - "src/composition/signals.ts"
+      - "src/cli/signalCommands.ts"
+      - "src/server/registerSignalTools.ts"
+      - "src/server/createServer.ts"
+      - "src/server/sessionAvailabilityPolicy.ts"
+      - "src/application/CapabilityInventory.ts"
+      - "src/contracts/toolContracts.ts"
+      - "src/contracts/toolEffects.ts"
+      - "src/cli.ts"
+      - "src/cliCommandNames.ts"
+      - "src/process/**"
+      - "src/artifacts/readStableArtifact.ts"
+      - "scripts/verify-signal-recording.mjs"
+      - ".github/workflows/real-signal-recording.yml"
+      - "package*.json"
+      - ".nvmrc"
+permissions:
+  contents: read
+concurrency:
+  group: real-signal-recording-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+jobs:
+  sigrok:
+    runs-on: ubuntu-24.04
+    timeout-minutes: 15
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
+      - run: sudo apt-get update && sudo apt-get install -y sigrok-cli util-linux python3
+      - run: npm ci
+      - run: npm run build:cached
+      - run: npm run test:focused -- tests/boundary/signals/decoderDiscovery.test.ts src/process/LinuxResourceLimits.test.ts src/domain/signals/annotationTextWindow.test.ts tests/boundary/signals/recordingProvider.test.ts tests/boundary/signals/recordingDescriptorCleanup.test.ts tests/boundary/signals/resourceLimits.test.ts tests/boundary/mcp/signalRecordingMcp.test.ts tests/boundary/cli/signalCleanup.test.ts
+      - run: node scripts/verify-signal-recording.mjs
+        env:
+          REA_SIGROK_COMMAND: /usr/bin/sigrok-cli
@@ -18,7 +18,7 @@ jobs:
    runs-on: ubuntu-latest
    timeout-minutes: 45
    steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
      - name: Download the pinned Termux app and verify its release checksum
        shell: bash
        run: |
@@ -38,7 +38,7 @@ jobs:
          sudo udevadm control --reload-rules
          sudo udevadm trigger --name-match=kvm
      - name: Build and capture a page inside Termux
-        uses: reactivecircus/android-emulator-runner@4c44018e59b437e86cdfc41da381398f93ed8808 # v2
+        uses: reactivecircus/android-emulator-runner@a421e43855164a8197daf9d8d40fe71c6996bb0d # v2.38.0
        with:
          api-level: 30
          arch: x86_64
@@ -50,7 +50,7 @@ jobs:
          script: bash scripts/verify/termux/emulator.sh
      - name: Upload Termux and Android diagnostics
        if: always()
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
        with:
          name: termux-browser-diagnostics
          path: ${{ runner.temp }}/rea-termux/diagnostics
@@ -3,14 +3,16 @@ name: Real historical web network captures
 on:
  workflow_dispatch:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""

 permissions:
  contents: read

-concurrency:
-  group: real-web-network-captures-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
 env:
  HUSKY: "0"
  NODE_OPTIONS: --max-old-space-size=768
@@ -20,14 +22,19 @@ jobs:
    runs-on: ubuntu-24.04
    timeout-minutes: 15
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
-          cache: npm
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - name: Use pinned package manager
        run: npm install --global npm@11.16.0
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - name: Acquire fixed upstream native profile in runner temporary storage
        shell: bash
        run: |
@@ -1,12 +1,15 @@
 name: Real website runtime attribution
 on:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""
  workflow_dispatch:
 permissions:
  contents: read
-concurrency:
-  group: real-web-runtime-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
 env:
  HUSKY: "0"
 jobs:
@@ -14,13 +17,18 @@ jobs:
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
-          node-version: 24
-          cache: npm
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
-      - run: npm run verify:browser:runtime
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
+      - run: env -u DBUS_SESSION_BUS_ADDRESS npm run verify:browser:runtime
        env:
          REA_BROWSER_EXECUTABLE: /usr/bin/google-chrome
          REA_BROWSER_NO_SANDBOX: "true"
@@ -1,12 +1,15 @@
 name: Real website source-map verification
 on:
  workflow_call:
+    inputs:
+      build_cache_artifact:
+        description: Portable build cache artifact from the calling CI run.
+        required: false
+        type: string
+        default: ""
  workflow_dispatch:
 permissions:
  contents: read
-concurrency:
-  group: real-web-source-map-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
 env:
  HUSKY: "0"
 jobs:
@@ -14,12 +17,17 @@ jobs:
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
-          node-version: 24
-          cache: npm
+          node-version-file: .nvmrc
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
      - run: npm ci
+      - uses: ./.github/actions/portable-build-cache
+        with:
+          build_cache_artifact: ${{ inputs.build_cache_artifact }}
      - name: Install the pinned compiler into an isolated fixture directory
        run: npm install --prefix "$RUNNER_TEMP/rea-web-source-map-compiler" --no-audit --no-fund --package-lock=false esbuild@0.25.10
      - run: npm run verify:browser:source-maps
@@ -48,7 +48,7 @@ jobs:
            echo "frozen=true" >> "$GITHUB_OUTPUT"
          fi

-      - uses: googleapis/release-please-action@v4
+      - uses: googleapis/release-please-action@5c625bfb5d1ff62eadeeb3772007f7f66fdcf071 # v4.4.1
        if: steps.editorial.outputs.frozen != 'true'
        with:
          token: ${{ secrets.RELEASE_PLEASE_TOKEN || secrets.GITHUB_TOKEN }}
@@ -70,13 +70,13 @@ jobs:
      sha: ${{ steps.release.outputs.sha }}
    steps:
      - name: Check out reviewed release merge
-        uses: actions/checkout@v6
+        uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0
        with:
          ref: ${{ github.event.pull_request.merge_commit_sha }}
          fetch-depth: 0
          persist-credentials: false

-      - uses: actions/setup-node@v6
+      - uses: actions/setup-node@249970729cb0ef3589644e2896645e5dc5ba9c38 # v6.5.0
        with:
          node-version-file: .nvmrc
          package-manager-cache: false
@@ -104,7 +104,7 @@ jobs:
          node scripts/release-notes.mjs publication --repository "$GH_REPO" \
            --source "$SOURCE_SHA" --pr "$RELEASE_PR"

-      - uses: googleapis/release-please-action@v4
+      - uses: googleapis/release-please-action@5c625bfb5d1ff62eadeeb3772007f7f66fdcf071 # v4.4.1
        id: release
        with:
          token: ${{ secrets.RELEASE_PLEASE_TOKEN || secrets.GITHUB_TOKEN }}
@@ -191,14 +191,14 @@ jobs:
          fi

      - name: Check out release controller and Git history
-        uses: actions/checkout@v6
+        uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0
        with:
          ref: ${{ github.sha }}
          fetch-depth: 0
          persist-credentials: false

      - name: Set up Node.js for checkpoint validation
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@249970729cb0ef3589644e2896645e5dc5ba9c38 # v6.5.0
        with:
          node-version-file: .nvmrc
          package-manager-cache: false
@@ -226,7 +226,7 @@ jobs:
          node scripts/release-notes.mjs publication --repository "$GH_REPO" \
            --source "$SOURCE_SHA"

-      - uses: googleapis/release-please-action@v4
+      - uses: googleapis/release-please-action@5c625bfb5d1ff62eadeeb3772007f7f66fdcf071 # v4.4.1
        id: release
        with:
          # A dedicated credential starts PR CI automatically; the built-in
@@ -255,7 +255,7 @@ jobs:

      - name: Check out release pull request
        if: inputs.phase == 'prepare' && steps.release.outputs.prs_created == 'true'
-        uses: actions/checkout@v6
+        uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0
        with:
          ref: ${{ fromJSON(steps.release.outputs.pr).headBranchName }}
          fetch-depth: 0
@@ -263,7 +263,7 @@ jobs:

      - name: Set up Node.js for generated documentation
        if: inputs.phase == 'prepare' && steps.release.outputs.prs_created == 'true'
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@249970729cb0ef3589644e2896645e5dc5ba9c38 # v6.5.0
        with:
          node-version-file: .nvmrc
          package-manager-cache: false
@@ -290,24 +290,36 @@ jobs:
        if: inputs.phase == 'prepare' && steps.release.outputs.prs_created == 'true'
        run: npm run docs:check && git diff --exit-code

-  publish:
-    name: Publish to npm
+  qualification:
+    name: Qualify exact release revision
    needs: [merged-release, release-please]
    if: >-
      always() && !cancelled() &&
      ((needs.merged-release.result == 'success' && needs.merged-release.outputs.release_created == 'true') ||
      (needs.release-please.result == 'success' && inputs.phase == 'publish' && needs.release-please.outputs.release_created == 'true'))
+    permissions:
+      contents: read
+      actions: read
+    uses: ./.github/workflows/ci.yml
+    with:
+      source_sha: ${{ needs.merged-release.outputs.sha || needs.release-please.outputs.sha }}
+
+  publish:
+    name: Publish to npm
+    needs: [merged-release, release-please, qualification]
+    if: >-
+      always() && !cancelled() && needs.qualification.result == 'success'
    runs-on: ubuntu-latest
    permissions:
      contents: read
      id-token: write
    steps:
-      - uses: actions/checkout@v6
+      - uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0
        with:
          ref: ${{ needs.merged-release.outputs.sha || needs.release-please.outputs.sha }}
          persist-credentials: false

-      - uses: actions/setup-node@v6
+      - uses: actions/setup-node@249970729cb0ef3589644e2896645e5dc5ba9c38 # v6.5.0
        with:
          node-version-file: .nvmrc
          package-manager-cache: false
@@ -315,17 +327,20 @@ jobs:

      - run: npm ci

-      - name: Prepare Windows native build toolchain
-        run: sudo apt-get update && sudo apt-get install --no-install-recommends -y g++-mingw-w64-x86-64
+      - name: Download the exact qualified package
+        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
+        with:
+          name: verified-release-package
+          path: .cache/verified-package/

-      - name: Build and verify bundled Windows native controls
-        run: npm run build:windows-native && npm run verify:windows-native:artifact
-
-      - name: Verify packaged release before publishing
-        run: npm run verify:package
+      - name: Verify qualified package provenance and digest
+        id: package
+        run: node scripts/ci/verified-package.mjs validate .cache/verified-package

      - name: Publish
        shell: bash
+        env:
+          VERIFIED_PACKAGE_TARBALL: ${{ steps.package.outputs.tarball }}
        run: |
          version="$(node --input-type=module -e 'import packageJson from "./package.json" with { type: "json" }; process.stdout.write(packageJson.version)')"
          published_version="$(npm view "rea-agents@${version}" version --json 2>/dev/null || true)"
@@ -333,7 +348,7 @@ jobs:
            echo "rea-agents@${version} is already published; verifying the existing package."
          else
            tag="$(node scripts/release-npm-tag.mjs "${version}")"
-            npm publish --access public --tag "${tag}"
+            npm publish "${VERIFIED_PACKAGE_TARBALL}" --access public --tag "${tag}"
          fi

      - name: Verify published CLI from npm
@@ -377,7 +392,7 @@ jobs:
      contents: read
      id-token: write
    steps:
-      - uses: actions/checkout@v6
+      - uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0
        with:
          ref: ${{ needs.merged-release.outputs.sha || needs.release-please.outputs.sha }}
          persist-credentials: false
@@ -0,0 +1,39 @@
+name: SQLite runtime compatibility
+
+on:
+  workflow_call:
+
+permissions:
+  contents: read
+
+env:
+  HUSKY: "0"
+
+jobs:
+  sqlite:
+    name: SQLite runtime compatibility (Node ${{ matrix.node }})
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    strategy:
+      fail-fast: false
+      matrix:
+        # Exercise unavailable-runtime diagnostics, the first supported release,
+        # and new patches on both supported lines without hand-editing versions.
+        # The repository-pinned Node 24 version also runs the complete test suite.
+        node: ["22.x", "24.15.0", "24.x", "26.x"]
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version: ${{ matrix.node }}
+          check-latest: true
+      - name: Cache npm package downloads
+        id: npm-cache
+        uses: ./.github/actions/npm-cache
+      - run: npm ci
+      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
+        with:
+          name: rea-test-runtime
+          path: .
+      - name: Verify SQLite discovery, inspection and cleanup
+        run: npm run test:sqlite:run
@@ -7,10 +7,6 @@ on:
 permissions:
  contents: read

-concurrency:
-  group: website-check-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
 jobs:
  check:
    name: Website checks
@@ -18,9 +14,9 @@ jobs:
    timeout-minutes: 10
    steps:
      - name: Check out the repository
-        uses: actions/checkout@v4
+        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
      - name: Set up Python for website assets
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
        with:
          python-version: "3.12"
          cache: pip
@@ -47,9 +47,9 @@ jobs:
            exit 1
          fi
      - name: Check out the repository
-        uses: actions/checkout@v4
+        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
      - name: Set up Python for website assets
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
        with:
          python-version: "3.12"
          cache: pip
@@ -68,17 +68,17 @@ jobs:
          import os
          from pathlib import Path

-          Path("website/public/deployment.json").write_text(
+          Path("website/_site/deployment.json").write_text(
              json.dumps({"commit": os.environ["GITHUB_SHA"]}) + "\n",
              encoding="utf-8",
          )
          PY
      - name: Configure Pages
-        uses: actions/configure-pages@v5
+        uses: actions/configure-pages@983d7736d9b0ae728b81ab479565c72886d7745b # v5.0.0
      - name: Upload the website
-        uses: actions/upload-pages-artifact@v4
+        uses: actions/upload-pages-artifact@7b1f4a764d45c48632c6b24a0339c27f5614fb0b # v4.0.0
        with:
-          path: website/public
+          path: website/_site

  deploy:
    needs: prepare
@@ -95,15 +95,15 @@ jobs:
    steps:
      - name: Check out publishing configuration
        if: inputs.target == 'both'
-        uses: actions/checkout@v4
+        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
      - name: Set up Node.js for Wrangler
        if: inputs.target == 'both'
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
        with:
          node-version-file: .nvmrc
      - name: Download the verified website
        if: inputs.target == 'both'
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
        with:
          name: github-pages
          path: ${{ runner.temp }}/website-archive
@@ -114,7 +114,7 @@ jobs:
          tar -xf "$RUNNER_TEMP/website-archive/artifact.tar" -C "$RUNNER_TEMP/rea-website-public"
      - name: Publish the website to Cloudflare
        if: inputs.target == 'both'
-        uses: cloudflare/wrangler-action@v4
+        uses: cloudflare/wrangler-action@953926a2e2182532811c01a25e53647d93bf07c0 # v4.1.3
        with:
          apiToken: ${{ secrets.CLOUDFLARE_API_TOKEN }}
          accountId: ${{ secrets.CLOUDFLARE_ACCOUNT_ID }}
@@ -123,7 +123,7 @@ jobs:
          command: deploy --cwd website --env production --assets "${{ runner.temp }}/rea-website-public"
      - name: Publish to GitHub Pages
        id: deployment
-        uses: actions/deploy-pages@v4
+        uses: actions/deploy-pages@d6db90164ac5ed86f2b6aed7e0febac5b3c0c03e # v4.0.5
      - name: Verify the selected published revisions
        env:
          PAGES_URL: ${{ steps.deployment.outputs.page_url }}
@@ -33,8 +33,12 @@ codedb.snapshot
 htmlcov/
 .coverage
 coverage/
+.stryker-tmp/
+reports/mutation/
+.vitest/
 .vitest-reports/
 test-results.xml
+test-timings.json
 .turbo/
 .cache/
 docs/.vitepress/cache/
@@ -46,6 +50,7 @@ Thumbs.db

 # Local reference sources and analysis artifacts
 /_reference/
+/experiments/analysis-efficiency/results/

 # Generated website download
 /website/public/examples/notes-example.zip
@@ -67,3 +72,8 @@ Thumbs.db
 /docs/public/product-catalog.json
 /docs/verification/managed-conformance-manifest.json
 /docs/verification/managed-conformance-ledger.json
+
+# Complete static website output, including generated language editions.
+website/_site/
+# Temporary translation output prepared before replacing the complete site.
+website/.i18n-*/
@@ -1,6 +1,5 @@
 {
-  "_comment": "Test files are intentionally NOT excluded: the gate used to report zero clones because it skipped **/*.test.ts, hiding all test-tree duplication. Current actual is ~2.47% lines across src+ and tests/, down from 3.4% when tests were included. The 2.6% limit leaves modest headroom for ordinary changes; lower it as duplication is removed.",
-  "threshold": 2.6,
+  "threshold": 3.5,
  "reporters": ["consoleFull"],
  "ignore": [
    "node_modules/**",
@@ -0,0 +1,10 @@
+{
+  "mcpServers": {
+    "rea": {
+      "command": "node",
+      "args": ["./scripts/plugin-mcp-launcher.mjs", "6.4.0"],
+      "cwd": ".",
+      "startup_timeout_sec": 30
+    }
+  }
+}
@@ -9,7 +9,9 @@
    "docs/api/**",
    "node_modules/**",
    "CHANGELOG.md",
-    "package-lock.json",
+    "npm-shrinkwrap.json",
+    ".claude-plugin/plugin.json",
+    ".codex-plugin/plugin.json",
    "_reference/**",
    "third_party/**"
  ]
@@ -22,7 +22,8 @@
    ],
    "max-params": ["warn", 4],
    "no-var": "warn",
-    "no-var-requires": "error"
+    "no-var-requires": "error",
+    "typescript/no-import-type-side-effects": "error"
  },
  "overrides": [
    {
@@ -1,3 +1,3 @@
 {
-  ".": "6.3.0"
+  ".": "6.4.0"
 }
@@ -4,6 +4,8 @@

 REA is a local-only reverse-engineering tool with shared CLI/MCP workflows. Distinguish observations, derivations, inferences, and unknowns; preserve Evidence provenance and session semantics through both adapters. Provider support and setup live in [README.md](README.md#choosing-a-deep-analysis-provider) and its linked guides.

+Follow the [contribution scope and fail-closed integration rule](CONTRIBUTING.md#contribution-scope). Do not integrate capabilities while material scope, authority, side-effect, or required verification questions remain unresolved. For new providers or capabilities involving target execution, injection, mutation, or protection recovery, seek the required maintainer scope decision before substantial implementation for integration. Welcome early proposals, small exploratory prototypes, and draft PRs that help resolve questions; they do not authorize integration.
+
 Ghidra uses ephemeral databases without modifying executable bytes or controlling a GUI; Windows Ghidra P0 has no mutation authority. IDA analysis is read-only; never save or close an attached GUI database.

 Configuration changes must be additive and idempotent, with backups. Installers must not install or upgrade unrelated software, including Homebrew, Node.js, npm, Java, or Ghidra. Ghidra and IDA are bring-your-own. `rea setup` must show its planned changes and require approval before writing files or installing Hopper.
@@ -14,14 +16,45 @@ Preserve caller-selected inputs and local evidence, including output, URLs, path

 Dependencies flow inward: domain semantics and contracts, provider adapters, shared application workflows, then CLI/MCP adapters. Keep engine protocols and provider-specific code out of domain/application layers. See [docs/architecture.mermaid](docs/architecture.mermaid) when changing composition.

-Give each interpretation and resource a clear owner. Preserve identity, omission, uncertainty, and source evidence across boundaries. Retain ownership of resources whose cleanup failed. Reuse `src/process/` supervision and identity primitives; a PID and executable path alone do not establish ownership after exit or reuse.
+Give each interpretation and resource a clear owner. Consumers must use that owner's parsed facts rather than reinterpret raw source. Preserve identity, omission, uncertainty, and source evidence across boundaries. Retain ownership of resources whose cleanup failed. Reuse `src/process/` supervision and identity primitives; a PID and executable path alone do not establish ownership after exit or reuse.
+
+Keep proof of ownership separate from validity, editability, and readiness. Malformed owned data may still be safely removable. Unresolved identity or applicability must remain unknown rather than be treated as absent, foreign, or aligned.

 Assess brittleness in touched code during implementation, debugging, review, and maintenance. Fix the producing representation or ownership rule before adding downstream repairs. Normalize once at the owning boundary using format-aware parsers. Remove superseded representations, aliases, wrappers, and redundant tests after verifying affected consumers. Introduce abstractions only when they remove observed complexity.

+When fixing a bug, inspect the producing boundary and affected consumers for duplicated interpretation or unclear ownership. Complete any directly related refactor in the same change and remove superseded code. Passing regression tests alone does not establish that the brittleness is resolved. If related cleanup remains, explain what remains and why.
+
+When addressing a review thread, inspect the producing boundary and affected consumers for brittleness. Complete directly related refactoring in the same change. Before resolving the thread, explain the root cause, structural fix, validation, and any remaining cleanup.
+
 Parse unknown boundary values and model expected failures with the tagged error algebra and `Result`. Preserve partial facts and meaningful failure reasons. Keep identity and lookup values separate from display formatting; preserve source values and explicit unknowns when normalization loses information. Absolute paths, file URLs, and HTTP paths have different semantics.

 Validate advertised JSON Schemas after SDK conversion against their declared dialect and actual clients. Bind handlers to named contracts with exact types; catalog ordering must never select an operation or schema.

+## Performance
+
+Design analysis workflows and agent interfaces for efficiency from the first
+implementation. Correctness, useful answers, Evidence provenance, caller choices,
+compatibility and safe resource bounds remain required.
+
+Optimize in this order:
+
+1. Agent tokens and context. Return the smallest sufficient answer with
+   necessary provenance, coverage and limitations. Use task selection,
+   summaries, paging and retained Evidence where supported; honor requested
+   detail and keep complete Evidence accessible. Count the whole task,
+   including schemas, extra calls and retries.
+2. Analysis time. Reuse validated results and shared work. Cache keys must
+   cover result-affecting inputs; define invalidation and lifetime. Validate
+   freshness before reporting cached live state as current.
+3. CPU and memory. Avoid unnecessary work and copies; bound concurrency,
+   retained state and serialization growth.
+
+For changes affecting these costs, measure representative completed tasks and
+compare equivalent workflows before and after where available. Report token
+usage, elapsed time, CPU time and peak process-family memory. Separate cold
+and warm runs, label estimates, and state unmeasured costs. Fewer bytes alone
+do not establish fewer tokens or faster analysis.
+
 ## Development

 Use [.nvmrc](.nvmrc) and `package.json#packageManager` for the pinned toolchain. Cursor Cloud may place older Node.js first on `PATH`; environment setup installs the pinned toolchain under `/usr/local/bin`.
@@ -1,5 +1,69 @@
 # Changelog

+## [6.4.0](https://github.com/morluto/rea/compare/rea-agents-6.3.0...rea-agents-6.4.0) (2026-10-11)
+
+### Highlights
+
+- Android: added device inspection, split-APK acquisition, logs, screenshots, file transfers and explicit app lifecycle actions through ADB, plus manifest and localized-string decoding with Apktool. ([#1552](https://github.com/morluto/rea/pull/1552), [#1556](https://github.com/morluto/rea/pull/1556)) Thanks @3QL.
+- JEB: browse project units and decompile selected items through a caller-started local JEB MCP server. ([#1536](https://github.com/morluto/rea/pull/1536)) Thanks @3QL.
+- Large analyses: discover retained Evidence without loading every payload, and inspect JavaScript graphs, binary layouts and function dossiers through summaries and selected views. ([#1682](https://github.com/morluto/rea/pull/1682), [#1568](https://github.com/morluto/rea/pull/1568), [#1570](https://github.com/morluto/rea/pull/1570), [#1345](https://github.com/morluto/rea/pull/1345), [#1507](https://github.com/morluto/rea/pull/1507), [#1247](https://github.com/morluto/rea/issues/1247)) Thanks @jhuang-tw, @kaoru0822-kitauji, @N0zoM1z0.
+- JavaScript: isolate source analysis in an owned process with configurable heap and timeout limits; retain completed facts and explicit unknowns when analysis runs out of resources. ([#1651](https://github.com/morluto/rea/pull/1651), [#1563](https://github.com/morluto/rea/issues/1563), [#1565](https://github.com/morluto/rea/issues/1565)) Thanks @N0zoM1z0, @peter-research.
+- Ghidra: configure import language/compiler specifications and seed functions, code and labels before auto-analysis. ([#1667](https://github.com/morluto/rea/pull/1667)) Thanks @sqmagellan.
+- RPC discovery: recognize supported tRPC calls, HTTP-link URLs and named GraphQL operations in JavaScript applications and web bundles. ([#1670](https://github.com/morluto/rea/pull/1670)) Thanks @pompiliodonatoapp.
+
+### Changes
+
+- Added Ghidra support for Windows PE DLLs and bounded MIPS ELF analysis, including carved MIPS inputs with incomplete section tables. Windows P0 remains read-only. ([#1303](https://github.com/morluto/rea/pull/1303), [#1362](https://github.com/morluto/rea/pull/1362), [#1629](https://github.com/morluto/rea/pull/1629), [#1290](https://github.com/morluto/rea/issues/1290), [#718](https://github.com/morluto/rea/issues/718), [#1619](https://github.com/morluto/rea/issues/1619)) Thanks @bqw371, @jaelliot, @kvalls, @N0zoM1z0.
+- Enabled `set_address_name` and `set_addresses_names` for Ghidra sessions with mutation authority, including namespace-aware data labels and per-address batch results. ([#1635](https://github.com/morluto/rea/pull/1635), [#1634](https://github.com/morluto/rea/issues/1634)) Thanks @giri256.
+- Recovered additional signed AArch64 relative jump-table patterns. ([#1592](https://github.com/morluto/rea/pull/1592)) Thanks @RishTheFish21.
+- Added PE resource inspection and read-only NativeAOT metadata parsing for supported x64 PE32+/.NET 8 RTR 9.1 inputs. The NativeAOT parser’s Windows-host Ghidra workflow remains unverified on a real Windows host. ([#1482](https://github.com/morluto/rea/pull/1482), [76d8f68](https://github.com/morluto/rea/commit/76d8f68cb918dfb43869f2b649a0b6c0c1cf266d), [#1254](https://github.com/morluto/rea/issues/1254)) Thanks @N0zoM1z0, @thangldw.
+- Added Pi, Hermes and Qwen Code setup, including Hermes active-profile support. ([#1376](https://github.com/morluto/rea/pull/1376), [#1403](https://github.com/morluto/rea/pull/1403), [#1415](https://github.com/morluto/rea/pull/1415), [#1409](https://github.com/morluto/rea/issues/1409)) Thanks @rasatpetabit, @yiliang114, @zaydiscold.
+- Added `REA_MCP_INPUT_SCHEMA_PROFILE=compact` for size-capped MCP clients such as Kimi/Moonshot, while retaining canonical runtime validation. ([#1514](https://github.com/morluto/rea/pull/1514), [#1580](https://github.com/morluto/rea/pull/1580), [#1484](https://github.com/morluto/rea/issues/1484), [#1581](https://github.com/morluto/rea/issues/1581)) Thanks @haythemsellami, @N0zoM1z0.
+- Added live process-capture progress and optional `finalization_ms` to capture shutdown output before timeout escalation; cancellation remains immediate. ([#1483](https://github.com/morluto/rea/pull/1483), [#1583](https://github.com/morluto/rea/pull/1583), [#1252](https://github.com/morluto/rea/issues/1252), [#1253](https://github.com/morluto/rea/issues/1253), [modelcontextprotocol/typescript-sdk#2997](https://github.com/modelcontextprotocol/typescript-sdk/issues/2997)) Thanks @N0zoM1z0, @nikolay-dementiev, @sathariels.
+- Improved Codex onboarding and recovery from observed client, PATH and provider blockers. ([#1613](https://github.com/morluto/rea/pull/1613)) Thanks @N0zoM1z0.
+- Reduced redirected JSON and terminal output size, and deferred heavy imports until their commands need them. ([#1413](https://github.com/morluto/rea/pull/1413), [#1450](https://github.com/morluto/rea/pull/1450), [#1439](https://github.com/morluto/rea/pull/1439), [21dbcf7](https://github.com/morluto/rea/commit/21dbcf7a1458219989ad291e44ce7d017378b727)) Thanks @kaoru0822-kitauji.
+
+### Fixes
+
+- Electron analysis: follow renamed and default-imported APIs and optional calls, ignore overwritten identities and media preload attributes, and resolve packaged `app.asar` pages and unpacked companions. ([#1391](https://github.com/morluto/rea/pull/1391), [#1393](https://github.com/morluto/rea/pull/1393), [#1400](https://github.com/morluto/rea/pull/1400), [#1664](https://github.com/morluto/rea/pull/1664), [#1673](https://github.com/morluto/rea/pull/1673)) Thanks @kaoru0822-kitauji, @marnixpluim, @N0zoM1z0, @suninrain086.
+- JavaScript data flow: preserve returned/yielded reference uncertainty, invalidate property-stored aliases correctly, and retain bundled-module runtime scope. ([#1437](https://github.com/morluto/rea/pull/1437), [#1445](https://github.com/morluto/rea/pull/1445), [#1471](https://github.com/morluto/rea/pull/1471), [#1577](https://github.com/morluto/rea/pull/1577), [#1428](https://github.com/morluto/rea/issues/1428), [#1442](https://github.com/morluto/rea/issues/1442), [#1449](https://github.com/morluto/rea/issues/1449), [#1576](https://github.com/morluto/rea/issues/1576)) Thanks @CassieFox23, @dakjdakd, @N0zoM1z0.
+- JavaScript packages: resolve exact exports subpaths, preserve trailing-slash directory requests and exact CommonJS keys, and respect shadowed `require` bindings. ([#1424](https://github.com/morluto/rea/pull/1424), [#1615](https://github.com/morluto/rea/pull/1615), [5485483](https://github.com/morluto/rea/commit/5485483f7da2f6cb66441bb784aaf2a1a5f8731a), [8b2f851](https://github.com/morluto/rea/commit/8b2f8514c50e469c3add8ce9ba7ebebef888b14f), [f998fcb](https://github.com/morluto/rea/commit/f998fcb1cca9994c3589fb70d6e0127fc215f76f), [1e51c73](https://github.com/morluto/rea/commit/1e51c732bbf08191688d6526bf58ab2650b601e3), [#1365](https://github.com/morluto/rea/issues/1365)) Thanks @akshitkhokhani, @AWUQAQ, @immanuwell, @rudycelekli.
+- Large JavaScript inputs: handle oversized inline source maps, deep returns and reassignment loops; avoid argument expansion limits and yield during large analysis passes. ([#1386](https://github.com/morluto/rea/pull/1386), [#1368](https://github.com/morluto/rea/pull/1368), [#1390](https://github.com/morluto/rea/pull/1390), [#1539](https://github.com/morluto/rea/pull/1539), [#1525](https://github.com/morluto/rea/pull/1525), [#1542](https://github.com/morluto/rea/pull/1542), [#1545](https://github.com/morluto/rea/pull/1545), [#1557](https://github.com/morluto/rea/pull/1557), [#1380](https://github.com/morluto/rea/issues/1380), [#1401](https://github.com/morluto/rea/issues/1401), [#1538](https://github.com/morluto/rea/issues/1538), [#1462](https://github.com/morluto/rea/issues/1462)) Thanks @cold-atom, @KadirJeager, @kaoru0822-kitauji, @marnixpluim, @N0zoM1z0, @rudycelekli.
+- JavaScript receivers: expand escaped methods once, index returned-object targets, and avoid replaying earlier writes or duplicating retained graph allocations. ([#1517](https://github.com/morluto/rea/pull/1517), [#1537](https://github.com/morluto/rea/pull/1537), [#1444](https://github.com/morluto/rea/pull/1444), [#1601](https://github.com/morluto/rea/pull/1601), [#1495](https://github.com/morluto/rea/issues/1495)) Thanks @CassieFox23, @geekwolf, @kaoru0822-kitauji, @N0zoM1z0, @zeynelacikgoez.
+- Browser captures: retain WebSocket URLs, empty connections and frame order; compare only common known payload coverage and preserve incomplete WebMCP replay coverage. ([db4e360](https://github.com/morluto/rea/commit/db4e3608465486466591b594cf77037ac150c988), [aa257ea](https://github.com/morluto/rea/commit/aa257ea4b16f2d2a35c12df63359d650b48b1154), [e52c0a7](https://github.com/morluto/rea/commit/e52c0a72c59c740315e89149881a27ce4d89a11c))
+- Browser scenarios: seed storage in the correct partition, normalize loopback attach identities, retain declared-secret redaction and require verified lineage before Windows Electron signals. ([#1604](https://github.com/morluto/rea/pull/1604), [#1590](https://github.com/morluto/rea/pull/1590), [#1588](https://github.com/morluto/rea/pull/1588), [8b4c3de](https://github.com/morluto/rea/commit/8b4c3de3b3a3b580adec266f94f29ccedf1da228), [#1497](https://github.com/morluto/rea/issues/1497), [#1591](https://github.com/morluto/rea/issues/1591), [#1589](https://github.com/morluto/rea/issues/1589)) Thanks @Adolanium, @akushonkamen, @dakjdakd.
+- Added Termux Android browser-scenario support. ([#1655](https://github.com/morluto/rea/pull/1655)) Thanks @tyloryang.
+- Web inputs: accept BOM-prefixed HAR, preserve the first duplicate DOM attribute and resolve SVG/XLink URLs in HTML and XHTML. ([#1509](https://github.com/morluto/rea/pull/1509), [#1516](https://github.com/morluto/rea/pull/1516), [#1602](https://github.com/morluto/rea/pull/1602), [#1683](https://github.com/morluto/rea/pull/1683), [#1496](https://github.com/morluto/rea/issues/1496), [#1586](https://github.com/morluto/rea/issues/1586)) Thanks @CassieFox23, @kaoru0822-kitauji, @sathariels, @taljeri.
+- Screenshot comparisons: validate PNG chunks, distinguish corruption from unsupported encodings and bound comparison memory. ([4e4b15b](https://github.com/morluto/rea/commit/4e4b15b8250b29e176d4191c9610a3931b693474), [b3e977e](https://github.com/morluto/rea/commit/b3e977e6d073ab1783cece8581279f87fc360411)) Thanks @rudycelekli.
+- Archives: preserve case-distinct names, report collisions at their diverging segment, detect nested formats from bytes and reject Windows alternate-stream extraction paths. ([#1480](https://github.com/morluto/rea/pull/1480), [#1505](https://github.com/morluto/rea/pull/1505), [#1553](https://github.com/morluto/rea/pull/1553), [573598c](https://github.com/morluto/rea/commit/573598c6324698cc48ceafd24ed3864670e08d00), [#1470](https://github.com/morluto/rea/issues/1470), [#1491](https://github.com/morluto/rea/issues/1491)) Thanks @CassieFox23, @kaoru0822-kitauji.
+- Cleanup: retain failed readers, ASAR handles, output trees and provider workers for retry without discarding completed output or partial observations. ([ffe0574](https://github.com/morluto/rea/commit/ffe0574e17b9d973860371083cb1a351be912131), [de0d341](https://github.com/morluto/rea/commit/de0d341524dc68250a45b41eb7e49fcecc1e32a9), [3f8c51e](https://github.com/morluto/rea/commit/3f8c51e9b506d88fc41aded7b1a22b9a09037c26), [d1235fa](https://github.com/morluto/rea/commit/d1235fadc901e1fb9ab3149d53063306c8545414), [026126e](https://github.com/morluto/rea/commit/026126eafe741e5a8b857433587fd615cdd44c19))
+- Providers: canonicalize IDA workspace roots before identity checks, respect IDA cancellation, serialize JEB acquisition and correct Ghidra no-return/imported-symbol recovery. ([#1481](https://github.com/morluto/rea/pull/1481), [#1419](https://github.com/morluto/rea/pull/1419), [5f30ea6](https://github.com/morluto/rea/commit/5f30ea6ff2dd89aab7af6da6e8dfa5e508ed459b), [f88c990](https://github.com/morluto/rea/commit/f88c99021b88ce67e513025901f57478d8107d0c), [49e6f00](https://github.com/morluto/rea/commit/49e6f00ed08b225abf9ddb7c762553607fdab115)) Thanks @AS9929, @JonahSimon, @Y0UW1SH.
+- Binary metadata: respect PE/.NET section extents, bound binary-plist expansion, scope keyed-archive traversal and decode app `Info.plist` files without `plutil`. ([#1678](https://github.com/morluto/rea/pull/1678), [#1515](https://github.com/morluto/rea/pull/1515), [#1378](https://github.com/morluto/rea/pull/1378), [#1476](https://github.com/morluto/rea/pull/1476), [975dff2](https://github.com/morluto/rea/commit/975dff2e25b14b634e14dd5e0db1537d0ca9815b), [76296c5](https://github.com/morluto/rea/commit/76296c5a3868ba3b54be6ed3896719283d6de5f8)) Thanks @jayhemnani9910, @kaoru0822-kitauji, @rudycelekli.
+- Windows runtimes: preserve drive-root paths when creating private provider directories. ([ee46323](https://github.com/morluto/rea/commit/ee46323b493994902b04824a7dc8aae2a8117903))
+- Compatibility: load Windows native controls under Node-API hosts with renamed executables, preserve dylib directory requirements, support Wakaru 1.x and reject directory-valued Hopper launchers. ([#1531](https://github.com/morluto/rea/pull/1531), [#1623](https://github.com/morluto/rea/pull/1623), [#1540](https://github.com/morluto/rea/pull/1540), [#1426](https://github.com/morluto/rea/pull/1426)) Thanks @evenafterallisaid, @Fazmin, @pionxzh, @xeropresence.
+- Reference imports: normalize internal symlink lookup consistently, preserve exact paths and exclusion provenance, report root access failures and reject FIFO replacements without blocking. ([#1535](https://github.com/morluto/rea/pull/1535), [2ca3ad2](https://github.com/morluto/rea/commit/2ca3ad238608adfca80a7ff1aaadea6e59a58225), [6636045](https://github.com/morluto/rea/commit/6636045c9c5b472bba51afee9d434394861e93a4), [3967cfa](https://github.com/morluto/rea/commit/3967cfa43e786c65a57e89c265343983d19ef9d0), [f1c58cc](https://github.com/morluto/rea/commit/f1c58cc53fbd83b968e11b3e2eccebca51299c54)) Thanks @anujsingh-cse.
+- Setup: honor active configuration and skill roots, preserve configured MCP environments, validate Hermes/OMP profiles before writing, recognize Omarchy and retain recovery backups. ([#1504](https://github.com/morluto/rea/pull/1504), [#1607](https://github.com/morluto/rea/pull/1607), [#1528](https://github.com/morluto/rea/pull/1528), [#1562](https://github.com/morluto/rea/pull/1562), [#1662](https://github.com/morluto/rea/pull/1662), [a011485](https://github.com/morluto/rea/commit/a01148553e2a37c72ae78405418ab648876a96a6), [#1606](https://github.com/morluto/rea/issues/1606), [#1521](https://github.com/morluto/rea/issues/1521), [#1561](https://github.com/morluto/rea/issues/1561)) Thanks @Adkid-Zephyr, @CassieFox23, @LuisTaffelli, @N0zoM1z0, @peter-research.
+- Process capture: retain descendant lineage after launcher exit, preserve identity uncertainty before signaling and release terminals after observation failures. ([80896a1](https://github.com/morluto/rea/commit/80896a1d64b216ef893907249cd4252f86215d49), [c916850](https://github.com/morluto/rea/commit/c9168508527302ac57bc86385ccf0d824697b5dc)) Thanks @Y0UW1SH.
+- CLI/MCP errors: accept JSON files for web comparisons, preserve native option help and union correction details, remove invalid aliases and validate doctor against its discovered output contract. ([#1397](https://github.com/morluto/rea/pull/1397), [#1395](https://github.com/morluto/rea/pull/1395), [#1425](https://github.com/morluto/rea/pull/1425), [#1461](https://github.com/morluto/rea/pull/1461), [8b4c96e](https://github.com/morluto/rea/commit/8b4c96e6c6b3ed6806e4cd72b556b5db6114304a), [#1421](https://github.com/morluto/rea/issues/1421), [#1454](https://github.com/morluto/rea/issues/1454)) Thanks @Fazmin, @kaoru0822-kitauji, @N0zoM1z0.
+
+### ⚠ BREAKING CHANGES
+
+- WebSocket results: use `network.websocket_connections` instead of `network.websocket_events`. Connections contain `request_id`, `url` and `events`; allow `payload_bytes: null` when size is unknown. No legacy alias remains. ([db4e360](https://github.com/morluto/rea/commit/db4e3608465486466591b594cf77037ac150c988))
+- Direct resource APIs: pass a caller-owned `ArtifactResourceScope` to inventory/extraction, JavaScript analysis, reference imports and web exports; retain it through cleanup/shutdown. For `readJavaScriptArtifactFiles`, scope is argument three and abort signal is argument four. Keep each root inventory source’s single cleanup owner. Use `OwnedFileHandle` owners, injected ASAR `fd`/`read` handles instead of `createReadStream`, VCS `Result` values, and `WebScriptExportError`’s options object. CLI/MCP inputs for the scoped operations are unchanged. ([ffe0574](https://github.com/morluto/rea/commit/ffe0574e17b9d973860371083cb1a351be912131), [53e7fad](https://github.com/morluto/rea/commit/53e7fadbd078422ba3d55b5499a70d237c5f4908), [684199b](https://github.com/morluto/rea/commit/684199ba0be7822266d57ddb8a40e95f087fdc42), [be75ed5](https://github.com/morluto/rea/commit/be75ed5e4b4bfb146af497f8fa5d172a1ea19462), [be6272f](https://github.com/morluto/rea/commit/be6272f26115bd8e94f2ab0df4cf092004587a18), [e0e49e9](https://github.com/morluto/rea/commit/e0e49e9dd53d9d1f1d2330b23f7f880dc1c81774), [95a77b6](https://github.com/morluto/rea/commit/95a77b6e25420b3879bd0e32bdbb14d452b38776))
+- Saved reference graphs: exclusions require their matching `pattern` and distinguish `project-ignored` from `default-ignored`. Re-import graphs missing that provenance. ([6585293](https://github.com/morluto/rea/commit/6585293c562853b95c3a70df02c12ad8cc17afc4))
+- Terminal captures: keep initial and resized buffers within 1,000,000 cells using `columns × (rows + scrollback)`; oversized scenarios are rejected before launch. ([f9ea9a9](https://github.com/morluto/rea/commit/f9ea9a9268c854d31048da9f8a50cfa37703fe5a))
+- Optional agent-verification reports: accept schema version 4, nullable `expectedFirstTool` and `targetClarificationPassed`. Production CLI/MCP contracts are unchanged. ([#1613](https://github.com/morluto/rea/pull/1613))
+- Capture comparisons: recapture both sides with matching `literal-replacements-v1` semantics when custom normalization patterns are used. Legacy unversioned pattern captures remain readable but cannot be compared with new pattern captures; captures without patterns are unaffected. ([#1456](https://github.com/morluto/rea/pull/1456), [2b6f14f](https://github.com/morluto/rea/commit/2b6f14f528297cf12c776c960af0859f5cfbefbf)) Thanks @Madhavan006, @N0zoM1z0.
+- Artifact graph edges: read child artifact IDs and paths from the occurrence selected by `occurrence_id`, not edge fields `child_artifact_id`/`logical_path`. Earlier inventory Evidence no longer validates against the new commitment. ([#1493](https://github.com/morluto/rea/pull/1493)) Thanks @kaoru0822-kitauji.
+- Semantic graph schema: nodes and relations no longer include `identifier_strategy`, and relation IDs no longer hash it. Earlier Evidence no longer validates against this schema. ([#1474](https://github.com/morluto/rea/pull/1474)) Thanks @kaoru0822-kitauji.
+- Semantic identities: select IDs from the parent graph; read exact event names from `event_name`, endpoints from `endpoint` and literal values from `properties.value`. New analyses use digest-based IDs for long/reserved event and endpoint names and `value-sha256` literal role keys, with shorter labels; string literal labels now read `string literal`. Previously stored graphs retain their original IDs and remain valid workflow inputs. ([#1427](https://github.com/morluto/rea/pull/1427), [#1423](https://github.com/morluto/rea/pull/1423), [#1407](https://github.com/morluto/rea/pull/1407)) Thanks @N0zoM1z0.
+
+### Contributors
+
+- Thanks also to @BaaRai, @commit-regrets, @coygeek, @Franky100-pig, @inii-sayaa, @jopemachine, @JosephDemarest, @keemsisi, @Totopo27 for contributions included in this release.
+
+Full contribution record: [336 commits and 163 merged PRs](https://github.com/morluto/rea/blob/rea-agents-6.4.0/docs/releases/6.4.0.contributions.json).
+
 ## [6.3.0](https://github.com/morluto/rea/compare/rea-agents-6.2.0...rea-agents-6.3.0) (2026-10-09)


@@ -17,8 +17,108 @@ targets; optional adapter failures must preserve successful peers. Run
 `npm run verify:test-discovery` after adding or moving tests. Resolved import
 boundaries are checked by `npm run verify:module-boundaries`, included in lint.

+For analysis workflows and agent interfaces, follow the
+[performance priorities](AGENTS.md#performance). Explain the selected output
+scope, reuse/cache decisions, and relevant before/after measurements when these
+costs change.
+
 When adding or changing an MCP tool, follow the [tool design guide](docs/tool-design.md) and preserve the canonical contracts and generated catalog.

+## Contribution scope
+
+REA focuses on software structure, runtime behavior, and evidence through
+general-purpose reverse-engineering workflows. Evaluate a proposal by its actual
+capabilities, side effects, defaults, and complete workflow. A tool name or a
+reference to vulnerability research, malware analysis, or CTFs does not determine
+whether it belongs in REA.
+
+Supported contributions include format inspection, ordinary archive and standard
+compression extraction, disassembly, decompilation, cross-references, general
+static deobfuscation, vulnerability analysis, and evidence reporting. Static
+protection detection is distinct from defeating that protection. Security fixes
+and isolated regression tests remain in scope.
+
+Runtime investigation needs explicit boundaries. Launching or attaching to a
+process, injecting instrumentation, running scripts, and changing memory or
+registers require review of target selection, execution and mutation authority,
+provenance, resource limits, cancellation, and cleanup. General Frida or debugger
+support is not inherently read-only, and existing support does not authorize
+every workflow built on it. General-purpose interfaces are not classified as
+dedicated bypass workflows solely because caller-provided code could be misused;
+their execution and mutation authority must still be reviewed.
+
+REA does not integrate or officially distribute:
+
+- Purpose-built workflows to defeat commercial packing or code-virtualization
+  protections and automatically recover protected code.
+- Anti-debugging or anti-instrumentation evasion, or DRM, license, anti-cheat, or
+  access-control bypass workflows.
+- Exploit or attack-payload automation, credential theft, covert persistence,
+  unauthorized access, or bulk attack workflows.
+
+These boundaries apply to first-party providers, adapters, plugins, bundled
+scripts, defaults, examples, and documentation. Moving a capability into an
+optional dependency or plugin does not exempt it. Review what the integration
+exposes, including upstream defaults and raw-command or script execution paths.
+Ordinary extraction and general static deobfuscation are not excluded merely
+because the input is packed or obfuscated; assess added bypass or execution
+behavior separately.
+
+This is a product and maintenance boundary, not a judgment about a contributor's
+intentions or a declaration that all excluded research is unlawful. Contributors
+must have the rights and authorization needed for their work. A local-only design
+or an "authorized use only" disclaimer does not establish containment or settle
+legal and scope questions.
+
+### Fail-closed integration
+
+If scope fit, target or authorization boundaries, side effects, or the verification
+needed to establish those boundaries remain materially unresolved, **do not merge,
+enable, or include the capability in official packages or distributions**. Keep
+the proposal pending, narrow it, or decline it. Record unknowns explicitly; do not
+turn them into "none" or assumed safety. Passing CI, completed implementation,
+upstream popularity, or silence from maintainers is not scope approval.
+
+For a new provider or a capability involving target execution, injection,
+mutation, or protection recovery, seek an explicit maintainer scope decision in
+the linked issue or design discussion before substantial implementation intended
+for integration. Small exploratory prototypes and draft PRs may help inform that
+decision; they do not authorize integration. The decision must cover the final
+proposed behavior; a scope expansion needs a new decision. Routine fixes and
+documentation changes within existing boundaries do not need a separate advance
+decision, but must not silently expand those boundaries.
+
+### Starting a proposal
+
+To open a feature request or design discussion, describe the problem and desired
+outcome or decision. Short answers and the forms' "Not sure yet" options are
+welcome. You do not need code, a complete design or test plan, legal analysis, or
+private proof-of-authorization documents to start a discussion. Share what you
+know and leave optional details blank when you need help.
+
+Form selections guide triage; they are not proof of authorization or scope
+approval. Unselected effects must not be treated as established absence.
+Maintainers and implementers work together to resolve the following before
+integration:
+
+- The concrete analyst question, requested inputs and outputs, and why existing
+  REA operations are insufficient.
+- Why the workflow fits this scope, its non-goals, and any protection recovery,
+  bypass, or attack automation it would expose.
+- The selected targets and every execution, injection, mutation, network, or
+  persistence effect, including behavior inherited from upstream tools.
+- How the boundaries and outcome will be verified with self-owned or appropriately
+  licensed samples; identify real CLI/MCP and provider coverage, cleanup and
+  resource checks, and remaining unknowns. Do not publish private binaries,
+  credentials, or sensitive target data; follow [SECURITY.md](SECURITY.md) for
+  private vulnerability reports.
+
+Mixed proposals should separate acceptable analysis capabilities from excluded
+workflows so each can be decided independently. An issue may discuss unresolved
+questions; opening an issue or PR does not authorize integration. This is a
+maintainer integration gate, not a per-call confirmation requirement or a sandbox
+guarantee.
+
 ## Development setup

 Use the development toolchain pinned in `.nvmrc` and
@@ -31,6 +131,9 @@ npm ci
 npm run check:fast
 ```

+`npm-shrinkwrap.json` is the sole dependency lock for development and the
+published plugins. Do not add a root `package-lock.json`; npm would ignore it.
+
 `npm ci` installs the exact dependencies and prepares the Husky hooks without
 building the project. Run `npm run build:cached` when you need the standalone
 CLI or MCP server; it compiles the runtime and bundles the authored skill.
@@ -42,6 +145,21 @@ caches deterministic builds and static checks across Git worktrees. After a pack
 Repository invariants live in [AGENTS.md](AGENTS.md). Keep provider-specific
 protocols in adapters and shared CLI/MCP workflows in the application layer.

+### Windows source checkouts
+
+The published npm package includes the Windows x64 native addon and its
+manifest. A source checkout does not: `native/windows/build/` is ignored by
+Git, and neither `npm ci` nor `npm run build:cached` (or `npm run build`)
+produces `rea-windows-x64.node` or `manifest.json` in that directory.
+Windows workflows and tests that require native controls need these artifacts;
+without them, those controls are unavailable even after a successful runtime build.
+
+Follow the [Windows native build guide](native/windows/README.md#build) before
+running those workflows from source. The artifact build requires Linux,
+MinGW-w64, and `tar`; the build script does not run on Windows. See that guide
+for the build and artifact-verification commands and the separate real Windows
+verification prerequisites. Keep the generated addon and manifest uncommitted.
+
 ## Documentation website

 The VitePress site uses the Markdown files in `docs/`. Run `npm run docs:dev`
@@ -66,6 +184,10 @@ The separate GitHub Pages host uses only the manual

 ## Development feedback and PR verification

+For routine merges, wait for the required checks to pass for the current PR
+revision. Administrators retain an override for deliberate maintainer recovery.
+If `main` is broken, prioritize a focused repair before updating affected PRs.
+
 Select checks for the behavior changed. For a source edit:

 ```bash
@@ -87,6 +209,14 @@ CI diagnosis. Packaging/setup/distribution changes additionally require
 `npm run verify:package` and `npm pack --dry-run`; provider changes require the
 matching real-provider lane. State which host/provider workflows ran.

+For CI/test correctness, drift, or performance investigations, use the
+[`rea-ci-tests-audit` contributor skill](.agents/skills/rea-ci-tests-audit/SKILL.md).
+It guides reproduction and comparable measurements before changing checks.
+
+`check:pr` reuses build and static-check outputs while executing the test suite
+fresh, including the fast Python producer checks. `docs:check` also runs the
+onboarding example consistency suite for README-only changes.
+
 Pre-commit formats and lints staged files; pre-push runs `check:fast`.

 ### Generated files and command ownership
@@ -0,0 +1,136 @@
+# Disclaimer
+
+## Responsible use
+
+REA is built for lawful analysis: security research, malware and incident
+analysis, vulnerability research and coordinated disclosure, CTFs and education,
+interoperability, software preservation, and debugging or auditing software you
+own or are authorized to examine.
+
+The REA project does not support, and will not help with:
+
+- circumventing DRM, license checks, anti-cheat systems or other technical
+  protection measures;
+- developing game cheats or tampering with online services;
+- bypassing authentication, bot protection or access controls on systems you are
+  not authorized to test;
+- copying proprietary software or reproducing its protected code and assets;
+- evading the usage policies or safety measures of AI model providers;
+- reselling or sharing AI accounts or API access.
+
+You are responsible for obtaining any required authorization and for complying
+with applicable laws, license agreements and terms of service. Reverse
+engineering may be restricted in your jurisdiction or by the license of the
+software you analyze.
+
+Requests for help with the uses above will be closed without response in issues
+and discussions, and may result in removal from the REA community.
+
+## User responsibility
+
+Do not use REA for unauthorized access or exploitation, credential theft,
+malicious deployment of malware, unlawful collection or disclosure of data,
+unauthorized service disruption, or unlawful circumvention of access controls or
+licensing restrictions. Legitimate security research and malware analysis must
+remain within the rights and permissions applicable to the work.
+
+You are responsible for your use of REA, including commands, configurations,
+scripts, and actions you direct an AI agent to perform. Establish the legal
+authority and scope required for the targets, accounts, devices, software, data,
+and operations involved. Comply with applicable laws, contracts,
+intellectual-property rights, privacy obligations, and tool licenses. Public
+availability of software, possession of a file, or access to a debugging
+endpoint does not by itself establish permission for every operation.
+
+## Runtime, network, and device operations
+
+Some REA workflows execute or observe programs, attach to runtime or browser
+endpoints, capture network traffic, interact with applications, or operate on
+connected devices. These actions can affect application state, files, accounts,
+devices, and services. Authorization to inspect an artifact does not necessarily
+authorize execution, interception, device changes, or interaction with another
+party's system.
+
+Review the operations you request, including agent-generated plans and scripts,
+and use an appropriate environment for untrusted targets. You are responsible
+for protecting credentials, personal data, captured traffic, and investigation
+artifacts, and for obtaining any permission required to collect, retain,
+process, or share them. Local operation does not establish authorization or
+guarantee that executed targets and third-party tools have no external effects.
+
+## Analysis results and generated output
+
+Decompilation, reconstruction, heuristics, and AI-assisted analysis may be
+incomplete or incorrect. Results do not guarantee accuracy, completeness,
+security, compatibility, or legal compliance. Review the underlying evidence and
+verify consequential conclusions before relying on them.
+
+Obtaining source-like output or an analysis report does not grant rights to the
+original program, its data, or the generated output. Check the target's rights
+and the producing tool's applicable terms before publishing, sharing, or using
+results in another product or service.
+
+## Third-party tools and licenses
+
+Third-party components and analysis tools retain their own licenses, copyright
+notices, and usage conditions. Consult
+[Third-party notices](THIRD_PARTY_NOTICES.md) for their roles, licensing
+records, and preserved upstream texts.
+
+You must legally obtain IDA Pro and Hex-Rays decompilers, Hopper, JEB, and any
+other proprietary tools, together with the rights required for your use. REA
+provides no commercial-tool license, key, or activation entitlement. Permitted
+trials and demos remain subject to vendor terms. An open-source bridge or
+accessible API does not grant additional user, machine, feature, automation,
+hosting, or output-sharing rights. Cracked copies, stolen keys, unauthorized key
+sharing, and bypassing license restrictions are not authorized by REA.
+
+## Disclaimer of warranty
+
+REA is provided **AS IS**, without express or implied warranties, including
+merchantability, fitness for a particular purpose, and noninfringement, as set
+out in the [MIT License](LICENSE). You assume the risks of using the software
+and deciding whether it is suitable for your task. No project description,
+example, analysis result, or third-party integration guarantees a particular
+outcome or establishes that your intended use is lawful or safe.
+
+## Limitation of liability
+
+The warranty and liability provisions of the [MIT License](LICENSE) govern REA's
+authors and copyright holders, subject to applicable law. That license disclaims
+liability for claims, damages, and other liability arising from or in connection
+with the software or its use, whether in contract, tort, or otherwise. This
+document does not expand those provisions or override rights and obligations
+that cannot be excluded under applicable law.
+
+## Disclaimer of endorsement
+
+References to third-party projects, contributors, products, services, and
+trademarks identify their work and REA's integrations. Credits and thanks do not
+imply affiliation, sponsorship, approval, or endorsement by the named parties,
+including the National Security Agency and commercial tool vendors.
+Redistribution, bundling, or hosting by a third party does not by itself imply
+affiliation with or endorsement by the REA project.
+
+## Relationship to software licenses
+
+This document explains REA's intended use, user responsibilities, and community
+support boundaries. It does not amend the MIT license, impose additional
+copyright-license conditions, replace upstream licenses, or grant rights to
+third-party software or targets. The applicable software licenses and agreements
+retain their own terms.
+
+## Versions, distribution channels, and platforms
+
+The project's position on lawful use and non-endorsement of illegal or
+unauthorized activity covers all REA versions, distribution forms, and
+platforms. This includes historical releases and Git commits, npm packages,
+bundled or separately distributed REA skills, source checkouts, release
+archives, and redistributed copies, regardless of installation method, agent
+client, operating system, or platform. Using an older version or another
+distribution channel does not imply project endorsement or remove obligations
+under applicable law or to respect third-party rights.
+
+Each copy remains subject to the licenses accompanying that copy. This statement
+does not retroactively amend existing license grants, add license conditions, or
+expand historical warranty or liability exclusions.
@@ -1,6 +1,7 @@
 <div align="center">

-**English** · [简体中文](README_zh.md) · [繁體中文](README_zh-TW.md) · [日本語](README_ja.md) · [한국어](README_ko.md) · [Türkçe](README_tr.md) · [Русский](README_ru.md) · [Tiếng Việt](README_vi.md) · [ไทย](README_th.md) · [Deutsch](README_de.md) · [Español](README_es.md) · [Français](README_fr.md) · [Українська](README_uk.md) · [Polski](README_pl.md) · [Português (Brasil)](README_pt-BR.md) · [العربية](README_ar.md) · [فارسی](README_fa.md) · [Bahasa Indonesia](README_id.md) · [Italiano](README_it.md)
+**English** · [简体中文](docs/readme/README_zh.md) · [繁體中文](docs/readme/README_zh-TW.md) · [日本語](docs/readme/README_ja.md) · [한국어](docs/readme/README_ko.md) · [Türkçe](docs/readme/README_tr.md) · [Русский](docs/readme/README_ru.md) · [Tiếng Việt](docs/readme/README_vi.md) · [ไทย](docs/readme/README_th.md) · [Deutsch](docs/readme/README_de.md) · [Español](docs/readme/README_es.md) · [Français](docs/readme/README_fr.md) · [Українська](docs/readme/README_uk.md) · [Polski](docs/readme/README_pl.md) · [Português (Brasil)](docs/readme/README_pt-BR.md) · [العربية](docs/readme/README_ar.md) · [فارسی](docs/readme/README_fa.md) · [Bahasa Indonesia](docs/readme/README_id.md) · [Italiano](docs/readme/README_it.md)
+· [Ελληνικά](docs/readme/README_el.md)

 # REA: Reverse Engineer Anything

@@ -9,7 +10,6 @@
 **See a feature you like. Understand how it works, down to the binary level.**

 [![npm version](https://img.shields.io/npm/v/rea-agents?style=flat-square&color=cb3837)](https://www.npmjs.com/package/rea-agents)
-[![CI](https://img.shields.io/github/actions/workflow/status/morluto/rea/ci.yml?branch=main&style=flat-square&label=CI)](https://github.com/morluto/rea/actions/workflows/ci.yml)
 [![MCP tool catalog](https://img.shields.io/badge/MCP-tool_catalog-5c4ee5?style=flat-square)](docs/mcp-contracts.md#generated-catalog)
 [![Node.js requirements](https://img.shields.io/badge/Node.js-requirements-339933?style=flat-square&logo=nodedotjs&logoColor=white)](#what-you-can-analyze)
 [![skills.sh](https://skills.sh/b/morluto/rea?style=flat-square)](https://skills.sh/morluto/rea/reverse-engineer-anything)
@@ -24,6 +24,12 @@

 <code>npx rea-agents setup</code>

+> [!IMPORTANT]
+> Get REA from [morluto/rea](https://github.com/morluto/rea),
+> [rea.tools](https://rea.tools/), or the [rea-agents npm package](https://www.npmjs.com/package/rea-agents).
+> Follow the installation and release links from these official sources. A similar
+> repository name or copied README does not establish that a download is official.
+
 <br />

 <img src="docs/assets/rea-hopper-analysis.png" alt="REA launching its analysis bridge inside Hopper while inspecting a native binary" width="1200" />
@@ -111,12 +117,15 @@ snapshots and scripting.
 REA changes quickly, and new releases include frequent bug fixes. Keep your
 installation up to date.

-For an npm-installed CLI:
+For a CLI installed with `npm install --global` or `bun add --global`:

 ```bash
 rea update
 ```

+Bun global updates require Bun 1.2.17 or later. With an older Bun release,
+upgrade Bun manually before running `rea update`.
+
 To refresh your agent registrations and skill, run the setup command printed
 by the update.

@@ -127,7 +136,8 @@ npx rea-agents@latest setup
 ```

 Review the setup changes and restart your agent. For one-off CLI commands,
-use `npx rea-agents@latest` followed by the command.
+use `npx rea-agents@latest` followed by the command. With Bun, use
+`bunx rea-agents@latest` in place of `npx rea-agents@latest`.

 ## How REA works

@@ -151,17 +161,25 @@ Additional tools and host support depend on the target:
 | ---------------------- | ------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------- |
 | Native binaries        | Pseudocode, assembly, strings, symbols, calls and references                         | Hopper, Ghidra or IDA; [native analysis](https://rea.tools/guides/native/)                                                            |
 | Offline ELF layout     | Sections, segments, original symbols/relocations and static mitigation candidates    | Caller-supplied pwntools on Linux x64; [binary diagnostics](docs/binary-diagnostics.md)                                               |
+| Go binaries            | Embedded compiler version, modules, replacements and build settings                  | Built-in static ELF/PE/thin Mach-O inspection; no Go installation required                                                            |
+| Rust binaries          | Embedded cargo-auditable crate names, versions, sources and dependency edges         | Built-in static ELF/PE/thin Mach-O inspection; [Rust binaries](docs/rust-binaries.md)                                                 |
+| WASM artifact          | Validation, sections, exact import/export WAT forms and decoded WAT                  | Caller-supplied WABT 1.0.42; `inspect_wasm_artifact`                                                                                  |
 | EVM bytecode           | Dispatch selectors, byte offsets, inferred arguments and mutability                  | Local raw/hex carrier on Linux x64; [offline EVM guide](docs/evm-bytecode.md)                                                         |
+| SQLite databases       | Tables, columns, indexes, views, triggers and selected ordinary-table rows           | Local database/WAL snapshot; Node.js 24.x >=24.15 or 26+ with native SQLite limits                                                    |
 | Recorded Linux crashes | Raw notes, every recorded thread's registers/signals and optional mapping candidates | Caller-supplied pwntools on Linux x64; optional GDB/pwndbg; [recorded crashes](docs/recorded-crashes.md)                              |
 | JavaScript / Electron  | Modules, imports, source maps, routes, IPC and native add-on relationships           | Node.js and npm; [application analysis](https://rea.tools/guides/javascript/)                                                         |
 | Websites               | Page structure, scripts, network observations and requested screenshots              | A Chrome-family browser; [browser analysis](https://rea.tools/guides/browser/)                                                        |
-| Saved network captures | Requests, responses, exposed payloads and source locations                           | HAR; mitmdump on Linux for native mitmproxy captures; [capture guide](docs/web-network-captures.md)                                   |
+| Saved network captures | Requests, responses, exposed payloads and source locations                           | HAR or Fiddler Everywhere SAZ; mitmdump on Linux for native mitmproxy; [capture guide](docs/web-network-captures.md)                  |
 | .NET assemblies        | Metadata, CIL instructions, declared native dependencies and build comparisons       | None (static only); [managed-code guide](docs/managed-code-analysis.md)                                                               |
 | Android APKs           | Manifest declarations, classes, decompiled methods and references                    | Headless JADX and a full JDK on Linux/macOS/Windows x64; [Android guide](docs/android-analysis.md)                                    |
 | Android devices        | Devices, packages, processes, logs, dumps, screen captures, and APK/file transfers   | Caller-supplied adb (emulator or hardware); [ADB guide](docs/adb-device-analysis.md)                                                  |
 | JEB projects           | Engine-backed project units, artifact digests and type/method pseudo-code            | A caller-started JEB client serving MCP, selected with REA_JEB_MCP_URL; [JEB guide](docs/jeb-analysis.md)                             |
 | Android resources      | Decoded manifest, string tables, locales and version facts                           | Caller-supplied Apktool launcher; [resource guide](docs/apktool-resource-analysis.md)                                                 |
+| Flutter apps           | Dart snapshot hash, per-ABI library digests and engine build provenance              | None (static APK parsing); [Flutter guide](docs/flutter-analysis.md)                                                                  |
 | Firmware               | Regions, extraction results and native-analysis handoffs                             | Binwalk / Unblob on Linux; [firmware guide](docs/firmware-analysis.md)                                                                |
+| Recorded packets       | Selected fields, repeated values, filters and decode-as rules                        | Linux: TShark 4.x + prlimit; [capture guide](docs/packet-captures.md)                                                                 |
+| Packet summaries       | Endpoints, DNS, TLS SNI, protocol distribution and HTTP requests                     | Caller-supplied tshark and capinfos (Linux-verified); [pcap guide](docs/pcap-analysis.md)                                             |
+| Recorded logic signals | Decoder annotations, sample ranges and diagnostics                                   | Linux: sigrok-cli 0.7.2 + prlimit; [signal recordings](docs/signal-recordings.md)                                                     |
 | Packages and resources | File inventories, digests, plists, Apple bundle anatomy and extracted resources      | [Artifact and JavaScript guide](docs/javascript-artifact-reconstruction.md), [Apple applications](docs/apple-application-analysis.md) |
 | Process behavior       | Terminal output, interactions, exit and filesystem observations, and run comparisons | Linux/macOS with a native PTY; [process capture](docs/process-capture.md)                                                             |

@@ -169,15 +187,20 @@ Static JavaScript and .NET inspection read the supplied files without running
 the application. Runtime capture runs or interacts with the selected target
 using your user permissions; each runtime guide describes its effects.

+For offline native-binary inspection, REA also supports caller-supplied GNU
+objdump and Rizin. See the [reverse-engineering tools guide](docs/reverse-engineering-tools.md)
+for provider setup, CLI commands, and Evidence limitations. REA does not install
+or update either provider.
+
 <a id="choosing-a-deep-analysis-provider"></a>

 Native formats and host support vary by provider. See
 [Hopper and Ghidra setup](docs/installation.md#hopper), the
-[IDA guide](docs/ida-provider.md), and
+[IDA guide](docs/ida-provider.md) (including an optional native MCP bridge), and
 [experimental Windows Ghidra support](docs/windows-ghidra-p0.md).
 Ghidra also supports [16-bit DOS analysis](docs/ghidra-dos.md).
-For large binaries, raise its startup deadline with
-`REA_GHIDRA_STARTUP_TIMEOUT_MS`.
+For large binaries, raise the provider's startup deadline with
+`REA_GHIDRA_STARTUP_TIMEOUT_MS` or `REA_HOPPER_STARTUP_TIMEOUT_MS`.
 For provider selection, see the [CLI guide](docs/cli.md#choose-a-provider).
 Check [release availability](docs/installation.md#released-package-and-main)
 for features added since the latest npm release.
@@ -240,7 +263,9 @@ Ghidra and IDA use your existing installations. See [provider setup](docs/instal
 <summary><strong>Do I need to start Hopper first?</strong></summary>

 REA starts Hopper when an operation needs it. On macOS, a first-run dialog may
-ask you to choose demo mode or activate your license. See
+ask you to choose demo mode or activate your license. Complete that prompt in the
+same user session before your first CLI/MCP operation or unattended job. Open
+Hopper manually if needed. See
 [Hopper startup and troubleshooting](docs/installation.md#launcher-paths-and-troubleshooting).

 </details>
@@ -276,7 +301,7 @@ model provider has its own data policy.

 Update first; a recent release may already fix it.

-For an npm-installed CLI:
+For a CLI installed with `npm install --global` or `bun add --global`:

 ```bash
 rea update
@@ -310,7 +335,7 @@ Report vulnerabilities through [SECURITY.md](SECURITY.md).

 ## Star history

-🎉 **60,000 GitHub stars — thank you!**
+🎉 **180,000 GitHub stars — thank you!**

 Thanks to everyone using REA, reporting bugs, requesting features, testing builds, and contributing fixes.

@@ -324,9 +349,13 @@ Thanks to everyone using REA, reporting bugs, requesting features, testing build

 ## Disclaimer

-REA provides tools for lawful reverse-engineering research, analysis, and reconstruction. You are responsible for obtaining any required authorization and complying with applicable laws. The project does not endorse illegal or unauthorized use.
+See [DISCLAIMER.md](DISCLAIMER.md) for responsible-use guidance.

-REA is an open-source software project. We have not issued or endorsed any cryptocurrency or token. Tokens using the REA name are not affiliated with the project.
+The project's position on illegal or unauthorized use covers all REA versions,
+distribution channels, and platforms; each copy retains its applicable licenses.
+
+REA is for lawful research, education, interoperability, software preservation,
+and analysis of software you own or are authorized to examine.

 ## Contributing

@@ -341,3 +370,16 @@ See [CONTRIBUTING.md](CONTRIBUTING.md) for development setup and checks,
 ## License

 [MIT](LICENSE)
+
+REA-authored code is licensed under MIT. Third-party material retains its own
+licenses and copyright notices; see
+[Third-party notices](THIRD_PARTY_NOTICES.md) for the dependency inventory,
+preserved license texts, and external-tool terms.
+
+**You must legally obtain and be authorized to use commercial tools**, including
+IDA Pro / Hex-Rays decompilers, Hopper, and JEB, under their applicable
+licenses. REA grants no tool license, key, or activation entitlement. Trials and
+demos remain subject to vendor terms, including user, feature, automation, and
+output sharing restrictions. Cracked copies, stolen keys, unauthorized key
+sharing, and bypassing license restrictions are not authorized. This guidance
+does not amend MIT or third-party licenses.
@@ -1,267 +0,0 @@
-<div align="center" dir="rtl">
-
-[English](README.md) · [简体中文](README_zh.md) · [繁體中文](README_zh-TW.md) · [日本語](README_ja.md) · [한국어](README_ko.md) · [Türkçe](README_tr.md) · [Русский](README_ru.md) · [Tiếng Việt](README_vi.md) · [ไทย](README_th.md) · [Deutsch](README_de.md) · [Español](README_es.md) · [Français](README_fr.md) · [Українська](README_uk.md) · [Polski](README_pl.md) · [Português (Brasil)](README_pt-BR.md) · [العربية](README_ar.md) · **فارسی** · [Bahasa Indonesia](README_id.md) · [Italiano](README_it.md)
-
-# REA: مهندسی معکوس هر چیزی
-
-### یک MCP برای مهندسی معکوس فایل‌های باینری، برنامه‌ها و رفتار زمان اجرا.
-
-**قابلیتی دیده‌اید که دوستش دارید؟ سازوکارش را تا سطح باینری درک کنید.**
-
-[![npm version](https://img.shields.io/npm/v/rea-agents?style=flat-square&color=cb3837)](https://www.npmjs.com/package/rea-agents)
-[![CI](https://img.shields.io/github/actions/workflow/status/morluto/rea/ci.yml?branch=main&style=flat-square&label=CI)](https://github.com/morluto/rea/actions/workflows/ci.yml)
-[![MCP tool catalog](https://img.shields.io/badge/MCP-tool_catalog-5c4ee5?style=flat-square)](docs/mcp-contracts.md#generated-catalog)
-[![Node.js requirements](https://img.shields.io/badge/Node.js-requirements-339933?style=flat-square&logo=nodedotjs&logoColor=white)](#current-status)
-[![skills.sh](https://skills.sh/b/morluto/rea?style=flat-square)](https://skills.sh/morluto/rea/reverse-engineer-anything)
-[![MIT license](https://img.shields.io/badge/license-MIT-f4c430?style=flat-square)](LICENSE)
-[![Discord](https://img.shields.io/discord/1556595354999332884?logo=discord&logoColor=white&label=Discord&color=5865F2)](https://discord.gg/GkcryMnJDM)
-
-<a href="https://trendshift.io/repositories/82054?utm_source=repository-badge&amp;utm_medium=badge&amp;utm_campaign=badge-repository-82054" target="_blank" rel="noopener noreferrer"><img src="https://trendshift.io/api/badge/repositories/82054" alt="morluto%2Frea | Trendshift" width="250" height="55" /></a>
-
-**[وب‌سایت](https://rea.tools/) · [راهنماها](https://rea.tools/guides/) · [نمونه‌های عملی](https://rea.tools/showcase/)**
-
-[شروع سریع](#شروع-سریع) · [REA چگونه کار می‌کند؟](#rea-چگونه-کار-میکند) · [چه چیزهایی را میتوان تحلیل کرد؟](#چه-چیزهایی-را-میتوان-تحلیل-کرد) · [نمونه‌های عملی](#نمونههای-عملی) · [پرسش‌های متداول](#پرسشهای-متداول) · [مستندات](#مستندات)
-
-<code>npx rea-agents setup</code>
-
-<br />
-
-<img src="docs/assets/rea-hopper-analysis.png" alt="اجرای پل تحلیل REA در Hopper هنگام بررسی یک فایل باینری بومی" width="1200" />
-
-<br /><br />
-
-<table aria-label="انجمن REA"><tr><td align="center" width="360"><a href="https://discord.gg/GkcryMnJDM"><img    src="docs/assets/discord.svg" height="42" alt="Discord" /><br /><strong>به انجمن مهندسی معکوس بپیوندید</strong></a><br /><sub>دیسکورد · پرسش و پاسخ · نمایش دستاوردها</sub></td></tr></table>
-
-<br />
-
-</div>
-
---
-
-<div dir="rtl">
-
-قابلیتی در یک برنامه دیده‌اید که می‌خواهید در محصول خودتان داشته باشید؟ از ایجنت خود بخواهید با REA آن را بررسی کند. ایجنت می‌تواند بدون دسترسی به کد منبع برنامه، نحوهٔ کار قابلیت را بررسی و توضیح دهد، شواهد را نشان دهد و نسخه‌ای مشابه برای پروژهٔ شما بسازد.
-
-REA ایجنت شما را به ابزارهایی برای بررسی باینری‌های بومی، برنامه‌های JavaScript و Electron، اسمبلی‌های .NET و وب‌سایت‌ها متصل می‌کند. همین ابزارها از ترمینال نیز قابل استفاده‌اند. تحلیل به‌صورت محلی انجام می‌شود و نتایج شامل شواهد و محدودیت‌های پشت هر نتیجه‌گیری هستند.
-
-فرایند راه‌اندازی، REA را در ایجنت شما ثبت می‌کند و دستورالعمل‌های گردش‌کار متناظر را نصب می‌کند. تحلیل بومی می‌تواند از نصب موجود Hopper یا Ghidra استفاده کند؛ همچنین راه‌اندازی می‌تواند با تأیید شما Hopper را به‌صورت اختیاری نصب کند. تحلیل ایستای JavaScript به هیچ‌یک از این موتورهای تحلیل نیاز ندارد.
-
-> برای دستورالعمل‌های راه‌اندازی، راهنماهای تصویری و مطالعات موردی واقعی، **[به وب‌سایت REA مراجعه کنید](https://rea.tools/)**.
-
-## شروع سریع
-
-### ایجنت خود را راه‌اندازی کنید
-
-پس از نصب Node.js و npm، دستور زیر را اجرا کنید:
-
-```bash
-npx rea-agents setup
-```
-
-ایجنت‌های موردنظر را انتخاب کنید، تغییرات پیشنهادی را بررسی و تأیید کنید. فرایند راه‌اندازی، سرور MCP مربوط به REA و دستورالعمل‌های گردش‌کار متناظر را اضافه می‌کند و از پیکربندی موجود نسخهٔ پشتیبان می‌گیرد. پس از آن ایجنت خود را مجدداً راه‌اندازی کنید.
-
-راه‌اندازی از Claude Code، Codex، Cursor، Gemini CLI، Grok Build و [ایجنت‌های دیگر](docs/installation.md#supported-agents) پشتیبانی می‌کند. برای پیکربندی ارائه‌دهنده و ثبت دستی MCP، [نصب و راه‌اندازی](docs/installation.md) را ببینید.
-
-### از ایجنت خود بپرسید
-
-```text
-بررسی کن جست‌وجو در برنامهٔ Notes چگونه کار می‌کند، شواهدش را نشان بده و قابلیتی مشابه برای پروژهٔ من بساز.
-```
-
-به‌جای Notes، نام برنامهٔ هدف و قابلیتی را قرار دهید که می‌خواهید درک کنید.
-
-### استفاده از ترمینال
-
-برای بررسی پوشهٔ استخراج‌شدهٔ یک برنامهٔ JavaScript/Electron یا فایل ASAR:
-
-```bash
-npx -y rea-agents@latest analyze-javascript-application /absolute/path/to/app --json
-```
-
-نتیجه شامل ماژول‌ها، importها، مرزهای Electron و شواهد مربوط به آن‌هاست. مسیر را با مسیر برنامهٔ هدف جایگزین کنید؛ مثلاً در ویندوز `"D:/apps/example"`.
-
-برای نصب دستور `rea` جهت استفادهٔ مداوم:
-
-```bash
-npm install --global rea-agents
-rea --help
-```
-
-برای تحلیل بومی ابتدا یک ارائه‌دهنده پیکربندی کنید. برای دستورات بومی، انتخاب ارائه‌دهنده، snapshotها و اسکریپت‌نویسی، [راهنمای CLI و شواهد](docs/cli.md) را ببینید.
-
-### به‌روزرسانی REA
-
-REA به‌سرعت تغییر می‌کند و نسخه‌های جدید معمولاً شامل رفع اشکال‌های متعددی هستند. نصب خود را به‌روز نگه دارید.
-
-برای CLI نصب‌شده از طریق npm:
-
-```bash
-rea update
-```
-
-برای به‌روزرسانی ثبت ایجنت‌ها و skill، دستور setup نمایش‌داده‌شده توسط به‌روزرسانی را اجرا کنید.
-
-اگر از `npx` استفاده می‌کنید، راه‌اندازی ایجنت را با دستور زیر به‌روز کنید:
-
-```bash
-npx rea-agents@latest setup
-```
-
-تغییرات راه‌اندازی را بررسی کنید و ایجنت را دوباره اجرا کنید. برای اجرای یک‌بارهٔ دستورات CLI، از `npx rea-agents@latest` و سپس دستور موردنظر استفاده کنید.
-
-## REA چگونه کار می‌کند؟
-
-ایجنت شما از طریق MCP، REA را فراخوانی می‌کند تا هدف را بررسی و کد مرتبط را ردیابی کند. REA یافته‌ها را همراه با شواهد برمی‌گرداند. ایجنت از این نتایج برای طرح پرسش‌های تکمیلی، توضیح رفتار یا نوشتن و آزمایش یک پیاده‌سازی استفاده می‌کند. دستورات CLI نیز از همین گردش‌کارها بهره می‌برند.
-
-![جریان بررسی REA: ایجنت دربارهٔ هدف محلی سؤال می‌پرسد، REA آن را با ابزارهای تحلیل بررسی و ردیابی می‌کند و ایجنت از کد، ارجاعات و ناشناخته‌های بازگردانده‌شده برای توضیح، پیاده‌سازی و آزمایش استفاده می‌کند.](website/public/assets/figures/rea-investigation-flow.svg)
-
-[نمایش تصویر در اندازهٔ کامل](website/public/assets/figures/rea-investigation-flow.svg).
-
-<a id="current-status"></a>
-
-## چه چیزهایی را می‌توان تحلیل کرد؟
-
-REA به Node.js 22.x (>=22.19)، 24.x (>=24.11) یا 26+، به‌همراه npm نیاز دارد. ابزارهای اضافی و پشتیبانی سیستم میزبان به نوع هدف بستگی دارند:
-
-| هدف                   | خروجی REA                                                                           | پیش‌نیازها و راهنما                                                                                                                  |
-| --------------------- | ----------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
-| باینری‌های بومی        | شبه‌کد، اسمبلی، رشته‌ها، نمادها، فراخوانی‌ها و ارجاعات                                 | Hopper، Ghidra یا IDA؛ [تحلیل بومی](https://rea.tools/guides/native/)                                                               |
-| ساختار ELF آفلاین      | بخش‌ها، سگمنت‌ها، نمادها و relocationهای اصلی و گزینه‌های احتمالی کاهش آسیب‌پذیری ایستا | pwntools تأمین‌شده توسط فراخواننده روی Linux x64؛ [تشخیص باینری](docs/binary-diagnostics.md)                                         |
-| بایت‌کد EVM            | selectorهای dispatch، offsetهای بایتی، آرگومان‌های استنباط‌شده و قابلیت تغییر وضعیت   | دادهٔ خام/hex محلی روی Linux x64؛ [راهنمای EVM آفلاین](docs/evm-bytecode.md)                                                          |
-| کرش‌های ثبت‌شدهٔ لینوکس  | یادداشت‌های خام، رجیسترها/سیگنال‌های تمام threadهای ثبت‌شده و گزینه‌های احتمالی نگاشت   | pwntools تأمین‌شده توسط فراخواننده روی Linux x64؛ GDB/pwndbg اختیاری؛ [کرش‌های ثبت‌شده](docs/recorded-crashes.md)                      |
-| JavaScript / Electron | ماژول‌ها، importها، source mapها، مسیرها، IPC و ارتباط افزونه‌های بومی                | Node.js و npm؛ [تحلیل برنامه](https://rea.tools/guides/javascript/)                                                                 |
-| وب‌سایت‌ها              | ساختار صفحه، اسکریپت‌ها، مشاهدات شبکه و اسکرین‌شات‌های درخواستی                        | مرورگر مبتنی بر Chrome؛ [تحلیل مرورگر](https://rea.tools/guides/browser/)                                                           |
-| داده‌های ضبط‌شدهٔ شبکه   | درخواست‌ها، پاسخ‌ها، payloadهای آشکارشده و مکان آن‌ها در منبع                          | HAR؛ ابزار mitmdump در لینوکس برای ضبط‌های بومی mitmproxy؛ [راهنمای ضبط شبکه](docs/web-network-captures.md)                          |
-| اسمبلی‌های .NET        | فراداده، دستورالعمل‌های CIL، وابستگی‌های بومی اعلام‌شده و مقایسهٔ buildها                | بررسی ایستا؛ [راهنمای کد مدیریت‌شده](docs/managed-code-analysis.md)                                                                  |
-| فایل‌های APK اندروید   | تعاریف Manifest، کلاس‌ها، متدهای دیکامپایل‌شده و ارجاعات                               | JADX بدون رابط گرافیکی و JDK کامل در Linux/macOS/Windows x64؛ [راهنمای اندروید](docs/android-analysis.md)                           |
-| دستگاه‌های اندروید     | دستگاه‌ها، بسته‌ها، فرایندها، لاگ‌ها، دامپ‌ها، اسکرین‌شات‌ها و انتقال APK/فایل             | adb تأمین‌شده توسط فراخوان‌کننده (شبیه‌ساز یا دستگاه)؛ [راهنمای ADB](docs/adb-device-analysis.md)                                      |
-| پروژه‌های JEB          | واحدهای پروژه از موتور، هش‌های مصنوع و شبه‌کدِ نوع/متد                                 | یک کلاینت JEB که توسط فراخوان‌کننده اجرا شده و MCP ارائه می‌دهد و با REA_JEB_MCP_URL انتخاب می‌شود؛ [راهنمای JEB](docs/jeb-analysis.md) |
-| منابع اندروید         | مانیفست رمزگشایی‌شده، جداول رشته‌ها، زبان‌ها و اطلاعات نسخه                             | راه‌انداز Apktool تأمین‌شده توسط فراخوان‌کننده؛ [راهنمای منابع](docs/apktool-resource-analysis.md)                                     |
-| Firmware              | ناحیه‌ها، نتایج استخراج و تحویل به ابزارهای تحلیل بومی                               | Binwalk / Unblob در لینوکس؛ [راهنمای Firmware](docs/firmware-analysis.md)                                                           |
-| بسته‌ها و منابع        | فهرست فایل‌ها، digestها، فایل‌های plist، ساختار bundle اپل و منابع استخراج‌شده         | [راهنمای آرتیفکت و JavaScript](docs/javascript-artifact-reconstruction.md)، [برنامه‌های اپل](docs/apple-application-analysis.md)     |
-| رفتار فرایند          | خروجی ترمینال، تعاملات، وضعیت خروج، مشاهدات سیستم فایل و مقایسهٔ اجراها               | Linux/macOS با PTY بومی؛ [ضبط فرایند](docs/process-capture.md)                                                                      |
-
-بررسی ایستای JavaScript و .NET فایل‌های ارائه‌شده را بدون اجرای برنامه می‌خواند. ضبط زمان اجرا، هدف انتخاب‌شده را با مجوزهای کاربری شما اجرا می‌کند یا با آن تعامل دارد؛ آثار این عملیات در راهنمای مربوط به هر نوع ضبط توضیح داده شده‌اند.
-
-<a id="choosing-a-deep-analysis-provider"></a>
-
-فرمت‌های بومی و پشتیبانی سیستم میزبان بر اساس ارائه‌دهنده متفاوت‌اند. [راه‌اندازی Hopper و Ghidra](docs/installation.md#hopper)، [راهنمای IDA](docs/ida-provider.md) و [پشتیبانی آزمایشی Ghidra در ویندوز](docs/windows-ghidra-p0.md) را ببینید. Ghidra همچنین از [تحلیل DOS شانزده‌بیتی](docs/ghidra-dos.md) پشتیبانی می‌کند. برای انتخاب ارائه‌دهنده، به [راهنمای CLI](docs/cli.md#choose-a-provider) مراجعه کنید. برای قابلیت‌هایی که پس از آخرین انتشار npm اضافه شده‌اند، [وضعیت انتشار](docs/installation.md#released-package-and-main) را بررسی کنید.
-
-## نمونه‌های عملی
-
-[![تصویرهایی از نمونه‌های پنینگ صدای DX-Ball، پل کلیپ‌بورد Notion و حلقهٔ گلوله‌های TH04](docs/assets/rea-showcases.png)](https://rea.tools/showcase/)
-
-### DX-Ball: بازسازی محاسبهٔ موقعیت استریوی صدا
-
-یک فراخوانی صوتی را تا تابع کمکی تبدیل موقعیت به pan دنبال کنید، دستورالعمل‌ها را بررسی کنید و شبه‌کد ناقص را به C تبدیل کنید. بازسازی حاصل، هر ۳٬۲۰۵ مورد آزمایش x86 اصلی را با موفقیت پشت سر می‌گذارد و هر ۶۳ بایت تابع کامپایل‌شده را دقیقاً بازتولید می‌کند.
-
-[مطالعهٔ موردی](https://rea.tools/showcase/dx-ball/) · [مخزن بازسازی](https://github.com/N0zoM1z0/dx-ball)
-
-### Notion: ردیابی پل کلیپ‌بورد Electron
-
-API کلیپ‌بورد در renderer را پیدا کنید، مسیر آن را از preload و IPC تا فرایند اصلی دنبال کنید و قالب غنی کلیپ‌بورد را بررسی کنید.
-
-[مطالعهٔ موردی](https://rea.tools/showcase/notion/)
-
-### TH04: بازیابی محاسبهٔ حلقهٔ گلوله در DOS
-
-دستورالعمل‌های ۱۶بیتی بازی اصلی PC-98 را بررسی کنید، محاسبات زاویهٔ ثابت و هدف‌گیری‌شده را بازیابی کنید و کد C++ بازسازی‌شده را با خروجی کامپایلر تاریخی مقایسه کنید.
-
-[مطالعهٔ موردی](https://rea.tools/showcase/th04/) · [مخزن بازسازی](https://github.com/N0zoM1z0/th04)
-
-اگر از REA برای بررسی موضوع جالبی استفاده کرده‌اید، خوشحال می‌شویم آن را به اشتراک بگذارید. یک [issue](https://github.com/morluto/rea/issues) یا [pull request](https://github.com/morluto/rea/pulls) ثبت کنید و هدف، پرسش خود، نقش REA و یافته‌هایتان را توضیح دهید.
-
-## پرسش‌های متداول
-
-<details><summary><strong>کدام ایجنت‌ها می‌توانند از REA استفاده کنند؟</strong></summary>
-
-هر ایجنتی که از سرورهای MCP محلی پشتیبانی کند. فرایند setup، [ایجنت‌های پشتیبانی‌شده](docs/installation.md#supported-agents) را پیکربندی می‌کند؛ سایر کلاینت‌ها می‌توانند از [ثبت دستی MCP](docs/installation.md#mcp-registry) استفاده کنند.
-
-</details>
-
-<details><summary><strong>آیا به Hopper، Ghidra یا IDA نیاز دارم؟</strong></summary>
-
-تحلیل عمیق بومی به یکی از آن‌ها نیاز دارد. بررسی ایستای JavaScript و .NET بدون موتور تحلیل بومی کار می‌کند. setup می‌تواند با تأیید شما Hopper را نصب کند؛ Ghidra و IDA از نصب‌های موجود شما استفاده می‌کنند. [راه‌اندازی ارائه‌دهنده](docs/installation.md#hopper) را ببینید.
-
-</details>
-
-<details><summary><strong>آیا باید ابتدا Hopper را اجرا کنم؟</strong></summary>
-
-REA زمانی که عملیاتی به Hopper نیاز داشته باشد آن را اجرا می‌کند. در macOS، ممکن است هنگام نخستین اجرا پنجره‌ای برای انتخاب حالت آزمایشی یا فعال‌سازی مجوز نمایش داده شود. [راه‌اندازی Hopper و رفع اشکال](docs/installation.md#launcher-paths-and-troubleshooting) را ببینید.
-
-</details>
-
-<details><summary><strong>نصب skill از skills.sh چه کاری انجام می‌دهد؟</strong></summary>
-
-این skill دستورالعمل‌های بررسی را برای ایجنت فراهم می‌کند. برای ثبت سرور MCP مربوط به REA و نصب دستورالعمل‌های متناظر، `npx rea-agents setup` را اجرا کنید و سپس ایجنت را مجدداً راه‌اندازی کنید. [نصب مستقل skill](docs/installation.md#skill-only-installation) را ببینید.
-
-</details>
-
-<details><summary><strong>REA چه نوع کدی برمی‌گرداند؟</strong></summary>
-
-تحلیل بومی، شبه‌کد و اسمبلی برمی‌گرداند. تحلیل JavaScript/Electron ماژول‌ها و روابط میان آن‌ها را بازیابی می‌کند. ایجنت شما از این یافته‌ها برای نوشتن و آزمایش یک پیاده‌سازی استفاده می‌کند؛ [نمونه‌های عملی](#نمونههای-عملی) مثال‌های گام‌به‌گام ارائه می‌دهند.
-
-</details>
-
-<details><summary><strong>آیا REA برنامهٔ من را آپلود می‌کند؟</strong></summary>
-
-REA اهداف را به‌صورت محلی تحلیل می‌کند. ایجنت شما نتایج ابزارها را دریافت می‌کند و ارائه‌دهندهٔ مدل آن سیاست مستقل خود را دربارهٔ داده‌ها دارد.
-
-</details>
-
-<details><summary><strong>اگر با خطا مواجه شدم چه کنم؟</strong></summary>
-
-ابتدا به‌روزرسانی کنید؛ ممکن است مشکل در نسخه‌ای جدید رفع شده باشد.
-
-برای CLI نصب‌شده با npm:
-
-```bash
-rea update
-```
-
-برای راه‌اندازی ایجنت با `npx`:
-
-```bash
-npx rea-agents@latest setup
-```
-
-اگر از ایجنت استفاده می‌کنید، [به‌روزرسانی راه‌اندازی](#بهروزرسانی-rea) را تکمیل و ایجنت را مجدداً اجرا کنید. همان کار را دوباره امتحان کنید. اگر مشکل ادامه داشت، یک [issue ثبت کنید](https://github.com/morluto/rea/issues) و نسخهٔ REA، نوع هدف، مراحل بازتولید و خروجی خطا را ذکر کنید.
-
-</details>
-
-## مستندات
-
-از [راهنماهای عملی وب‌سایت](https://rea.tools/guides/) شروع کنید. برای گزینه‌های دقیق، پیش‌نیازها و قراردادهای خروجی:
-
- [نصب و راه‌اندازی](docs/installation.md): ثبت ایجنت، پیکربندی ارائه‌دهنده، به‌روزرسانی و حذف نصب.
- [آمادگی و رفع اشکال](docs/installation.md#check-readiness-for-your-task): عیب‌یابی یک ایجنت یا موتور تحلیل.
- [CLI و شواهد](docs/cli.md): دستورات، انتخاب ارائه‌دهنده، snapshotها، ورود/خروج داده و وضعیت‌های خروج.
- [قراردادهای MCP](docs/mcp-contracts.md) و [پرامپت‌های ایجنت](docs/mcp-prompts.md): نتایج ابزارها، نشست‌ها و بررسی‌های هدایت‌شده.
- [فهرست ابزارها](docs/mcp-contracts.md#generated-catalog): فهرست تولیدشده هنگام build از ابزارها، ارائه‌دهندگان و دستورات CLI.
- [نقشهٔ راه](docs/roadmap.md): برنامه‌های آینده و پیگیری قابلیت‌ها.
-
-آسیب‌پذیری‌ها را از طریق [SECURITY.md](SECURITY.md) گزارش کنید.
-
-## تاریخچهٔ ستاره‌ها
-
-🎉 **۵۰٬۰۰۰ ستارهٔ GitHub — سپاسگزاریم!**
-
-از همهٔ کسانی که از REA استفاده می‌کنند، خطاها را گزارش می‌دهند، قابلیت‌های جدید پیشنهاد می‌کنند، buildها را آزمایش می‌کنند و در بهبود پروژه مشارکت دارند سپاسگزاریم.
-
-<a href="https://www.star-history.com/?repos=morluto%2Frea&amp;type=date"><picture><source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/chart?repos=morluto/rea&amp;type=date&amp;theme=dark&amp;legend=top-left" /><source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/chart?repos=morluto/rea&amp;type=date" /><img alt="تاریخچهٔ ستاره‌های GitHub پروژهٔ REA" src="https://api.star-history.com/chart?repos=morluto/rea&amp;type=date" /></picture></a>
-
-## سلب مسئولیت
-
-REA ابزارهایی برای پژوهش، تحلیل و بازسازی قانونی در حوزهٔ مهندسی معکوس فراهم می‌کند. مسئولیت دریافت مجوزهای لازم و رعایت قوانین مربوط بر عهدهٔ شماست. این پروژه استفادهٔ غیرقانونی یا بدون مجوز را تأیید نمی‌کند.
-
-## مشارکت
-
-از کمک شما به REA استقبال می‌کنیم! برای گزارش خطا یا پیشنهاد قابلیت، یک [issue باز کنید](https://github.com/morluto/rea/issues) یا برای بهبود کد و مستندات [pull request بفرستید](https://github.com/morluto/rea/pulls).
-
-برای راه‌اندازی محیط توسعه و بررسی‌ها، [CONTRIBUTING.md](CONTRIBUTING.md)، برای مراحل اعتبارسنجی [آزمایش‌ها](docs/testing.md) و برای ساختار پروژه [نقشهٔ معماری](docs/architecture.mermaid) را ببینید.
-
-## مجوز
-
-[MIT](LICENSE)
-
-</div>
@@ -24,9 +24,12 @@ provider-dependent behavior to be reproduced safely.

 ## Security boundary

-REA authenticates each local provider bridge session with a random capability
-token and a current-user Unix socket. Tokens are transferred through private
-session descriptors rather than process arguments or environment variables.
+REA authenticates its Hopper and Ghidra bridges with per-session random
+capability tokens. POSIX uses current-user Unix sockets; Windows Ghidra uses
+authenticated loopback TCP with a private NTFS runtime whose DACL admits only
+the current user and SYSTEM. See [Windows Ghidra P0](docs/windows-ghidra-p0.md#private-runtime-transport-and-lifecycle).
+Tokens are transferred through private session descriptors rather than process
+arguments or environment variables.
 This is not a sandbox and does not protect against malicious processes already
 running as the same operating-system user. Opening an untrusted binary
 delegates parsing and analysis to the selected local provider with that user's
@@ -0,0 +1,134 @@
+# Third-party notices
+
+REA-authored code is licensed under [MIT](LICENSE). Third-party material retains
+its own copyright, license, attribution, and NOTICE requirements. REA's MIT
+license grants no rights to proprietary analysis tools. Project and product
+names identify integrations; they do not imply affiliation or endorsement.
+
+This inventory covers the locked npm dependency graph, code or assets derived
+from upstream projects, and separately obtained engines used by REA. A listed
+integration is not a statement that its executable is bundled. Upstream
+components and caller-selected versions may have additional terms; retain the
+notices supplied with the software you actually obtain or redistribute.
+
+## Distributed dependencies, derived code, and assets
+
+| Material                 | Use and licensing records                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
+| ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| npm dependencies         | [Complete inventory](licenses/npm-dependencies.txt) of every entry in `npm-shrinkwrap.json`, including transitive, optional platform, development, and website dependencies. [Runtime notices](licenses/npm-runtime-notices.txt) preserve the copyright, license, and NOTICE documents from integrity-verified release archives, plus the MIT notice for Rolldown's generated runtime helpers. Development dependencies are inventoried but are not all distributed in REA's npm package.                                                                                                                                                                                                                     |
+| NativeAOT parser         | REA's derived parser uses the MIT-licensed [ghidra-nativeaot](https://github.com/Washi1337/ghidra-nativeaot/tree/effeb734fc570c32650f88b159608979dc7b423e) work of Washi and the NativeAOT layout definitions in [dotnet/runtime](https://github.com/dotnet/runtime/tree/a3b2d40328be0be3f8dd43f0e5cc925b8827b044). Preserve [Washi's MIT notice](licenses/nativeaot-parser-MIT.txt) and the [.NET Foundation MIT notice](licenses/dotnet-runtime-MIT.txt).                                                                                                                                                                                                                                                   |
+| EVM analysis             | `evmole@0.9.3`, MIT, copyright 2023 Maxim Andreev. The published npm archive omits its license document; REA retains the [original upstream license](third_party/evmole/LICENSE) and [provenance](third_party/evmole/README.md). It is also included in the runtime notices.                                                                                                                                                                                                                                                                                                                                                                                                                                  |
+| Windows native addon     | The release addon is compiled with MinGW-w64 and statically links GCC runtime libraries. Preserve the [MinGW-w64 runtime notices](licenses/MinGW-w64-runtime.txt), [GPLv3 text](licenses/GPL-3.0.txt), and [GCC Runtime Library Exception 3.1](licenses/GCC-Runtime-Library-Exception-3.1.txt). The exception governs eligible compiled combinations; it does not relicense the runtime itself under MIT. The Node-API build headers are from Node 22.19.0; its [original license and component notices](licenses/Node-22.19.0-LICENSE.txt) are retained. Node itself is supplied by the user. Snapshot provenance and release-toolchain review requirements are in [licenses/README.md](licenses/README.md). |
+| Dinosaur website example | Selected excerpts from [wayou/t-rex-runner](https://github.com/wayou/t-rex-runner/tree/5455bfa408ec6b707c7300ff194b7390733a766d), derived from Chromium's dinosaur game. The website retains the [original BSD-3-Clause notices](website/public/assets/licenses/dinosaur.txt), including Chromium authors and wayou. This asset is not part of the npm runtime.                                                                                                                                                                                                                                                                                                                                               |
+
+Two locked runtime releases, [`clean-git-ref@2.0.1`](https://github.com/TheSavior/clean-git-ref/tree/bf3c8afc8032bcdc6a53634c18b601dc6df478fb) (declared Apache-2.0) and
+[`minimisted@2.0.1`](https://github.com/kt3k/minimisted/tree/95fdfb8d6efb9546c1117854fe3a58105f6332ce) (declared MIT), omit a full license/copyright document in their
+published archives and inspected upstream release sources. Their runtime notice
+entries preserve the declared license and author metadata and include a clearly
+labeled standard license text. These are documented upstream notice gaps, not
+recovered copyright notices or a claim that attribution has been fully verified.
+
+The optional `frida@17.23.2` npm archive and tagged Node binding source also
+omit a standalone license/copyright document. Its declared
+`LGPL-2.0 WITH WxWindows-exception-3.1` license is recorded with unchanged
+[Frida licensing text](licenses/frida-17.23.2-COPYING.txt) and the
+[GNU Library GPL v2](licenses/LGPL-2.0.txt). These are licensing supplements,
+not recovered Node binding or prebuild-component copyright notices.
+
+## Open-source engine credits and licenses
+
+We thank the maintainers and contributors of the projects below. Their
+decompilers, binary-analysis engines, debuggers, parsers, and capture tools
+provide capabilities used through REA's integrations. Each row identifies the
+upstream project, its own license, and its role in REA. In particular, JADX
+provides the Android decompiler, while jadx-headless-mcp provides the separate
+headless integration used by REA; both projects deserve credit.
+
+These acknowledgments are voluntary thanks. Required upstream copyright and
+NOTICE texts remain in their original form and are not replaced by this
+summary. Credits do not imply endorsement or change any license.
+
+REA's adapters do not grant, replace, or waive these licenses. The upstream
+executables and their dependency graphs are supplied separately. Git submodules
+pin inspection sources; they are not included in the npm package.
+
+| Project                        | License and source                                                                                                                                                                                                                                       | REA relationship                                                                                                                                                                                                            |
+| ------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| Ghidra                         | [Apache-2.0](https://github.com/NationalSecurityAgency/ghidra/blob/Ghidra_12.1.4_build/LICENSE), with the government-work and component qualifications in its [NOTICE](https://github.com/NationalSecurityAgency/ghidra/blob/Ghidra_12.1.4_build/NOTICE) | Bring your own Ghidra 12.1.4. Copies of its [LICENSE](licenses/ghidra-12.1.4-LICENSE.txt), [NOTICE](licenses/ghidra-12.1.4-NOTICE.txt), and [DISCLAIMER](licenses/ghidra-12.1.4-DISCLAIMER.txt) are retained for reference. |
+| ghidra-nativeaot               | [MIT](https://github.com/Washi1337/ghidra-nativeaot/blob/effeb734fc570c32650f88b159608979dc7b423e/LICENSE.md)                                                                                                                                            | Pinned inspection source; derived parser attribution above.                                                                                                                                                                 |
+| JADX                           | [Apache-2.0](https://github.com/skylot/jadx/blob/v1.5.6/LICENSE)                                                                                                                                                                                         | The audited jadx-headless-mcp source uses JADX 1.5.6 for Android decompilation; the engine JAR is separately supplied.                                                                                                      |
+| jadx-headless-mcp              | [Apache-2.0](https://github.com/1013503897/jadx-headless-mcp/blob/5844800a486d2248fa949b74c9896285f5b54de2/LICENSE)                                                                                                                                      | v0.7.1; pinned submodule and user-supplied JAR.                                                                                                                                                                             |
+| Apktool                        | [Apache-2.0](https://github.com/iBotPeaches/Apktool/blob/master/LICENSE.md)                                                                                                                                                                              | User-supplied Android resource decoder.                                                                                                                                                                                     |
+| Binwalk                        | [MIT](https://github.com/ReFirmLabs/binwalk/blob/4fdab3d464d97b68e0af9088df3f9e2e1545b21c/LICENSE)                                                                                                                                                       | v3.1.0; pinned submodule and user-supplied firmware engine.                                                                                                                                                                 |
+| Unblob                         | [MIT](https://github.com/onekey-sec/unblob/blob/1fcc7a0a584a70a96c31f5a276c20944d199a089/LICENSE)                                                                                                                                                        | 26.6.4; pinned submodule and user-supplied firmware engine.                                                                                                                                                                 |
+| Wakaru                         | [Apache-2.0](https://github.com/pionxzh/wakaru/blob/8219bf5016a063d7fb215101a8c45d5959ec733f/LICENSE)                                                                                                                                                    | v1.14.0; pinned source and separately obtained JavaScript reconstruction tooling.                                                                                                                                           |
+| WABT                           | [Apache-2.0](https://github.com/WebAssembly/wabt/blob/ff0ef7e0009402740c805a9744c09b05be063e48/LICENSE)                                                                                                                                                  | v1.0.42; external WebAssembly tools; [integration provenance](third_party/wabt/README.md).                                                                                                                                  |
+| pwntools                       | [MIT, with GPL/BSD component qualifications](https://github.com/Gallopsled/pwntools/blob/3ad86ec5adab1abf9781ca7552a6f5d7cff7a33c/LICENSE-pwntools.txt)                                                                                                  | 4.15.0; caller-supplied offline ELF analysis package.                                                                                                                                                                       |
+| pyelftools                     | [Public-domain dedication and component notices](https://github.com/eliben/pyelftools/blob/v0.33/LICENSE)                                                                                                                                                | 0.33; includes separately licensed bundled material, including Construct.                                                                                                                                                   |
+| Unicorn                        | [GPL-2.0 and component notices](https://github.com/unicorn-engine/unicorn/blob/2.1.2/COPYING)                                                                                                                                                            | 2.1.2; used by the caller-supplied ELF analysis profile for instruction emulation.                                                                                                                                          |
+| pwndbg                         | [MIT](third_party/pwndbg/LICENSE.md)                                                                                                                                                                                                                     | 2026.09.15, commit `55596aa676a2756efb8d67034f31675170eabe48`; separately obtained debugger extension.                                                                                                                      |
+| GNU GDB                        | [GPL-3.0-or-later](https://sourceware.org/git/?p=binutils-gdb.git;a=blob;f=gdb/COPYING)                                                                                                                                                                  | Caller-supplied debugger for the optional pwndbg core-file workflow.                                                                                                                                                        |
+| LLDB / LLVM                    | [Apache-2.0 with LLVM exceptions](https://github.com/llvm/llvm-project/blob/main/LICENSE.TXT)                                                                                                                                                            | Caller-supplied debugger; the installed distribution's component and historical notices also apply.                                                                                                                         |
+| mitmproxy                      | [MIT](https://github.com/mitmproxy/mitmproxy/blob/6c09d56e4c29a92f5ad01b03199977584b8ea14f/LICENSE)                                                                                                                                                      | 12.2.3; separately supplied network capture proxy.                                                                                                                                                                          |
+| Android Debug Bridge (adb)     | [Apache-2.0](https://android.googlesource.com/platform/packages/modules/adb/+/refs/heads/main/NOTICE)                                                                                                                                                    | User-supplied Android device tool. A complete Android SDK download has its own [SDK license terms](https://developer.android.com/studio/terms); adb's source license does not replace them.                                 |
+| ILSpy / ICSharpCode.Decompiler | [MIT](https://github.com/icsharpcode/ILSpy/blob/master/ICSharpCode.Decompiler/Properties/AssemblyInfo.cs)                                                                                                                                                | Optional separately installed .NET decompiler; not a redistributed REA dependency.                                                                                                                                          |
+
+Ghidra's original notice credits software developed at the National Security
+Agency. It explains that US-government-created portions are not subject to US
+copyright protection, while other portions retain their authors' rights. Its
+modules include additional third-party licenses, and standalone programs under
+`GPL/` have GPLv3 terms. The three retained top-level documents do not replace
+the module notices and license directories in a complete Ghidra distribution.
+REA is not affiliated with or endorsed by the NSA.
+
+## Commercial tools and bridges
+
+**You must legally obtain the tools and the rights needed for your use.** REA
+does not include an IDA Pro, Hex-Rays decompiler, Hopper, or JEB license, license
+key, or activation entitlement. A permitted trial/demo may be used only within
+its official terms. Cracked copies, stolen keys, unauthorized sharing, and
+bypassing activation or license limits are not authorized by REA.
+
+| Project / product                | Applicable terms and distinction                                                                                                                                                                                                                                                                                                                                                                                                                           |
+| -------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| IDA Pro and Hex-Rays decompilers | Proprietary Hex-Rays software governed by the [EULA](https://hex-rays.com/eula), purchase agreement, and applicable license scheme. Obtain rights for the required product, user, machine, decompiler, and automation interfaces. Availability of an API or a bridge does not grant hosting, service, or additional-user rights.                                                                                                                           |
+| ida-pro-mcp                      | The bridge's [MIT license](https://github.com/mrexodia/ida-pro-mcp/blob/main/LICENSE) belongs to Duncan Ogilvie. It licenses the bridge code, not IDA or Hex-Rays products. Its [README](https://github.com/mrexodia/ida-pro-mcp#prerequisites) requires IDA Pro and explicitly states that IDA Free is unsupported. Follow the prerequisites for the version you actually install. The bridge is separately supplied, not bundled by REA.                 |
+| Hopper Disassembler              | Proprietary Cryptic Apps software under its [license agreement](https://www.hopperapp.com/license_agreement.html). Personal and computer licenses have different permitted-user and installation rules; an official evaluation trial is available. REA setup may obtain the official installer after approval, but this grants no paid license or activation entitlement.                                                                                  |
+| JEB                              | Proprietary PNF Software product under its [EULA](https://www.pnfsoftware.com/jeb/LICENSE.TXT) and edition-specific agreement. Check user/seat/server rights and whether your edition permits automation, backend/service integration, and disclosure of generated output. The published EULA restricts public or third-party disclosure of decompiled code, reconstructed artifacts, and analysis reports unless expressly permitted by the license type. |
+
+The vendor's applicable agreement controls. These summaries do not interpret
+your individual contract or expand the rights it grants. Verify permissions
+before publishing analysis output, sharing tool access, or operating a service.
+Vendor terms above were checked on 2026-10-12.
+
+## Runtimes, development tooling, and fixtures
+
+REA uses a caller-supplied Node.js runtime ([MIT and component notices](https://github.com/nodejs/node/blob/main/LICENSE)).
+Optional workflows may also use Python ([PSF and component licenses](https://docs.python.org/3/license.html)),
+OpenJDK ([GPLv2 with Classpath Exception and component notices](https://openjdk.org/legal/)),
+Electron ([MIT and component notices](https://github.com/electron/electron/blob/main/LICENSE)),
+or a Chromium-based browser ([BSD-3-Clause and third-party licenses](https://chromium.googlesource.com/chromium/src/+/main/LICENSE)).
+These external installations are not licensed or bundled by REA's MIT license.
+
+Development, test, and website npm packages are included in the locked inventory.
+GitHub Actions, operating-system utilities, downloaded verification fixtures,
+and their runtimes retain the terms supplied by their publishers. Fixture
+identities and provenance are recorded under `scripts/fixtures/` and the
+existing `third_party/` records; downloaded executable fixtures are not tracked
+or included in the npm package. This notice does not claim to inventory every
+dependency of a caller's operating system, agent, or separately installed tool.
+
+## Maintaining and redistributing notices
+
+Refresh the dependency inventory and runtime notice collection when the lockfile
+changes. Review the actual versions and distributed contents, retain the
+upstream license and copyright texts, and account for optional platform
+packages. See [licenses/README.md](licenses/README.md) for source provenance,
+collection scope, and known upstream notice gaps.
+
+The npm package includes this document, the `licenses/` directory, and the
+existing redistributed license records under `third_party/`. If you additionally
+bundle an external engine or modify upstream code, review that actual
+distribution and retain its full applicable licenses, copyright notices,
+NOTICE files, modification notices, and source obligations. REA's inventory is
+not a substitute for those obligations.
@@ -3,8 +3,14 @@ import com.google.gson.*;
 import io.modelcontextprotocol.kotlin.sdk.server.*;
 import io.modelcontextprotocol.kotlin.sdk.types.*;
 import jadx.api.*;
+import jadx.api.plugins.input.data.ICodeReader;
+import jadx.api.plugins.input.data.IMethodRef;
+import jadx.api.plugins.input.insns.InsnIndexType;
+import jadx.api.plugins.input.insns.Opcode;
+import jadx.core.dex.info.MethodInfo;
 import jadx.core.dex.nodes.*;
 import java.io.PrintStream;
+import java.nio.charset.StandardCharsets;
 import java.util.*;
 import java.util.concurrent.CountDownLatch;
 import java.util.stream.Collectors;
@@ -79,11 +85,17 @@ class ReaJadxBridge {
          ordered.sort(Comparator.comparing(MethodNode::getAlias));
          List<MethodSnapshot> methods = ordered.stream().map(method -> {
            String name = method.getAlias();
+            String shortId = method.getMethodInfo().getShortId();
+            String rawName = method.getMethodInfo().getName();
+            if (!shortId.startsWith(rawName + "("))
+              throw new IllegalStateException("JADX method ID has no raw-name prefix: " + shortId);
+            String descriptor = shortId.substring(rawName.length());
            String signature = name + "(" + method.getArgTypes().stream()
                .map(Object::toString).collect(Collectors.joining(", "))
                + "): " + method.getReturnType().toString();
-            return new MethodSnapshot(name, method.getMethodInfo().getShortId(), Map.of(
+            return new MethodSnapshot(name, shortId, Map.of(
                "name", name, "signature", signature,
+                "dex_descriptor", descriptor,
                "is_constructor", method.isConstructor(), "def_pos", 0));
          }).toList();
          List<Map<String, Object>> fields = node.getFields().stream()
@@ -126,6 +138,71 @@ class ReaJadxBridge {
    return result;
  }

+  private JadxSession.SmartCode methodBody(JadxSession current, ClassSnapshot cls,
+      MethodSnapshot selected, JavaMethod method) {
+    if (!selected.shortId().equals("<clinit>()V"))
+      return current.getMethodBodySmart(method, current.getMaxSourceBytes(), true);
+    // Java code generation can fold all or part of <clinit> into field declarations.
+    // Its method text cannot establish coverage, even when a static block remains.
+    JadxSession.MethodSmali smali = current.getMethodSmali(cls.node(), method.getName());
+    if (!smali.getFound() || smali.getBlocks().size() != 1)
+      throw new IllegalArgumentException("Static initializer disassembly requires one exact method: " + selected.shortId());
+    return new JadxSession.SmartCode(
+        TextUtilKt.truncateToBytes(smali.getBlocks().get(0), current.getMaxSourceBytes()),
+        "smali", true, List.of("static_initializer_disassembly"));
+  }
+
+  // The local transport admits 8 MiB frames. Reserve 1 KiB for the fixed
+  // JSON-RPC/CallToolResult envelope and generated numeric request id, and
+  // account for JSON text being escaped again inside TextContent.
+  private static final int MAX_XREF_TEXT_BYTES = 8 * 1024 * 1024 - 1024;
+  private static int escapedJsonBytes(Object value) {
+    return JSON.toJson(JSON.toJson(value)).getBytes(StandardCharsets.UTF_8).length;
+  }
+  private static final class InvokeBudget {
+    long remaining;
+    InvokeBudget(long remaining) { this.remaining = Math.max(0, remaining); }
+    boolean admit(Map<String, Object> site) {
+      // Check a conservative two-layer escape bound before serialization.
+      long bound = 256;
+      for (Object value : site.values())
+        if (value instanceof String text) bound += 12L * text.length();
+      if (bound > remaining) return false;
+      long bytes = escapedJsonBytes(site) - 2L + 1L;
+      if (bytes > remaining) return false;
+      remaining -= bytes;
+      return true;
+    }
+  }
+  private record DirectInvokes(List<Map<String, Object>> sites, boolean exhaustive) {}
+
+  /** Raw DEX invokes within one JADX using method; offsets are code units from its code start. */
+  private static DirectInvokes directInvokes(JavaMethod source, MethodInfo target, InvokeBudget budget) {
+    MethodNode node = source.getMethodNode();
+    ICodeReader code = node.getCodeReader();
+    if (code == null) return null;
+    List<Map<String, Object>> sites = new ArrayList<>();
+    boolean[] exhaustive = {true};
+    code.visitInstructions(insn -> {
+      insn.decode();
+      if (insn.getIndexType() != InsnIndexType.METHOD_REF) return;
+      Opcode opcode = insn.getOpcode();
+      if (!opcode.name().startsWith("INVOKE_")) return;
+      IMethodRef ref = insn.getIndexAsMethod();
+      ref.load();
+      if (!MethodInfo.fromRef(node.root(), ref).getRawFullId().equals(target.getRawFullId())) return;
+      Map<String, Object> site = Map.of(
+          "dex_entry", node.getInputFileName(),
+          "source_raw_method_id", node.getMethodInfo().getRawFullId(),
+          "dex_code_unit_offset", insn.getOffset(),
+          "dex_method_index", insn.getIndex(),
+          "opcode", opcode.name());
+      if (budget.admit(site)) sites.add(site);
+      else exhaustive[0] = false;
+    });
+    return new DirectInvokes(sites, exhaustive[0]);
+  }
+
  private CallToolResult handle(String name, JsonObject args) {
    if (name.equals("rea_jvm_status")) return json(Map.of(
        "engine_reported_version", BuildInfo.VERSION,
@@ -150,7 +227,7 @@ class ReaJadxBridge {
      int index = integer(args, "overload_index", 0);
      if (index >= overloads.size()) throw new InputFailure("overload_index", "overload_index out of range: " + index);
      JavaMethod method = method(cls, overloads.get(index));
-      JadxSession.SmartCode code = current.getMethodBodySmart(method, current.getMaxSourceBytes(), true);
+      JadxSession.SmartCode code = methodBody(current, cls, overloads.get(index), method);
      return json(Map.of("class_name", cls.summary().get("full_name"),
          "method_name", overloads.get(index).name(), "full_name", method.getFullName(),
          "overload_index", index, "overload_count", overloads.size(),
@@ -159,17 +236,52 @@ class ReaJadxBridge {
    }
    List<JavaNode> uses;
    String target = string(args, "class_name");
+    MethodSnapshot selectedMethod = null;
+    int overloadIndex = 0;
    if (name.equals("get_xrefs_to_class")) uses = cls.node().getUseIn();
    else {
      List<MethodSnapshot> overloads = methods(cls, args);
-      if (overloads.size() != 1) throw new InputFailure("method_name", "Method references require a unique overload");
-      target += "." + overloads.get(0).name();
-      uses = method(cls, overloads.get(0)).getUseIn();
+      if (overloads.size() != 1 && !args.containsKey("overload_index"))
+        throw new InputFailure("overload_index", "Select one of " + overloads.size() + " method overloads");
+      overloadIndex = integer(args, "overload_index", 0);
+      if (overloadIndex >= overloads.size())
+        throw new InputFailure("overload_index", "overload_index out of range: " + overloadIndex);
+      selectedMethod = overloads.get(overloadIndex);
+      target += "." + selectedMethod.name();
+      uses = method(cls, selectedMethod).getUseIn();
    }
    // REA consumes all references without materializing source line information.
-    List<Map<String, Object>> items = uses.stream().map(node -> current.describeUsage(node, false)).toList();
-    return json(Map.of("target", target, "total", uses.size(), "count", items.size(),
-        "resolve_line", false, "items", items));
+    MethodInfo targetInfo = selectedMethod == null ? null : method(cls, selectedMethod).getMethodNode().getMethodInfo();
+    List<Map<String, Object>> items = uses.stream().map(node -> {
+      Map<String, Object> item = new LinkedHashMap<>(current.describeUsage(node, false));
+      if (targetInfo != null && node instanceof JavaMethod) {
+        item.put("direct_invokes", List.of());
+        item.put("direct_invokes_exhaustive", false);
+      }
+      return item;
+    }).toList();
+    Map<String, Object> result = new LinkedHashMap<>(Map.of("target", target,
+        "total", uses.size(), "count", items.size(), "resolve_line", false, "items", items));
+    if (selectedMethod != null) {
+      result.put("overload_index", overloadIndex);
+      result.put("reported_signature", selectedMethod.summary().get("signature"));
+    }
+    if (targetInfo != null) {
+      InvokeBudget budget = new InvokeBudget(MAX_XREF_TEXT_BYTES - escapedJsonBytes(result));
+      for (int index = 0; index < uses.size(); index++) {
+        if (!(uses.get(index) instanceof JavaMethod source)) continue;
+        DirectInvokes invokes = directInvokes(source, targetInfo, budget);
+        Map<String, Object> item = items.get(index);
+        if (invokes == null) {
+          item.remove("direct_invokes");
+          item.remove("direct_invokes_exhaustive");
+        } else {
+          item.put("direct_invokes", invokes.sites());
+          item.put("direct_invokes_exhaustive", invokes.exhaustive());
+        }
+      }
+    }
+    return json(result);
  }

  public static void main(String[] args) throws Exception {
@@ -203,7 +315,7 @@ class ReaJadxBridge {
      List<String> fields = switch (name) {
        case "list_classes" -> List.of("offset", "limit");
        case "get_method_body" -> List.of("class_name", "method_name", "overload_index");
-        case "get_xrefs_to_method" -> List.of("class_name", "method_name");
+        case "get_xrefs_to_method" -> List.of("class_name", "method_name", "overload_index");
        default -> List.of("class_name");
      };
      Map<String, kotlinx.serialization.json.JsonElement> properties = new LinkedHashMap<>();
@@ -211,8 +323,8 @@ class ReaJadxBridge {
      String description = switch (name) {
        case "list_classes" -> "Indexed pages of all top-level class names, with offset and limit.";
        case "get_class_summary" -> "Parsed member metadata for one exact class name; no source generation.";
-        case "get_method_body" -> "Source for an exact class and stable same-name overload index, with smali fallback.";
-        default -> "Complete static references to an exact class or unique method; source lines are unresolved.";
+        case "get_method_body" -> "Source for an exact class and stable same-name overload index; static initializers use smali to preserve folded field initialization.";
+        default -> "Complete static using-node inventory for an exact class or selected method overload; one node may contain multiple invoke instructions, and source lines are unresolved.";
      };
      Tool narrowed = new Tool(name, new ToolSchema(null, new JsonObject(properties),
          tool.getInputSchema().getRequired(), null), description, null, null, null, null, null, null);
@@ -0,0 +1,816 @@
+"""REA's experimental local bridge plugin for Cutter.
+
+Install this file in Cutter's user plugin directory. The bridge is REA-owned
+IPC, not an upstream Cutter transport. It binds to loopback and requires a
+per-process random token stored in a user-only descriptor file.
+"""
+
+from __future__ import annotations
+
+import json
+import hashlib
+import hmac
+import os
+import secrets
+import socket
+import stat
+import tempfile
+import threading
+import time
+import uuid
+from pathlib import Path
+from typing import Any
+
+if os.name == "nt":
+    import ctypes
+    from ctypes import wintypes
+
+import cutter
+
+try:
+    from PySide6.QtCore import QObject, Signal
+except ImportError:
+    from PySide2.QtCore import QObject, Signal
+
+
+MAX_FRAME_BYTES = 16 * 1024 * 1024
+SOCKET_TIMEOUT_SECONDS = 20
+AUTH_FRAME_TIMEOUT_SECONDS = 1
+MAX_DESCRIPTOR_BYTES = 64 * 1024
+MAX_COMPLETION_RECEIPTS = 64
+
+
+def _windows_api() -> tuple[Any, Any]:
+    kernel32 = ctypes.WinDLL("kernel32", use_last_error=True)
+    advapi32 = ctypes.WinDLL("advapi32", use_last_error=True)
+    kernel32.GetCurrentProcess.restype = wintypes.HANDLE
+    kernel32.LocalFree.argtypes = [wintypes.HLOCAL]
+    kernel32.LocalFree.restype = wintypes.HLOCAL
+    kernel32.CloseHandle.argtypes = [wintypes.HANDLE]
+    kernel32.CloseHandle.restype = wintypes.BOOL
+    kernel32.GetFileAttributesW.argtypes = [wintypes.LPCWSTR]
+    kernel32.GetFileAttributesW.restype = wintypes.DWORD
+    kernel32.CreateDirectoryW.argtypes = [wintypes.LPCWSTR, wintypes.LPVOID]
+    kernel32.CreateDirectoryW.restype = wintypes.BOOL
+    kernel32.CreateFileW.argtypes = [wintypes.LPCWSTR, wintypes.DWORD, wintypes.DWORD,
+                                     wintypes.LPVOID, wintypes.DWORD, wintypes.DWORD, wintypes.HANDLE]
+    kernel32.CreateFileW.restype = wintypes.HANDLE
+    kernel32.WriteFile.argtypes = [wintypes.HANDLE, wintypes.LPVOID, wintypes.DWORD,
+                                   ctypes.POINTER(wintypes.DWORD), wintypes.LPVOID]
+    kernel32.WriteFile.restype = wintypes.BOOL
+    kernel32.FlushFileBuffers.argtypes = [wintypes.HANDLE]
+    kernel32.FlushFileBuffers.restype = wintypes.BOOL
+    kernel32.MoveFileExW.argtypes = [wintypes.LPCWSTR, wintypes.LPCWSTR, wintypes.DWORD]
+    kernel32.MoveFileExW.restype = wintypes.BOOL
+    kernel32.DeleteFileW.argtypes = [wintypes.LPCWSTR]
+    kernel32.DeleteFileW.restype = wintypes.BOOL
+    advapi32.OpenProcessToken.argtypes = [wintypes.HANDLE, wintypes.DWORD, ctypes.POINTER(wintypes.HANDLE)]
+    advapi32.OpenProcessToken.restype = wintypes.BOOL
+    advapi32.GetTokenInformation.argtypes = [wintypes.HANDLE, ctypes.c_int, wintypes.LPVOID,
+                                              wintypes.DWORD, ctypes.POINTER(wintypes.DWORD)]
+    advapi32.GetTokenInformation.restype = wintypes.BOOL
+    advapi32.ConvertSidToStringSidW.argtypes = [wintypes.LPVOID, ctypes.POINTER(wintypes.LPWSTR)]
+    advapi32.ConvertSidToStringSidW.restype = wintypes.BOOL
+    advapi32.GetNamedSecurityInfoW.argtypes = [wintypes.LPWSTR, ctypes.c_int, wintypes.DWORD,
+                                                ctypes.POINTER(wintypes.LPVOID), ctypes.POINTER(wintypes.LPVOID),
+                                                ctypes.POINTER(wintypes.LPVOID), ctypes.POINTER(wintypes.LPVOID),
+                                                ctypes.POINTER(wintypes.LPVOID)]
+    advapi32.GetNamedSecurityInfoW.restype = wintypes.DWORD
+    advapi32.GetSecurityDescriptorControl.argtypes = [wintypes.LPVOID, ctypes.POINTER(wintypes.WORD),
+                                                       ctypes.POINTER(wintypes.DWORD)]
+    advapi32.GetSecurityDescriptorControl.restype = wintypes.BOOL
+    advapi32.EqualSid.argtypes = [wintypes.LPVOID, wintypes.LPVOID]
+    advapi32.EqualSid.restype = wintypes.BOOL
+    advapi32.GetAclInformation.argtypes = [wintypes.LPVOID, wintypes.LPVOID, wintypes.DWORD, ctypes.c_int]
+    advapi32.GetAclInformation.restype = wintypes.BOOL
+    advapi32.GetAce.argtypes = [wintypes.LPVOID, wintypes.DWORD, ctypes.POINTER(wintypes.LPVOID)]
+    advapi32.GetAce.restype = wintypes.BOOL
+    advapi32.IsValidSid.argtypes = [wintypes.LPVOID]
+    advapi32.IsValidSid.restype = wintypes.BOOL
+    advapi32.GetLengthSid.argtypes = [wintypes.LPVOID]
+    advapi32.GetLengthSid.restype = wintypes.DWORD
+    advapi32.ConvertStringSecurityDescriptorToSecurityDescriptorW.argtypes = [
+        wintypes.LPCWSTR, wintypes.DWORD, ctypes.POINTER(wintypes.LPVOID), ctypes.POINTER(wintypes.DWORD),
+    ]
+    advapi32.ConvertStringSecurityDescriptorToSecurityDescriptorW.restype = wintypes.BOOL
+    return kernel32, advapi32
+
+
+def _windows_user_sid_text() -> str:
+    kernel32, advapi32 = _windows_api()
+    token = wintypes.HANDLE()
+    if not advapi32.OpenProcessToken(kernel32.GetCurrentProcess(), 0x0008, ctypes.byref(token)):
+        raise ctypes.WinError(ctypes.get_last_error())
+    try:
+        required = wintypes.DWORD()
+        advapi32.GetTokenInformation(token, 1, None, 0, ctypes.byref(required))
+        if required.value == 0:
+            raise ctypes.WinError(ctypes.get_last_error())
+        buffer = ctypes.create_string_buffer(required.value)
+        if not advapi32.GetTokenInformation(token, 1, buffer, required, ctypes.byref(required)):
+            raise ctypes.WinError(ctypes.get_last_error())
+
+        class SID_AND_ATTRIBUTES(ctypes.Structure):
+            _fields_ = [("Sid", wintypes.LPVOID), ("Attributes", wintypes.DWORD)]
+
+        class TOKEN_USER(ctypes.Structure):
+            _fields_ = [("User", SID_AND_ATTRIBUTES)]
+
+        sid = ctypes.cast(buffer, ctypes.POINTER(TOKEN_USER)).contents.User.Sid
+        text = wintypes.LPWSTR()
+        if not advapi32.ConvertSidToStringSidW(sid, ctypes.byref(text)):
+            raise ctypes.WinError(ctypes.get_last_error())
+        try:
+            return text.value
+        finally:
+            kernel32.LocalFree(text)
+    finally:
+        kernel32.CloseHandle(token)
+
+
+def _verify_windows_private_path(path: Path, directory: bool) -> None:
+    kernel32, advapi32 = _windows_api()
+    attributes = kernel32.GetFileAttributesW(str(path))
+    if attributes == 0xFFFFFFFF:
+        raise ctypes.WinError(ctypes.get_last_error())
+    if attributes & 0x400 or bool(attributes & 0x10) != directory:
+        raise RuntimeError("Cutter bridge path is a reparse point or has the wrong object type")
+
+    class ACL_SIZE_INFORMATION(ctypes.Structure):
+        _fields_ = [("AceCount", wintypes.DWORD), ("AclBytesInUse", wintypes.DWORD), ("AclBytesFree", wintypes.DWORD)]
+
+    class ACE_HEADER(ctypes.Structure):
+        _fields_ = [("AceType", wintypes.BYTE), ("AceFlags", wintypes.BYTE), ("AceSize", wintypes.WORD)]
+
+    class SID_HEADER(ctypes.Structure):
+        _fields_ = [("Revision", wintypes.BYTE), ("SubAuthorityCount", wintypes.BYTE),
+                    ("IdentifierAuthority", wintypes.BYTE * 6)]
+
+    class ACCESS_ALLOWED_ACE(ctypes.Structure):
+        _fields_ = [("Header", ACE_HEADER), ("Mask", wintypes.DWORD), ("SidStart", wintypes.DWORD)]
+
+    owner = wintypes.LPVOID()
+    dacl = wintypes.LPVOID()
+    descriptor = wintypes.LPVOID()
+    result = advapi32.GetNamedSecurityInfoW(
+        str(path), 1, 0x1 | 0x4, ctypes.byref(owner), None, ctypes.byref(dacl), None, ctypes.byref(descriptor),
+    )
+    if result != 0:
+        raise ctypes.WinError(result)
+    try:
+        control = wintypes.WORD()
+        revision = wintypes.DWORD()
+        if not advapi32.GetSecurityDescriptorControl(descriptor, ctypes.byref(control), ctypes.byref(revision)):
+            raise ctypes.WinError(ctypes.get_last_error())
+        if not control.value & 0x1000 or not dacl:
+            raise RuntimeError("Cutter bridge ACL must be present and protected from inheritance")
+
+        token = wintypes.HANDLE()
+        if not advapi32.OpenProcessToken(kernel32.GetCurrentProcess(), 0x0008, ctypes.byref(token)):
+            raise ctypes.WinError(ctypes.get_last_error())
+        try:
+            required = wintypes.DWORD()
+            advapi32.GetTokenInformation(token, 1, None, 0, ctypes.byref(required))
+            token_user = ctypes.create_string_buffer(required.value)
+            if not advapi32.GetTokenInformation(token, 1, token_user, required, ctypes.byref(required)):
+                raise ctypes.WinError(ctypes.get_last_error())
+
+            class SID_AND_ATTRIBUTES(ctypes.Structure):
+                _fields_ = [("Sid", wintypes.LPVOID), ("Attributes", wintypes.DWORD)]
+
+            class TOKEN_USER(ctypes.Structure):
+                _fields_ = [("User", SID_AND_ATTRIBUTES)]
+
+            current_sid = ctypes.cast(token_user, ctypes.POINTER(TOKEN_USER)).contents.User.Sid
+            if not advapi32.EqualSid(owner, current_sid):
+                raise RuntimeError("Cutter bridge path owner does not match the current user")
+        finally:
+            kernel32.CloseHandle(token)
+
+        info = ACL_SIZE_INFORMATION()
+        if not advapi32.GetAclInformation(dacl, ctypes.byref(info), ctypes.sizeof(info), 2) or info.AceCount != 2:
+            raise RuntimeError("Cutter bridge ACL must allow only the current user and SYSTEM")
+        expected_user = _windows_user_sid_text()
+        user_allowed = False
+        system_allowed = False
+        for index in range(info.AceCount):
+            ace_pointer = wintypes.LPVOID()
+            if not advapi32.GetAce(dacl, index, ctypes.byref(ace_pointer)):
+                raise ctypes.WinError(ctypes.get_last_error())
+            header = ctypes.cast(ace_pointer, ctypes.POINTER(ACE_HEADER)).contents
+            if header.AceType != 0 or header.AceSize < ctypes.sizeof(ACE_HEADER) + ctypes.sizeof(wintypes.DWORD) + ctypes.sizeof(SID_HEADER):
+                raise RuntimeError("Cutter bridge ACL contains a malformed or unsupported ACE")
+            ace_address = ctypes.cast(ace_pointer, ctypes.c_void_p).value
+            if ace_address is None:
+                raise RuntimeError("Cutter bridge ACL returned a null ACE")
+            sid_pointer = ctypes.c_void_p(ace_address + ctypes.sizeof(ACE_HEADER) + ctypes.sizeof(wintypes.DWORD))
+            sid_header = ctypes.cast(sid_pointer, ctypes.POINTER(SID_HEADER)).contents
+            sid_size = ctypes.sizeof(SID_HEADER) + sid_header.SubAuthorityCount * ctypes.sizeof(wintypes.DWORD)
+            if sid_size > header.AceSize - ctypes.sizeof(ACE_HEADER) - ctypes.sizeof(wintypes.DWORD) or not advapi32.IsValidSid(sid_pointer) or advapi32.GetLengthSid(sid_pointer) != sid_size:
+                raise RuntimeError("Cutter bridge ACL contains an invalid SID")
+            ace = ctypes.cast(ace_pointer, ctypes.POINTER(ACCESS_ALLOWED_ACE)).contents
+            sid_text = wintypes.LPWSTR()
+            if ace.Mask != 0x001F01FF or ace.Header.AceFlags != 0x03:
+                raise RuntimeError("Cutter bridge ACL contains an unsupported ACE")
+            if not advapi32.ConvertSidToStringSidW(sid_pointer, ctypes.byref(sid_text)):
+                raise ctypes.WinError(ctypes.get_last_error())
+            try:
+                if sid_text.value == expected_user:
+                    user_allowed = True
+                elif sid_text.value == "S-1-5-18":
+                    system_allowed = True
+                else:
+                    raise RuntimeError("Cutter bridge ACL allows an unrelated principal")
+            finally:
+                kernel32.LocalFree(sid_text)
+        if not user_allowed or not system_allowed:
+            raise RuntimeError("Cutter bridge ACL must allow only the current user and SYSTEM")
+    finally:
+        kernel32.LocalFree(descriptor)
+
+
+def _windows_security_attributes() -> tuple[Any, Any]:
+    kernel32, advapi32 = _windows_api()
+    sid = _windows_user_sid_text()
+    sddl = f"O:{sid}G:{sid}D:P(A;OICI;FA;;;{sid})(A;OICI;FA;;;SY)"
+    descriptor = wintypes.LPVOID()
+    if not advapi32.ConvertStringSecurityDescriptorToSecurityDescriptorW(sddl, 1, ctypes.byref(descriptor), None):
+        raise ctypes.WinError(ctypes.get_last_error())
+
+    class SECURITY_ATTRIBUTES(ctypes.Structure):
+        _fields_ = [("nLength", wintypes.DWORD), ("lpSecurityDescriptor", wintypes.LPVOID), ("bInheritHandle", wintypes.BOOL)]
+
+    attributes = SECURITY_ATTRIBUTES(ctypes.sizeof(SECURITY_ATTRIBUTES), descriptor, False)
+    return kernel32, (descriptor, attributes)
+
+
+def _ensure_windows_bridge_directory(directory: Path) -> None:
+    directory.parent.mkdir(parents=True, exist_ok=True)
+    if not directory.exists():
+        kernel32, (descriptor, attributes) = _windows_security_attributes()
+        try:
+            if not kernel32.CreateDirectoryW(str(directory), ctypes.byref(attributes)):
+                error = ctypes.get_last_error()
+                if error != 183:
+                    raise ctypes.WinError(error)
+        finally:
+            kernel32.LocalFree(descriptor)
+    _verify_windows_private_path(directory, True)
+
+
+def _ensure_posix_bridge_directory(directory: Path) -> None:
+    directory.mkdir(mode=0o700, parents=True, exist_ok=True)
+    if not hasattr(os, "O_NOFOLLOW") or not hasattr(os, "O_DIRECTORY"):
+        raise RuntimeError("Cutter bridge directory requires no-follow directory handles")
+    descriptor = os.open(directory, os.O_RDONLY | os.O_DIRECTORY | os.O_NOFOLLOW)
+    try:
+        directory_info = os.fstat(descriptor)
+        if directory_info.st_uid != os.getuid():
+            raise RuntimeError("Cutter bridge directory must be owned by the current user")
+        os.fchmod(descriptor, 0o700)
+        directory_info = os.fstat(descriptor)
+        if stat.S_IMODE(directory_info.st_mode) != 0o700:
+            raise RuntimeError("Cutter bridge directory must have mode 0700")
+    finally:
+        os.close(descriptor)
+
+
+def _write_windows_descriptor(directory: Path, destination: Path, payload: bytes) -> None:
+    if len(payload) > MAX_DESCRIPTOR_BYTES:
+        raise RuntimeError("Cutter bridge descriptor exceeds its byte limit")
+    kernel32, (security_descriptor, attributes) = _windows_security_attributes()
+    temporary = directory / f".{destination.name}.{uuid.uuid4().hex}.tmp"
+    handle = wintypes.HANDLE()
+    try:
+        handle = kernel32.CreateFileW(str(temporary), 0x40000000 | 0x00020000 | 0x00040000,
+                                      0x1 | 0x2 | 0x4, ctypes.byref(attributes), 1, 0x80, None)
+        if handle == wintypes.HANDLE(-1).value:
+            raise ctypes.WinError(ctypes.get_last_error())
+        _verify_windows_private_path(temporary, False)
+        if payload:
+            data = ctypes.create_string_buffer(payload)
+            written = wintypes.DWORD()
+            if not kernel32.WriteFile(handle, data, len(payload), ctypes.byref(written), None) or written.value != len(payload):
+                raise ctypes.WinError(ctypes.get_last_error())
+        if not kernel32.FlushFileBuffers(handle):
+            raise ctypes.WinError(ctypes.get_last_error())
+        kernel32.CloseHandle(handle)
+        handle = wintypes.HANDLE()
+        if not kernel32.MoveFileExW(str(temporary), str(destination), 0x1 | 0x8):
+            raise ctypes.WinError(ctypes.get_last_error())
+    finally:
+        if handle and handle != wintypes.HANDLE(-1).value:
+            kernel32.CloseHandle(handle)
+        kernel32.DeleteFileW(str(temporary))
+        kernel32.LocalFree(security_descriptor)
+
+
+def _bridge_directory() -> Path:
+    configured = os.environ.get("REA_CUTTER_BRIDGE_DIR")
+    if configured:
+        return Path(configured).expanduser()
+    local_app_data = os.environ.get("LOCALAPPDATA")
+    if os.name == "nt" and local_app_data:
+        return Path(local_app_data) / "REA" / "CutterBridge"
+    return Path.home() / ".cache" / "rea" / "cutter-bridge"
+
+
+class _RequestDispatcher(QObject):
+    dispatch = Signal(object, object)
+
+    def __init__(self) -> None:
+        super().__init__()
+        self.dispatch.connect(self._handle)
+        self.generation = 0
+        self.completed_commands: dict[str, str] = {}
+        self._receipt_lock = threading.Lock()
+        self.current_file = self._read_current_file()
+        try:
+            self.cutter_version = str(cutter.core().getVersionInformation())
+        except Exception:
+            self.cutter_version = None
+
+    @staticmethod
+    def _read_current_file() -> str | None:
+        try:
+            value = cutter.core().currentlyOpenFile()
+            return str(value) if value else None
+        except Exception:
+            return None
+
+    def _handle(self, request: dict[str, Any], reply: dict[str, Any]) -> None:
+        kind = request.get("kind")
+        request_id = request.get("request_id")
+        if kind in ("command", "ack"):
+            try:
+                valid_request_id = (
+                    isinstance(request_id, str)
+                    and str(uuid.UUID(request_id)) == request_id.lower()
+                )
+            except (ValueError, AttributeError):
+                valid_request_id = False
+            if not valid_request_id:
+                reply.update({
+                    "ok": False,
+                    "error": "invalid-command-request-id",
+                    "execution_state": "not_started",
+                })
+                request["_event"].set()
+                return
+        if kind == "ack":
+            with self._receipt_lock:
+                acknowledged = self.completed_commands.pop(request_id, None) is not None
+            reply.update({"ok": True, "acknowledged": acknowledged})
+            request["_event"].set()
+            return
+        if request["_cancelled"].is_set():
+            reply.update({
+                "ok": False,
+                "error": "request-cancelled-before-execution",
+                "execution_state": "not_started",
+            })
+            if kind == "command":
+                self._record_command_completion(request_id, "not_started")
+            request["_event"].set()
+            return
+        self._refresh_document_identity()
+        if kind == "status":
+            with self._receipt_lock:
+                completed_commands = [
+                    {"request_id": receipt_id, "execution_state": state}
+                    for receipt_id, state in self.completed_commands.items()
+                ]
+            reply.update({
+                "ok": True,
+                "session_id": request.get("session_id"),
+                "current_file": self.current_file,
+                "document_generation": self.generation,
+                "cutter_version": self.cutter_version,
+                "identity_status": "partial",
+                "completed_commands": completed_commands,
+            })
+            request["_event"].set()
+            return
+        if kind != "command":
+            reply.update({"ok": False, "error": "unsupported-request-kind"})
+            request["_event"].set()
+            return
+        if request.get("expected_generation") != self.generation:
+            reply.update({
+                "ok": False,
+                "error": "active-document-generation-changed",
+                "execution_state": "not_started",
+            })
+            self._record_command_completion(request_id, "not_started")
+            request["_event"].set()
+            return
+        command = request.get("command")
+        if not isinstance(command, str) or not command.strip():
+            reply.update({
+                "ok": False,
+                "error": "invalid-command",
+                "execution_state": "not_started",
+            })
+            self._record_command_completion(request_id, "not_started")
+            request["_event"].set()
+            return
+        reply["execution_state"] = "running"
+        try:
+            output = cutter.cmdj(command) if request.get("json") is True else cutter.cmd(command)
+            command_error = None
+        except Exception as error:
+            output = None
+            command_error = str(error)
+        self._refresh_document_identity()
+        if command_error is None:
+            reply.update({
+                "ok": True,
+                "output": output,
+                "current_file": self.current_file,
+                "document_generation": self.generation,
+                "cutter_version": self.cutter_version,
+                "identity_status": "partial",
+                "execution_state": "complete",
+            })
+            self._record_command_completion(request_id, "complete")
+        else:
+            reply.update({
+                "ok": False,
+                "error": command_error,
+                "current_file": self.current_file,
+                "document_generation": self.generation,
+                "cutter_version": self.cutter_version,
+                "identity_status": "partial",
+                "execution_state": "unknown",
+                "message": "The command raised an exception; it may have produced partial effects. Do not retry automatically.",
+            })
+            self._record_command_completion(request_id, "failed")
+        request["_event"].set()
+
+    def _record_command_completion(
+        self, request_id: str, execution_state: str
+    ) -> None:
+        with self._receipt_lock:
+            if request_id in self.completed_commands or len(self.completed_commands) < MAX_COMPLETION_RECEIPTS:
+                self.completed_commands[request_id] = execution_state
+
+    def can_accept_command(self) -> bool:
+        with self._receipt_lock:
+            return len(self.completed_commands) < MAX_COMPLETION_RECEIPTS
+
+    def _refresh_document_identity(self) -> None:
+        current_file = self._read_current_file()
+        if current_file != self.current_file:
+            self.current_file = current_file
+            self.generation += 1
+
+
+class _BridgeServer(threading.Thread):
+    def __init__(self, dispatcher: _RequestDispatcher) -> None:
+        super().__init__(name="rea-cutter-bridge", daemon=True)
+        self.dispatcher = dispatcher
+        self.server = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+        self.server.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+        self.server.bind(("127.0.0.1", 0))
+        self.server.listen(8)
+        self.server.settimeout(0.5)
+        self.port = self.server.getsockname()[1]
+        self.session_id = str(uuid.uuid4())
+        self.token = secrets.token_urlsafe(32)
+        self.poisoned = False
+        self.connection_ordinal = 0
+        self.directory = _bridge_directory()
+        self.descriptor = self.directory / f"cutter-{os.getpid()}-{self.session_id}.json"
+
+    def publish(self) -> None:
+        if os.name == "nt":
+            _ensure_windows_bridge_directory(self.directory)
+        else:
+            _ensure_posix_bridge_directory(self.directory)
+        if os.name == "nt" and (self.directory.is_symlink() or not self.directory.is_dir()):
+            raise RuntimeError("Cutter bridge directory must be a real directory")
+        payload = {
+            "session_id": self.session_id,
+            "pid": os.getpid(),
+            "host": "127.0.0.1",
+            "port": self.port,
+            "token": self.token,
+            "document_generation": self.dispatcher.generation,
+            "current_file": self.dispatcher.current_file,
+            "cutter_version": self.dispatcher.cutter_version,
+            "identity_status": "partial",
+        }
+        if os.name == "nt":
+            _write_windows_descriptor(self.directory, self.descriptor, json.dumps(payload).encode("utf-8"))
+            return
+        descriptor, temporary_path = tempfile.mkstemp(prefix=f".{self.descriptor.name}.", dir=self.directory)
+        try:
+            os.fchmod(descriptor, 0o600)
+            descriptor_info = os.fstat(descriptor)
+            if descriptor_info.st_uid != os.getuid() or stat.S_IMODE(descriptor_info.st_mode) & 0o077:
+                raise RuntimeError("Cutter bridge descriptor must be owned by the current user with mode 0600")
+            with os.fdopen(descriptor, "w", encoding="utf-8") as stream:
+                stream.write(json.dumps(payload))
+                stream.flush()
+                os.fsync(stream.fileno())
+            os.replace(temporary_path, self.descriptor)
+        except Exception:
+            try:
+                os.close(descriptor)
+            except OSError:
+                pass
+            try:
+                os.unlink(temporary_path)
+            except OSError:
+                pass
+            raise
+
+    def run(self) -> None:
+        try:
+            self.publish()
+        except Exception:
+            self.server.close()
+            return
+        while True:
+            try:
+                connection, _ = self.server.accept()
+            except socket.timeout:
+                if getattr(self, "stopping", False):
+                    break
+                continue
+            except OSError:
+                break
+            self.connection_ordinal += 1
+            connection_ordinal = self.connection_ordinal
+            with connection:
+                connection.settimeout(AUTH_FRAME_TIMEOUT_SECONDS)
+                authenticated = False
+                try:
+                    hello = self._receive_frame(connection)
+                    if not isinstance(hello, dict) or hello.get("kind") != "rea-auth-v2" or not isinstance(hello.get("nonce"), str) or len(hello["nonce"]) != 64:
+                        self._send(connection, {"ok": False, "error": "authentication-failed"})
+                        continue
+                    client_nonce = hello["nonce"]
+                    server_nonce = f"{connection_ordinal:016x}{secrets.token_hex(24)}"
+                    server_proof = hmac.new(self.token.encode(), f"rea-cutter-server-v2:{self.session_id}:{client_nonce}:{server_nonce}".encode(), hashlib.sha256).hexdigest()
+                    self._send(connection, {"session_id": self.session_id, "nonce": server_nonce, "proof": server_proof})
+                    authenticated_frame = self._receive_frame(connection)
+                    request_bytes = authenticated_frame.get("request") if isinstance(authenticated_frame, dict) else None
+                    expected_proof = hmac.new(self.token.encode(), f"rea-cutter-client-v2:{self.session_id}:{client_nonce}:{server_nonce}:{request_bytes}".encode(), hashlib.sha256).hexdigest()
+                    if not isinstance(request_bytes, str) or not isinstance(authenticated_frame.get("proof"), str) or not hmac.compare_digest(authenticated_frame["proof"], expected_proof):
+                        self._send(connection, {"ok": False, "error": "authentication-failed"})
+                        continue
+                    request = json.loads(request_bytes)
+                    if not isinstance(request, dict) or request.get("session_id") != self.session_id:
+                        self._send(connection, {"ok": False, "error": "authentication-failed"})
+                        continue
+                    authenticated = True
+                    connection.settimeout(SOCKET_TIMEOUT_SECONDS)
+                    if self.poisoned and request.get("kind") not in ("status", "ack"):
+                        self._send_authenticated(connection, {
+                            "ok": False,
+                            "error": "session-poisoned-after-ui-timeout",
+                            "execution_state": "not_started",
+                        }, client_nonce, server_nonce)
+                        continue
+                    if (
+                        request.get("kind") == "command"
+                        and not self.dispatcher.can_accept_command()
+                    ):
+                        self._send_authenticated(connection, {
+                            "ok": False,
+                            "error": "completion-receipt-capacity-exceeded",
+                            "execution_state": "not_started",
+                        }, client_nonce, server_nonce)
+                        continue
+                    request["session_id"] = self.session_id
+                    event = threading.Event()
+                    cancelled = threading.Event()
+                    reply: dict[str, Any] = {}
+                    request["_event"] = event
+                    request["_cancelled"] = cancelled
+                    self.dispatcher.dispatch.emit(request, reply)
+                    if not event.wait(SOCKET_TIMEOUT_SECONDS):
+                        cancelled.set()
+                        self.poisoned = True
+                        self._send_authenticated(connection, {
+                            "ok": False,
+                            "error": "ui-thread-request-timeout",
+                            "execution_state": "unknown",
+                            "message": "The UI operation may still complete after this timeout; restart Cutter before issuing further commands.",
+                        }, client_nonce, server_nonce)
+                    else:
+                        if request.get("kind") == "status" and reply.get("ok") is True:
+                            reply["completed_connection"] = f"{connection_ordinal - 1:016x}"
+                        self._send_authenticated(connection, reply, client_nonce, server_nonce)
+                except socket.timeout:
+                    if authenticated:
+                        self._send_authenticated(connection, {
+                            "ok": False,
+                            "error": "request-timeout",
+                            "execution_state": "unknown",
+                            "message": "The command may have produced partial or persistent effects. Do not retry automatically.",
+                        }, client_nonce, server_nonce)
+                except Exception as error:
+                    self._send(connection, {"ok": False, "error": str(error)})
+        self.server.close()
+        try:
+            self.descriptor.unlink(missing_ok=True)
+        except OSError:
+            pass
+
+    @staticmethod
+    def _receive_frame(connection: socket.socket) -> dict[str, Any]:
+        data = bytearray()
+        deadline = time.monotonic() + AUTH_FRAME_TIMEOUT_SECONDS
+        while b"\n" not in data:
+            if len(data) >= MAX_FRAME_BYTES:
+                raise ValueError("frame-limit-exceeded")
+            remaining = deadline - time.monotonic()
+            if remaining <= 0:
+                raise socket.timeout("authentication frame timed out")
+            connection.settimeout(remaining)
+            part = connection.recv(min(65536, MAX_FRAME_BYTES + 1 - len(data)))
+            if not part:
+                raise ValueError("incomplete authentication frame")
+            data.extend(part)
+        line, remainder = bytes(data).split(b"\n", 1)
+        if len(line) > MAX_FRAME_BYTES or remainder:
+            raise ValueError("invalid framed request")
+        value = json.loads(line)
+        return value if isinstance(value, dict) else {}
+
+    def _send_authenticated(self, connection: socket.socket, value: dict[str, Any], client_nonce: str, server_nonce: str) -> None:
+        encoded = self._encode_authenticated_response(value, client_nonce, server_nonce)
+        if encoded is None:
+            value = self._truncated_response(value)
+            encoded = self._encode_authenticated_response(value, client_nonce, server_nonce)
+        if encoded is None:
+            # This fixed-size fallback is checked through the same signed framing path.
+            value = {
+                "ok": False,
+                "error": "response-limit-exceeded",
+                "execution_state": value.get("execution_state", "unknown"),
+                "output": None,
+                "output_truncated": True,
+                "message": "The result exceeded the bridge frame limit; command completion state is preserved.",
+            }
+            encoded = self._encode_authenticated_response(value, client_nonce, server_nonce)
+        if encoded is None:
+            return
+        try:
+            connection.sendall(encoded)
+        except OSError:
+            pass
+
+    def _encode_authenticated_response(self, value: dict[str, Any], client_nonce: str, server_nonce: str) -> bytes | None:
+        if not _within_response_limit(value):
+            return None
+        prefix = b'{"response":"'
+        suffix = b'","proof":"' + (b"0" * 64) + b'"}\n'
+        frame_size = len(prefix) + len(suffix)
+        response_parts: list[bytes] = []
+        escaped_parts: list[bytes] = []
+        for fragment in json.JSONEncoder(ensure_ascii=False, separators=(",", ":")).iterencode(value):
+            raw = fragment.encode("utf-8")
+            escaped = json.dumps(fragment, ensure_ascii=False)[1:-1].encode("utf-8")
+            frame_size += len(escaped)
+            if frame_size > MAX_FRAME_BYTES:
+                return None
+            response_parts.append(raw)
+            escaped_parts.append(escaped)
+        proof = hmac.new(self.token.encode(), digestmod=hashlib.sha256)
+        proof.update(f"rea-cutter-response-v2:{self.session_id}:{client_nonce}:{server_nonce}:".encode())
+        for part in response_parts:
+            proof.update(part)
+        return prefix + b"".join(escaped_parts) + b'","proof":"' + proof.hexdigest().encode() + b'"}\n'
+
+    @staticmethod
+    def _truncated_response(value: dict[str, Any]) -> dict[str, Any]:
+        response: dict[str, Any] = {
+            "ok": value.get("ok") is True,
+            "output": None,
+            "execution_state": value.get("execution_state", "unknown"),
+            "output_truncated": True,
+            "message": "The command result exceeded the bridge frame limit; completion state is preserved and output was omitted.",
+        }
+        for key in ("current_file", "document_generation", "cutter_version", "identity_status"):
+            if key in value:
+                response[key] = value[key]
+        if value.get("ok") is not True and isinstance(value.get("error"), str):
+            response["error"] = value["error"]
+        return response
+
+    @staticmethod
+    def _send(connection: socket.socket, value: dict[str, Any]) -> None:
+        if not _within_response_limit(value):
+            value = {
+                "ok": False,
+                "error": "response-limit-exceeded",
+                "execution_state": value.get("execution_state", "unknown"),
+                "output_truncated": True,
+                "message": "The command completed or may have completed, but its result exceeded the bridge response limit. Do not retry automatically.",
+            }
+        parts: list[bytes] = []
+        encoded_size = 0
+        for part in json.JSONEncoder(ensure_ascii=False).iterencode(value):
+            encoded_part = part.encode("utf-8")
+            encoded_size += len(encoded_part)
+            if encoded_size > MAX_FRAME_BYTES:
+                fallback = {
+                    "ok": False,
+                    "error": "response-limit-exceeded",
+                    "execution_state": value.get("execution_state", "unknown"),
+                    "output_truncated": True,
+                    "message": "The command completed or may have completed, but its result exceeded the bridge response limit. Do not retry automatically.",
+                }
+                parts = [json.dumps(fallback, separators=(",", ":")).encode("utf-8")]
+                break
+            parts.append(encoded_part)
+        encoded = b"".join(parts)
+        try:
+            connection.sendall(encoded + b"\n")
+        except OSError:
+            # A disconnected caller must not terminate the bridge thread or
+            # prevent later commands from being served.
+            pass
+
+    def stop(self) -> None:
+        self.stopping = True
+        try:
+            self.server.close()
+        except OSError:
+            pass
+
+
+class ReaCutterPlugin(cutter.CutterPlugin):
+    name = "REA Local Bridge"
+    description = "Local REA analysis bridge for the active Cutter session"
+    version = "1.0"
+    author = "REA"
+
+    def __init__(self) -> None:
+        super().__init__()
+        self._dispatcher: _RequestDispatcher | None = None
+        self._bridge: _BridgeServer | None = None
+
+    def setupPlugin(self) -> None:
+        return
+
+    def setupInterface(self, main: Any) -> None:
+        self._dispatcher = _RequestDispatcher()
+        self._bridge = _BridgeServer(self._dispatcher)
+        self._bridge.start()
+
+    def terminate(self) -> None:
+        if self._bridge is not None:
+            self._bridge.stop()
+            self._bridge.join(timeout=2)
+            self._bridge = None
+        self._dispatcher = None
+
+
+def create_cutter_plugin() -> ReaCutterPlugin:
+    return ReaCutterPlugin()
+
+
+def _within_response_limit(value: Any) -> bool:
+    pending = [value]
+    seen: set[int] = set()
+    size = 0
+    while pending:
+        current = pending.pop()
+        if isinstance(current, str):
+            size += 2
+            if current.isascii():
+                size += len(current)
+                size += current.count('"') + current.count("\\")
+                size += sum(5 for character in current if ord(character) < 0x20)
+            else:
+                size += 4 * len(current)
+                size += current.count('"') + current.count("\\")
+                size += sum(5 for character in current if ord(character) < 0x20)
+        elif isinstance(current, dict):
+            identity = id(current)
+            if identity in seen:
+                return False
+            seen.add(identity)
+            size += 2 + 4 * len(current)
+            if size > MAX_FRAME_BYTES:
+                return False
+            pending.extend(current.keys())
+            pending.extend(current.values())
+        elif isinstance(current, (list, tuple)):
+            identity = id(current)
+            if identity in seen:
+                return False
+            seen.add(identity)
+            size += 2 + 2 * len(current)
+            if size > MAX_FRAME_BYTES:
+                return False
+            pending.extend(current)
+        else:
+            size += 16
+        if size > MAX_FRAME_BYTES:
+            return False
+    return True
@@ -0,0 +1,38 @@
+import java.util.Map;
+
+import ghidra.app.plugin.core.analysis.AutoAnalysisManager;
+import ghidra.app.util.headless.HeadlessScript;
+
+/** Disable the named auto-analyzers before default auto-analysis. */
+public final class ReaGhidraAnalysisOptions extends HeadlessScript {
+    @Override
+    public void run() throws Exception {
+        String[] arguments = getScriptArgs();
+        if (arguments.length != 1 || arguments[0].isBlank()) {
+            throw new IllegalArgumentException(
+                "ReaGhidraAnalysisOptions requires one comma-separated analyzer-name argument");
+        }
+        String[] names = arguments[0].split(",");
+        AutoAnalysisManager analysis = AutoAnalysisManager.getAnalysisManager(currentProgram);
+        Map<String, String> options = getCurrentAnalysisOptionsAndValues(currentProgram);
+        for (String requestedName : names) {
+            monitor.checkCancelled();
+            String name = requestedName.trim();
+            if (analysis.getAnalyzer(name) == null || !options.containsKey(name)) {
+                throw new IllegalArgumentException("Unknown Ghidra analyzer for this program: " + name);
+            }
+        }
+        for (String requestedName : names) {
+            monitor.checkCancelled();
+            setAnalysisOption(currentProgram, requestedName.trim(), "false");
+        }
+        options = getCurrentAnalysisOptionsAndValues(currentProgram);
+        for (String requestedName : names) {
+            monitor.checkCancelled();
+            String name = requestedName.trim();
+            if (!"false".equals(options.get(name))) {
+                throw new IllegalStateException("Ghidra analyzer remained enabled: " + name);
+            }
+        }
+    }
+}
@@ -0,0 +1,14 @@
+import org.osgi.framework.BundleActivator;
+import org.osgi.framework.BundleContext;
+
+// Compile the no-op activator with REA's scripts. Otherwise Ghidra generates
+// one afterward and opens a second Java compiler and file manager.
+public final class ReaGhidraBundleActivator implements BundleActivator {
+    @Override
+    public void start(BundleContext context) {
+    }
+
+    @Override
+    public void stop(BundleContext context) {
+    }
+}
@@ -0,0 +1,176 @@
+import ghidra.app.script.GhidraScript;
+import ghidra.program.database.SpecExtension;
+import ghidra.program.model.data.AbstractFloatDataType;
+import ghidra.program.model.data.DataType;
+import ghidra.program.model.data.DoubleDataType;
+import ghidra.program.model.data.FloatDataType;
+import ghidra.program.model.data.FunctionDefinition;
+import ghidra.program.model.data.IntegerDataType;
+import ghidra.program.model.data.ParameterDefinition;
+import ghidra.program.model.data.TypeDef;
+import ghidra.program.model.lang.CompilerSpec;
+import ghidra.program.model.lang.PrototypeModel;
+import ghidra.program.model.lang.Register;
+import ghidra.program.model.listing.Function;
+import ghidra.program.model.listing.Parameter;
+import ghidra.program.model.listing.Program;
+import ghidra.program.model.listing.VariableStorage;
+import ghidra.program.model.pcode.XmlEncode;
+import java.io.StringReader;
+import java.io.StringWriter;
+import java.util.ArrayList;
+import java.util.Iterator;
+import java.util.List;
+import java.util.regex.Pattern;
+import javax.xml.parsers.DocumentBuilderFactory;
+import javax.xml.transform.OutputKeys;
+import javax.xml.transform.TransformerFactory;
+import javax.xml.transform.dom.DOMSource;
+import javax.xml.transform.stream.StreamResult;
+import org.w3c.dom.Document;
+import org.w3c.dom.Element;
+import org.w3c.dom.Node;
+import org.w3c.dom.NodeList;
+import org.xml.sax.InputSource;
+
+/**
+ * Soft-float o32 passes and returns float/double values exactly where same-size integers go.
+ * Ghidra's MIPS32 compiler specifications model only hard-float o32, so every typed float
+ * prototype (functions, externals, thunks via their targets, and function-pointer definitions)
+ * would otherwise read $f12/$f14/$f0. Derive a program-specific prototype model from the
+ * program's own default o32 model with its floating-point entries and rules removed, install it
+ * as a specification extension, and assign it to every callable whose formal types use floats.
+ */
+public final class ReaGhidraMipsSoftFloat extends GhidraScript {
+  static final String REPORT_KEY = "REA MIPS soft-float storage";
+  static final String POLICY = "o32-soft-float-prototype-v1";
+  static final String PROTOTYPE = "__rea_soft_float";
+  private static final Pattern FPU_REGISTER = Pattern.compile("f\\d+(_\\d+)?");
+
+  @Override
+  public void run() throws Exception {
+    String processor = currentProgram.getLanguage().getProcessor().toString();
+    if (!processor.equals("MIPS") || currentProgram.getDefaultPointerSize() != 4)
+      throw new IllegalStateException(
+          "REA soft-float storage requires a 32-bit MIPS program, not " + processor);
+    CompilerSpec spec = currentProgram.getCompilerSpec();
+    PrototypeModel base = spec.getDefaultCallingConvention();
+    new SpecExtension(currentProgram)
+        .addReplaceCompilerSpecExtension(softFloatPrototype(base, spec), monitor);
+    PrototypeModel soft = currentProgram.getCompilerSpec().getCallingConvention(PROTOTYPE);
+    if (soft == null)
+      throw new IllegalStateException("REA soft-float prototype was not installed");
+    requireIntegerStorage(
+        soft.getStorageLocations(
+            currentProgram,
+            new DataType[] {
+              FloatDataType.dataType, IntegerDataType.dataType, DoubleDataType.dataType,
+              FloatDataType.dataType, DoubleDataType.dataType
+            },
+            false),
+        "soft-float prototype self-check");
+    int functions = 0;
+    List<Function> callables = new ArrayList<>();
+    currentProgram.getFunctionManager().getFunctions(true).forEach(callables::add);
+    currentProgram.getFunctionManager().getExternalFunctions().forEach(callables::add);
+    for (Function function : callables) {
+      // A thunk's signature is its target's; the target is in this list.
+      if (function.isThunk() || !usesFloat(function)) continue;
+      if (function.hasCustomVariableStorage()) function.setCustomVariableStorage(false);
+      function.setCallingConvention(PROTOTYPE);
+      List<VariableStorage> storage = new ArrayList<>();
+      storage.add(function.getReturn().getVariableStorage());
+      for (Parameter parameter : function.getParameters()) storage.add(parameter.getVariableStorage());
+      requireIntegerStorage(storage.toArray(VariableStorage[]::new), function.getName(true));
+      functions += 1;
+    }
+    int definitions = 0;
+    Iterator<DataType> types = currentProgram.getDataTypeManager().getAllDataTypes();
+    while (types.hasNext()) {
+      if (!(types.next() instanceof FunctionDefinition definition) || !usesFloat(definition))
+        continue;
+      definition.setCallingConvention(PROTOTYPE);
+      definitions += 1;
+    }
+    // Written last: any failure above leaves no report, and REA refuses the session.
+    currentProgram.getOptions(Program.PROGRAM_INFO).setString(
+        REPORT_KEY,
+        "{\"policy\":\"" + POLICY + "\",\"functions\":" + functions +
+            ",\"definitions\":" + definitions + "}");
+    println("REA MIPS soft-float prototype: functions=" + functions + " definitions=" + definitions);
+  }
+
+  /** The default o32 model without floating-point registers or float-consuming rules. */
+  private static String softFloatPrototype(PrototypeModel base, CompilerSpec spec) throws Exception {
+    XmlEncode encoder = new XmlEncode();
+    base.encode(encoder, spec.getPcodeInjectLibrary());
+    DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
+    factory.setFeature("http://apache.org/xml/features/disallow-doctype-decl", true);
+    Document document =
+        factory.newDocumentBuilder().parse(new InputSource(new StringReader(encoder.toString())));
+    Element root = document.getDocumentElement();
+    if (!root.getTagName().equals("prototype"))
+      throw new IllegalStateException("Unexpected default prototype encoding " + root.getTagName());
+    root.setAttribute("name", PROTOTYPE);
+    List<Node> removed = new ArrayList<>();
+    collect(root.getElementsByTagName("pentry"), "metatype", "float", null, removed);
+    collect(root.getElementsByTagName("consume"), "storage", "float", "rule", removed);
+    if (removed.isEmpty())
+      throw new IllegalStateException("Default prototype has no floating-point entries to remove");
+    for (Node node : removed) node.getParentNode().removeChild(node);
+    StringWriter xml = new StringWriter();
+    var transformer = TransformerFactory.newInstance().newTransformer();
+    transformer.setOutputProperty(OutputKeys.OMIT_XML_DECLARATION, "yes");
+    transformer.transform(new DOMSource(document), new StreamResult(xml));
+    return xml.toString();
+  }
+
+  /** Select matching elements, or their nearest `ancestor` element when one is named. */
+  private static void collect(
+      NodeList nodes, String attribute, String value, String ancestor, List<Node> selected) {
+    for (int index = 0; index < nodes.getLength(); index += 1) {
+      Element element = (Element) nodes.item(index);
+      if (!value.equals(element.getAttribute(attribute))) continue;
+      Node target = element;
+      while (ancestor != null && target != null && !ancestor.equals(target.getNodeName()))
+        target = target.getParentNode();
+      if (target != null && !selected.contains(target)) selected.add(target);
+    }
+  }
+
+  private static void requireIntegerStorage(VariableStorage[] storage, String owner) {
+    for (VariableStorage item : storage) {
+      if (item.isBadStorage() || item.isUnassignedStorage())
+        throw new IllegalStateException("REA soft-float storage unresolved for " + owner);
+      List<Register> registers = item.getRegisters();
+      if (registers == null) continue; // stack-only storage
+      for (Register register : registers) {
+        if (FPU_REGISTER.matcher(register.getName()).matches())
+          throw new IllegalStateException(
+              "REA soft-float storage still uses " + register.getName() + " for " + owner);
+      }
+    }
+  }
+
+  private static boolean usesFloat(Function function) {
+    if (isFloat(function.getReturn().getFormalDataType())) return true;
+    for (Parameter parameter : function.getParameters()) {
+      if (isFloat(parameter.getFormalDataType())) return true;
+    }
+    return false;
+  }
+
+  private static boolean usesFloat(FunctionDefinition definition) {
+    if (isFloat(definition.getReturnType())) return true;
+    for (ParameterDefinition argument : definition.getArguments()) {
+      if (isFloat(argument.getDataType())) return true;
+    }
+    return false;
+  }
+
+  private static boolean isFloat(DataType type) {
+    DataType current = type;
+    while (current instanceof TypeDef typeDef) current = typeDef.getBaseDataType();
+    return current instanceof AbstractFloatDataType;
+  }
+}
@@ -0,0 +1,49 @@
+import ghidra.app.script.GhidraScript;
+import ghidra.program.model.address.Address;
+import ghidra.program.model.listing.Program;
+import ghidra.program.model.mem.MemoryBlock;
+import ghidra.program.model.mem.MemoryBlockSourceInfo;
+
+/** Seed the caller-supplied raw-binary entry before default auto-analysis. */
+public final class ReaGhidraPrepareRaw extends GhidraScript {
+    @Override
+    public void run() throws Exception {
+        String[] arguments = getScriptArgs();
+        if (arguments.length != 2) {
+            throw new IllegalArgumentException(
+                "Raw binary preparation requires the committed language and entry address"
+            );
+        }
+        String expectedLanguage = arguments[0];
+        String entryHex = arguments[1];
+        if (!currentProgram.getLanguageID().getIdAsString().equals(expectedLanguage) ||
+            !currentProgram.getExecutableFormat().equals("Raw Binary")) {
+            throw new IllegalArgumentException(
+                "Raw binary preparation requires the committed language and BinaryLoader import"
+            );
+        }
+        Address entry = currentProgram.getAddressFactory().getDefaultAddressSpace().getAddress(entryHex);
+        MemoryBlock block = entry == null ? null : currentProgram.getMemory().getBlock(entry);
+        boolean fileStart = false;
+        if (block != null) {
+            // Processor-defined zero page/stack blocks can precede the imported file.
+            for (MemoryBlockSourceInfo source : block.getSourceInfos()) {
+                if (source.getFileBytes().isPresent() && source.getFileBytesOffset() == 0 &&
+                    entry.equals(source.getMinAddress())) {
+                    fileStart = true;
+                    break;
+                }
+            }
+        }
+        if (!fileStart) {
+            throw new IllegalArgumentException(
+                "Raw binary import must map file offset 0 at the caller-supplied entry"
+            );
+        }
+        addEntryPoint(entry);
+        if (!disassemble(entry) || createFunction(entry, "entry") == null) {
+            throw new IllegalStateException("Raw binary entry cannot be decoded as a function");
+        }
+        currentProgram.getOptions(Program.PROGRAM_INFO).setBoolean("REA raw binary prepared", true);
+    }
+}
@@ -644,7 +644,7 @@ def _analyze_function(document, params):
            "end": _hex(_basic_block_end(procedure, block)),
            "successors": sorted(set(successors), key=lambda value: int(value, 16)),
        })
-    pseudo = _pseudocode(document, procedure) or ""
+    pseudo = _pseudocode(document, procedure)
    assembly_lines = _assembly(procedure).splitlines()
    callers = sorted((_procedure_identity(item) for item in procedure.getAllCallerProcedures()), key=lambda item: int(item["address"], 16))
    callees = sorted((_procedure_identity(item) for item in procedure.getAllCalleeProcedures()), key=lambda item: int(item["address"], 16))
@@ -706,7 +706,7 @@ def _analyze_function(document, params):
            "signature": procedure.signatureString(),
            "locals": _procedure_locals(procedure),
        },
-        "pseudocode": pseudo,
+        "pseudocode": pseudo or "",
        "assembly": assembly_lines,
        "comments": comments,
        "callers": callers, "callees": callees,
@@ -718,6 +718,12 @@ def _analyze_function(document, params):
        "basic_blocks": blocks,
        "native_api": None,
        "native_value_flow": None,
+        "facet_availability": {
+            "pseudocode": {"available": True} if pseudo is not None else _unavailable("Hopper did not supply pseudocode."),
+            **{facet: {"available": True} for facet in (
+                "assembly", "comments", "callers", "callees", "references", "strings_names", "cfg"
+            )},
+        },
        "limitations": [
            "Native CallReference classifications are observed where available; detailed reference flags remain unknown.",
            "Hopper's public Python API does not expose equivalent external or thunk classification in this dossier.",
@@ -1016,73 +1022,87 @@ def _dispatch(method, params):
 def _serve_connection(connection):
    """Serve one capability-authenticated NDJSON connection."""
    file = connection.makefile("rwb")
-    while True:
-        line = file.readline()
-        if not line:
-            break
-        request_id = None
-        authenticated = False
-        should_stop = False
-        try:
+    should_stop = False
+    try:
+        while True:
+            line = file.readline()
+            if not line:
+                break
+            request_id = None
+            authenticated = False
+            should_stop = False
            try:
-                request = json.loads(line.decode("utf-8"))
-            except (UnicodeDecodeError, json.JSONDecodeError) as error:
-                raise InvalidRequestError("Invalid bridge request JSON") from error
-            if not isinstance(request, dict) or set(request) != {"id", "token", "method", "params"}:
-                raise InvalidRequestError("Invalid bridge request shape")
-            request_id = request["id"]
-            if isinstance(request_id, bool) or not isinstance(request_id, int) or request_id < 0:
-                raise InvalidRequestError("Invalid bridge request id")
-            if (
-                not isinstance(request["method"], str) or not request["method"]
-                or not isinstance(request["params"], dict)
-            ):
-                raise InvalidRequestError("Invalid bridge method or parameters")
-            if (
-                not isinstance(request["token"], str) or not request["token"].isascii()
-                or not hmac.compare_digest(request["token"], REA_TOKEN)
-            ):
-                raise PermissionError("Invalid bridge capability")
-            authenticated = True
-            _write_message(file, {"id": request_id, "event": {
-                "type": "progress",
-                "phase": "hopper_bridge",
-                "completed": 0,
-                "total": 1,
-                "message": "Hopper bridge started request",
-            }})
-            result = _dispatch(request["method"], request["params"])
-            should_stop = request["method"] == "shutdown_document" or (
-                request["method"] == "shutdown" and not result.get("cleanup_required", False)
-            )
-            safe_result = _json_safe(result)
-            _write_message(file, {"id": request_id, "event": {
-                "type": "progress",
-                "phase": "hopper_bridge",
-                "completed": 1,
-                "total": 1,
-                "message": "Hopper bridge completed request",
-                "terminal": True,
-            }})
-            response = {"id": request_id, "result": safe_result}
-        except Exception as error:
-            diagnostic = _safe_diagnostic(error)
-            if authenticated:
+                try:
+                    request = json.loads(line.decode("utf-8"))
+                except (UnicodeDecodeError, json.JSONDecodeError) as error:
+                    raise InvalidRequestError("Invalid bridge request JSON") from error
+                if not isinstance(request, dict) or set(request) != {"id", "token", "method", "params"}:
+                    raise InvalidRequestError("Invalid bridge request shape")
+                request_id = request["id"]
+                if isinstance(request_id, bool) or not isinstance(request_id, int) or request_id < 0:
+                    raise InvalidRequestError("Invalid bridge request id")
+                if (
+                    not isinstance(request["method"], str) or not request["method"]
+                    or not isinstance(request["params"], dict)
+                ):
+                    raise InvalidRequestError("Invalid bridge method or parameters")
+                if (
+                    not isinstance(request["token"], str) or not request["token"].isascii()
+                    or not hmac.compare_digest(request["token"], REA_TOKEN)
+                ):
+                    raise PermissionError("Invalid bridge capability")
+                authenticated = True
                _write_message(file, {"id": request_id, "event": {
-                    "type": "diagnostic",
-                    "error": diagnostic,
+                    "type": "progress",
+                    "phase": "hopper_bridge",
+                    "completed": 0,
+                    "total": 1,
+                    "message": "Hopper bridge started request",
                }})
-            response_id = request_id if (
-                isinstance(request_id, int)
-                and not isinstance(request_id, bool)
-                and request_id >= 0
-            ) else 0
-            response = {"id": response_id, "error": diagnostic}
-        _write_message(file, response)
-        if should_stop:
-            break
-    file.close()
-    connection.close()
+                result = _dispatch(request["method"], request["params"])
+                should_stop = request["method"] in ("shutdown", "shutdown_document") and (
+                    result.get("shutdown") is True and (
+                        (result.get("document_closed") is True and result.get("analysis_stopped") is True)
+                        or (result.get("document_closed") is False and result.get("document_retained") is True)
+                    )
+                )
+                safe_result = _json_safe(result)
+                _write_message(file, {"id": request_id, "event": {
+                    "type": "progress",
+                    "phase": "hopper_bridge",
+                    "completed": 1,
+                    "total": 1,
+                    "message": "Hopper bridge completed request",
+                    "terminal": True,
+                }})
+                response = {"id": request_id, "result": safe_result}
+            except Exception as error:
+                diagnostic = _safe_diagnostic(error)
+                if authenticated:
+                    _write_message(file, {"id": request_id, "event": {
+                        "type": "diagnostic",
+                        "error": diagnostic,
+                    }})
+                response_id = request_id if (
+                    isinstance(request_id, int)
+                    and not isinstance(request_id, bool)
+                    and request_id >= 0
+                ) else 0
+                response = {"id": response_id, "error": diagnostic}
+            _write_message(file, response)
+            if should_stop:
+                break
+    except (ConnectionError, OSError):
+        # A lost client cannot confirm document closure or revoke this run.
+        pass
+    finally:
+        try:
+            file.close()
+        except OSError:
+            # Closing a failed writer can repeat the same disconnected-pipe error.
+            pass
+        connection.close()
+    return should_stop


 def _write_message(file, message):
@@ -1124,8 +1144,17 @@ def _run():
    os.chmod(REA_SOCKET, 0o600)
    server.listen(1)
    try:
-        connection, _ = server.accept()
-        _serve_connection(connection)
+        while True:
+            try:
+                connection, _ = server.accept()
+            except socket.timeout:
+                break
+            connection.settimeout(None)
+            if _serve_connection(connection):
+                break
+            # Match REA's cleanup handshake budget without pinning Hopper's
+            # Python thread indefinitely after a client disappears.
+            server.settimeout(REA_RECONNECT_TIMEOUT_SECONDS)
    finally:
        server.close()
        if os.path.exists(REA_SOCKET):
@@ -93,7 +93,7 @@ def project_record(state, secrets):
            safe_bytes, credential_url = transport_value(value, path, transport_url)
            content_base64 = base64.b64encode(value).decode("ascii")
            digest = hashlib.sha256(value).hexdigest()
-            hidden = any(secret in value or secret in safe_bytes for secret in byte_secrets) or any(secret in content_base64 or secret in digest for secret in secrets)
+            hidden = any(secret in value or secret in safe_bytes for secret in byte_secrets) or any(secret in content_base64 or secret in digest or secret in str(len(value)) for secret in secrets)
            safe_text = None
            try:
                safe_text = safe_bytes.decode("utf-8", errors="strict")
@@ -112,6 +112,11 @@ enum RunTokenReadError: Error {
  case ambiguousEnvironmentBoundary
 }

+struct ParsedRunToken {
+  let runId: String?
+  let environmentEntryCount: Int
+}
+
@_silgen_name("csops")
 private func csops(
  _ pid: Int32,
@@ -122,19 +127,67 @@ private func csops(

 private let csOpsStatus: UInt32 = 0
 private let csPlatformBinary: UInt32 = 0x0400_0000
+private let csRestrict: UInt32 = 0x0000_0800

-/**
- * Whether the target image is an Apple platform binary. Current macOS omits
- * the environment of platform binaries from KERN_PROCARGS2 for every caller,
- * so this distinguishes a policy-withheld environment from a process whose
- * environment is genuinely absent or transiently unreadable.
- */
-func isPlatformBinary(pid: Int32) -> Bool {
+/** Read the target's current code-signing flags when csops permits inspection. */
+private func codeSigningFlags(pid: Int32) -> UInt32? {
  var flags: UInt32 = 0
  let result = withUnsafeMutablePointer(to: &flags) { pointer in
    csops(pid, csOpsStatus, UnsafeMutableRawPointer(pointer), MemoryLayout<UInt32>.size)
  }
-  return result == 0 && (flags & csPlatformBinary) != 0
+  return result == 0 ? flags : nil
+}
+
+/** Derive one process-token observation from parser facts and native flags. */
+func processRunTokenObservation(
+  pid: Int32,
+  parse: Result<ParsedRunToken, RunTokenReadError>,
+  codeSigningFlags: () -> UInt32?
+) -> RunTokenObservation {
+  switch parse {
+  case .success(let parsed):
+    if let runId = parsed.runId {
+      return RunTokenObservation(pid: pid, state: "readable", run_id: runId, reason: nil)
+    }
+    // A non-empty environment without REA's token proves this candidate is
+    // unrelated. CS_RESTRICT cannot turn that readable fact into uncertainty.
+    if parsed.environmentEntryCount > 0 {
+      return RunTokenObservation(pid: pid, state: "readable", run_id: nil, reason: nil)
+    }
+    let flags = codeSigningFlags()
+    if let flags, (flags & csPlatformBinary) != 0 {
+      return unavailable(pid: pid, reason: "platform_binary_environment_withheld")
+    }
+    if let flags, (flags & csRestrict) != 0 {
+      return unavailable(pid: pid, reason: "restricted_process_environment_withheld")
+    }
+    return RunTokenObservation(pid: pid, state: "readable", run_id: nil, reason: nil)
+  case .failure(let failure):
+    switch failure {
+    case .environmentUnavailable:
+      let flags = codeSigningFlags()
+      if let flags, (flags & csPlatformBinary) != 0 {
+        return unavailable(pid: pid, reason: "platform_binary_environment_withheld")
+      }
+      if let flags, (flags & csRestrict) != 0 {
+        return unavailable(pid: pid, reason: "restricted_process_environment_withheld")
+      }
+      return unavailable(pid: pid, reason: "environment_unavailable")
+    case .ambiguousEnvironmentBoundary:
+      if let flags = codeSigningFlags(), (flags & csRestrict) != 0 {
+        return unavailable(pid: pid, reason: "restricted_process_environment_withheld")
+      }
+      return unavailable(pid: pid, reason: "ambiguous_environment_boundary")
+    case .duplicateToken:
+      return unavailable(pid: pid, reason: "duplicate_run_token")
+    case .invalidTokenEncoding:
+      return unavailable(pid: pid, reason: "invalid_run_token_encoding")
+    case .appleVectorUnavailable:
+      return unavailable(pid: pid, reason: "apple_vector_unavailable")
+    case .invalidBuffer:
+      return unavailable(pid: pid, reason: "malformed_procargs")
+    }
+  }
 }

 private func readRunToken(pid: Int32, pointerSize: Int) -> RunTokenObservation {
@@ -151,33 +204,19 @@ private func readRunToken(pid: Int32, pointerSize: Int) -> RunTokenObservation {
    return unavailable(pid: pid, reason: "sysctl_failed_\(errno)")
  }
  bytes = Array(bytes.prefix(size))
+  let parse: Result<ParsedRunToken, RunTokenReadError>
  do {
-    let runId = try token(in: bytes, pointerSize: pointerSize)
-    // A complete empty environment is readable for an ordinary image. Apple
-    // platform images can expose the same layout after the kernel withholds
-    // their environment, so they remain explicitly unverified without a token.
-    if runId == nil && isPlatformBinary(pid: pid) {
-      return unavailable(pid: pid, reason: "platform_binary_environment_withheld")
-    }
-    return RunTokenObservation(pid: pid, state: "readable", run_id: runId, reason: nil)
-  } catch RunTokenReadError.duplicateToken {
-    return unavailable(pid: pid, reason: "duplicate_run_token")
-  } catch RunTokenReadError.invalidTokenEncoding {
-    return unavailable(pid: pid, reason: "invalid_run_token_encoding")
-  } catch RunTokenReadError.environmentUnavailable {
-    return unavailable(
-      pid: pid,
-      reason: isPlatformBinary(pid: pid)
-        ? "platform_binary_environment_withheld"
-        : "environment_unavailable"
-    )
-  } catch RunTokenReadError.appleVectorUnavailable {
-    return unavailable(pid: pid, reason: "apple_vector_unavailable")
-  } catch RunTokenReadError.ambiguousEnvironmentBoundary {
-    return unavailable(pid: pid, reason: "ambiguous_environment_boundary")
+    parse = .success(try parseProcessEnvironment(in: bytes, pointerSize: pointerSize))
+  } catch let failure as RunTokenReadError {
+    parse = .failure(failure)
  } catch {
-    return unavailable(pid: pid, reason: "malformed_procargs")
+    parse = .failure(.invalidBuffer)
  }
+  return processRunTokenObservation(
+    pid: pid,
+    parse: parse,
+    codeSigningFlags: { codeSigningFlags(pid: pid) }
+  )
 }

 func readRunTokens(pids: [Int32]) -> [RunTokenObservation] {
@@ -201,7 +240,10 @@ func readRunTokens(pids: [Int32]) -> [RunTokenObservation] {
  }
 }

-func token(in bytes: [UInt8], pointerSize: Int = MemoryLayout<UnsafeRawPointer>.size) throws -> String? {
+func parseProcessEnvironment(
+  in bytes: [UInt8],
+  pointerSize: Int = MemoryLayout<UnsafeRawPointer>.size
+) throws -> ParsedRunToken {
  guard bytes.count >= MemoryLayout<Int32>.size else { throw RunTokenReadError.invalidBuffer }
  let argc = bytes.withUnsafeBytes { $0.loadUnaligned(as: Int32.self).littleEndian }
  guard argc >= 0, Int(argc) <= bytes.count else { throw RunTokenReadError.invalidBuffer }
@@ -277,7 +319,7 @@ func token(in bytes: [UInt8], pointerSize: Int = MemoryLayout<UnsafeRawPointer>.
      found = decoded
    }
  }
-  return found
+  return ParsedRunToken(runId: found, environmentEntryCount: environmentEntries.count)
 }

 private func readString(
@@ -206,6 +206,11 @@ def inspect_elf(path, cache):
                            "visibility": raw.st_other.visibility, "section_index": raw.st_shndx,
                        })
                if h.sh_type in ("SHT_REL", "SHT_RELA"):
+                    if h.sh_size % h.sh_entsize != 0:
+                        raise LayoutFailure(
+                            "format",
+                            f"ELF relocation section {index} size is not a multiple of its {h.sh_entsize}-byte entries.",
+                        )
                    if image.header.e_type == "ET_REL" and h.sh_info >= len(all_sections):
                        raise LayoutFailure("format", f"ELF relocation section {index} references missing target section {h.sh_info}.")
                    symbol_table = None
@@ -11,6 +11,12 @@ import sys

 PROFILE = "pwntools@4.15.0;pyelftools@0.33;unicorn@2.1.2"

+# Linux ET_CORE machine identifiers whose PRSTATUS layout is interpreted with
+# unchanged pwntools 4.15.0 ctypes. Both share pr_cursig@12, pr_pid@32 and
+# pr_reg@112; only the register window differs: amd64 is 27 x 8 bytes for a
+# 336-byte descriptor, AArch64 is 34 x 8 bytes for a 392-byte descriptor.
+CORE_MACHINES = (62, 183)
+

 class CoreFailure(Exception):
    def __init__(self, reason, message):
@@ -51,12 +57,27 @@ def inspect_core(path, cache):
    # Other declared OSABIs must never select Linux note ctypes.
    if content[7] not in (0, 3) or content[8] != 0:
        raise CoreFailure("unsupported-target", "Initial recorded-core profile requires System V or GNU/Linux OSABI with ABI version zero; selected OSABI=" + str(content[7]) + ", ABI version=" + str(content[8]) + ".")
-    if int.from_bytes(content[16:18], "little") != 4 or int.from_bytes(content[18:20], "little") != 62:
-        raise CoreFailure("unsupported-target", "Initial recorded-core profile requires x86-64 ET_CORE.")
+    machine = int.from_bytes(content[18:20], "little")
+    if int.from_bytes(content[16:18], "little") != 4 or machine not in CORE_MACHINES:
+        raise CoreFailure("unsupported-target", "Initial recorded-core profile requires ET_CORE for x86-64 or AArch64; selected ET_TYPE=" + str(int.from_bytes(content[16:18], "little")) + ", e_machine=" + str(machine) + ".")
    from elftools.common.exceptions import ELFError
    from elftools.common.utils import struct_parse
    from elftools.elf.elffile import ELFFile
-    from pwnlib.elf.datatypes import elf_prstatus_amd64, elf_siginfo_64, user_regs_struct_amd64
+    from pwnlib.elf.datatypes import (
+        elf_prstatus_aarch64,
+        elf_prstatus_amd64,
+        elf_siginfo_64,
+        user_regs_struct_aarch64,
+        user_regs_struct_amd64,
+    )
+    if machine == 62:
+        profile_label = "x86-64"
+        architecture = "x86_64-little-endian"
+        prstatus_type, registers_type = elf_prstatus_amd64, user_regs_struct_amd64
+    else:
+        profile_label = "AArch64"
+        architecture = "aarch64-little-endian"
+        prstatus_type, registers_type = elf_prstatus_aarch64, user_regs_struct_aarch64
    try:
        image = ELFFile(io.BytesIO(content))
        header = image.header
@@ -107,17 +128,17 @@ def inspect_core(path, cache):
                if owner != b"CORE\0":
                    continue
                if note.n_type == "NT_PRSTATUS":
-                    if len(descriptor) < ctypes.sizeof(elf_prstatus_amd64):
-                        raise CoreFailure("format", "Linux amd64 NT_PRSTATUS descriptor is truncated.")
-                    status = elf_prstatus_amd64.from_buffer_copy(descriptor)
+                    if len(descriptor) < ctypes.sizeof(prstatus_type):
+                        raise CoreFailure("format", "Linux " + profile_label + " NT_PRSTATUS descriptor is truncated.")
+                    status = prstatus_type.from_buffer_copy(descriptor)
                    registers = []
-                    for name, kind in user_regs_struct_amd64._fields_:
-                        relative = elf_prstatus_amd64.pr_reg.offset + getattr(user_regs_struct_amd64, name).offset
+                    for name, kind in registers_type._fields_:
+                        relative = prstatus_type.pr_reg.offset + getattr(registers_type, name).offset
                        registers.append({"name": name, "value": hex(getattr(status.pr_reg, name)), "location": location(descriptor_start + relative, ctypes.sizeof(kind), end)})
                    threads.append({"note_index": note_index, "historical_pid": ctypes.c_int32(status.pr_pid).value, "recorded_current_signal": status.pr_cursig, "registers": registers})
                elif note.n_type == "NT_SIGINFO":
                    if len(descriptor) < ctypes.sizeof(elf_siginfo_64):
-                        raise CoreFailure("format", "Linux amd64 NT_SIGINFO descriptor is truncated.")
+                        raise CoreFailure("format", "Linux " + profile_label + " NT_SIGINFO descriptor is truncated.")
                    signal = elf_siginfo_64.from_buffer_copy(descriptor)
                    number = ctypes.c_int32(signal.si_signo).value
                    code = ctypes.c_int32(signal.si_code).value
@@ -125,7 +146,7 @@ def inspect_core(path, cache):
                    signals.append({"note_index": note_index, "number": number, "code": code, "errno": ctypes.c_int32(signal.si_errno).value,
                                    "fault_address": hex(signal.sigfault_addr) if known else None,
                                    "fault_address_meaning": "recorded-sigsegv-address" if known else "unknown", "thread_association": "unknown"})
-        return {"format": "elf-core", "architecture": "x86_64-little-endian", "target_execution": "not-performed", "live_process_identity": "unknown",
+        return {"format": "elf-core", "architecture": architecture, "target_execution": "not-performed", "live_process_identity": "unknown",
                "segments": segments, "notes": notes, "note_padding": padding, "threads": threads, "signals": signals,
                "note_interpretation_completeness": "unknown",
                "limitations": ["The initial Linux note profile accepts System V/unspecified or GNU/Linux OSABI with ABI version zero; an unspecified OSABI alone does not prove operating-system origin.",
@@ -6,6 +6,7 @@ export default defineConfig({
    "Reverse engineer binaries, applications, and runtime behavior with one CLI and MCP server.",
  lang: "en-US",
  base: "/rea/",
+  srcExclude: ["readme/**"],
  themeConfig: {
    nav: [
      { text: "Get started", link: "/installation" },
@@ -17,6 +18,7 @@ export default defineConfig({
        text: "Getting started",
        items: [
          { text: "Installation and setup", link: "/installation" },
+          { text: "Client configuration", link: "/client-configuration" },
          { text: "CLI and Evidence", link: "/cli" },
          { text: "Agent prompts", link: "/mcp-prompts" },
          { text: "Roadmap", link: "/roadmap" },
@@ -26,6 +28,8 @@ export default defineConfig({
        text: "Investigation guides",
        items: [
          { text: "Native binaries", link: "/native-investigation" },
+          { text: "Go build metadata", link: "/go-binaries" },
+          { text: "Rust dependencies", link: "/rust-binaries" },
          {
            text: "JavaScript and Electron",
            link: "/javascript-artifact-reconstruction",
@@ -35,16 +39,19 @@ export default defineConfig({
            link: "/javascript-application-workflows",
          },
          { text: "Browser observation", link: "/browser-observation" },
+          {
+            text: "Web network captures",
+            link: "/web-network-captures",
+          },
          { text: "Electron observation", link: "/electron-observation" },
          { text: "Process capture", link: "/process-capture" },
          { text: ".NET assemblies", link: "/managed-code-analysis" },
          { text: "Android applications", link: "/android-analysis" },
          { text: "Apple applications", link: "/apple-application-analysis" },
          { text: "Firmware", link: "/firmware-analysis" },
-          {
-            text: "Reconstruction readiness",
-            link: "/reconstruction-readiness",
-          },
+          { text: "SQLite databases", link: "/sqlite-databases" },
+          { text: "Packet captures", link: "/packet-captures" },
+          { text: "Signal recordings", link: "/signal-recordings" },
        ],
      },
      {
@@ -54,7 +61,9 @@ export default defineConfig({
          { text: "IDA", link: "/ida-provider" },
          { text: "Ghidra on Windows", link: "/windows-ghidra-p0" },
          { text: "Ghidra DOS analysis", link: "/ghidra-dos" },
+          { text: "Ghidra raw binary", link: "/raw-binary" },
          { text: "Ghidra NativeAOT", link: "/ghidra-nativeaot" },
+          { text: "Ghidra PowerPC", link: "/ghidra-ppc" },
          { text: "Ghidra semantics", link: "/ghidra-provider" },
        ],
      },
@@ -65,6 +74,11 @@ export default defineConfig({
          { text: "MCP contracts", link: "/mcp-contracts" },
          { text: "Tool design", link: "/tool-design" },
          { text: "Testing", link: "/testing" },
+          { text: "Client verification", link: "/client-verification" },
+          {
+            text: "Reconstruction readiness",
+            link: "/reconstruction-readiness",
+          },
          { text: "Releasing", link: "/releasing" },
        ],
      },
@@ -88,7 +88,7 @@ existing analysis families directly, for example
 | `list_adb_packages`     | read               | `pm list packages -f` with scope; base APK device paths                                                                                                                   |
 | `resolve_adb_packages`  | read               | Case-insensitive package-name substring search with exact-match flag                                                                                                      |
 | `inspect_adb_package`   | read               | `dumpsys package` projection: versions, install times, installer, flags, permission count                                                                                 |
-| `list_adb_directory`    | read               | `ls -la`: kinds, permissions, owner, group, sizes, symlink targets                                                                                                        |
+| `list_adb_directory`    | read               | NUL-framed immediate children, including hidden entries: kinds, permissions, owner, group, sizes, symlink targets                                                         |
 | `pull_adb_package`      | read               | Complete split APK set from `pm path`, each file digested                                                                                                                 |
 | `pull_adb_file`         | read               | One absolute device path pulled and digested                                                                                                                              |
 | `capture_adb_screen`    | read               | One `screencap -p` frame (exec-out, no device-side file) with dimensions                                                                                                  |
@@ -102,13 +102,22 @@ existing analysis families directly, for example

 ## Interpret the result

+- **Directory names retain whitespace and newlines.** Symlink targets do too,
+  including trailing newlines. Records containing invalid UTF-8 remain as
+  labelled hex in `unparsed_lines` and make directory coverage partial.
 - **Inferences carry their basis.** Device `kind` reports
  `kind_basis` (serial prefix or USB transport); APK `role` reports
  `role_basis: file_name`; `emulator_observed` is the build's own
  `ro.kernel.qemu` value, `null` when absent.
 - **Device paths are not host paths.** `base_apk_device_path` values live on
  the device filesystem. Only pulls and captures produce host-local paths,
-  each paired with byte size and SHA-256 digest.
+  each paired with byte size and SHA-256 digest. A pulled file's host name is
+  the device path's final POSIX segment. REA preserves the original device
+  path and filename; a missing final segment or `.`/`..` is refused before
+  output creation. Windows hosts follow [Windows filename rules](https://learn.microsoft.com/en-us/windows/win32/fileio/naming-a-file),
+  while POSIX hosts retain valid POSIX filenames such as names with spaces or
+  colons. A package pull checks reported names against the destination
+  filesystem before pulling, and refuses names that would share a local path.
 - **Splits are part of the answer.** `pm path` enumerates the base APK and
  every installed split; pulls cover the whole set. System packages whose
  APKs keep other names report the `unknown` role rather than guessing.
@@ -127,8 +136,8 @@ existing analysis families directly, for example
 ## Resource and lifecycle limits

 - Caller values reach adb only as positional arguments in fixed vectors;
-  no tool accepts an arbitrary shell command line, and UI interaction
-  (`input`, `am`, `monkey`) is deliberately not exposed.
+  no tool accepts an arbitrary shell command line, `input`, or `monkey`.
+  The app/activity lifecycle tools use `am start -W` and `am force-stop`.
 - Per-invocation budgets: client/devices 10-20 s, shell reads 20-60 s with
  64 KiB-8 MiB capture caps (logcat 4 MiB, screen 32 MiB), transfers and
  bugreports 600 s. Exceeded budgets report `resource_constraint`; stopped
@@ -139,6 +148,11 @@ existing analysis families directly, for example
 - Cancellation during a package pull preserves caller-selected output. The
  error includes partial observations identifying completed artifacts and the
  interrupted local path; remaining APKs were not attempted.
+- Bugreports collect into isolated staging and publish with exclusive creation;
+  filename collisions select the next numeric suffix. If publication fails, the
+  completed archive remains in staging. The error's partial observation gives
+  its recoverable `local_path`, the original `adb_reported_path`, and an unknown
+  digest status. Later operations and shutdown preserve this retained evidence.
 - Failed package acquisition rollback retains its directory identity and retries
  cleanup before later work and during shutdown. Cleanup refuses a replacement
  directory. Terminal shutdown stops admission and drains active operations.
@@ -9,13 +9,64 @@ This family is present on repository main and in npm 4.1.0. Check the
 [released package boundary](installation.md#released-package-and-main) and the
 connected server's tool list before selecting it.

+## Optional package delivery
+
+The MCP `inspect_android_package` request keeps complete output when `detail` is
+omitted or `"complete"`. Select `detail: "summary"` explicitly to receive all
+package, SDK and permission declarations, engine identity and class/resource
+counts without decoded manifest text or duplicate raw provider replies. A
+retaining session keeps the complete original Evidence and returns a distinct
+linked `inspect_analysis_view` summary. This changes delivery, not JADX analysis.
+
+Use the returned `normalized_result.parent_evidence_id` with
+`inspect_analysis_view`, source kind `retained-evidence`, and
+`view: {kind: "android", facet: "manifest", offset: 0, limit: 4096}` to read the
+exact decoded manifest without reopening the APK. Offsets and limits count UTF-16
+code units; follow `coverage.next_offset`. The page preserves `source_status` and
+`reported_total_bytes`: exhausting retained text does not make a partial decoded
+manifest complete. Complete source Evidence remains exportable under the existing
+session lifetime. Servers without retention reject summary before starting JADX.
+One-shot CLI output remains complete. Older servers require the existing complete
+request after schema discovery. APK authenticity and runtime behavior remain
+unverified.
+
 ## Supply tools

 Use a **full JDK 17 or newer** (including its compiler module) and
-**jadx-headless-mcp 0.7.1**. REA runs its packaged Java metadata bridge in
-source-file mode against the unmodified engine JAR. REA does not install Java,
-an Android SDK, JADX, a device service or an emulator. Set the path in the shell
-or MCP server environment:
+[**jadx-headless-mcp 0.7.1**](https://github.com/1013503897/jadx-headless-mcp/releases/download/v0.7.1/jadx-headless-mcp-0.7.1-all.jar).
+Verify the downloaded JAR's SHA-256 against
+[REA's pinned release metadata](https://github.com/morluto/rea/blob/main/src/android/JadxRelease.ts)
+before using it. The following download and verification commands work for npm
+package consumers as well as repository checkouts; the repository-only
+`npm run fixtures:android` helper is not included in the npm package.
+
+On macOS or Linux, choose a destination and verify the pinned digest:
+
+```sh
+mkdir -p "$HOME/.local/share/rea/android" &&
+cd "$HOME/.local/share/rea/android" &&
+curl -fL -o jadx-headless-mcp-0.7.1-all.jar https://github.com/1013503897/jadx-headless-mcp/releases/download/v0.7.1/jadx-headless-mcp-0.7.1-all.jar &&
+if command -v sha256sum >/dev/null; then
+  echo '6e5eacf500b64292bfb73c49797c1958f6ee44646e43e868039ae7feb573ff75  jadx-headless-mcp-0.7.1-all.jar' | sha256sum -c -
+else
+  echo '6e5eacf500b64292bfb73c49797c1958f6ee44646e43e868039ae7feb573ff75  jadx-headless-mcp-0.7.1-all.jar' | shasum -a 256 -c -
+fi &&
+export REA_JADX_MCP_JAR="$PWD/jadx-headless-mcp-0.7.1-all.jar"
+```
+
+On Windows, run the equivalent in PowerShell:
+
+```powershell
+$jar = Join-Path $HOME '.local\share\rea\android\jadx-headless-mcp-0.7.1-all.jar'
+New-Item -ItemType Directory -Force (Split-Path $jar) | Out-Null
+Invoke-WebRequest -Uri 'https://github.com/1013503897/jadx-headless-mcp/releases/download/v0.7.1/jadx-headless-mcp-0.7.1-all.jar' -OutFile $jar
+if ((Get-FileHash -Algorithm SHA256 $jar).Hash.ToLowerInvariant() -ne '6e5eacf500b64292bfb73c49797c1958f6ee44646e43e868039ae7feb573ff75') { throw 'JADX JAR SHA-256 does not match REA pinned metadata' }
+$env:REA_JADX_MCP_JAR = $jar
+```
+
+REA runs its packaged Java metadata bridge in source-file mode against the
+unmodified engine JAR. REA does not install Java, an Android SDK, JADX, a device
+service or an emulator. Set the path in the shell or MCP server environment:

 ```sh
 export REA_JADX_MCP_JAR=/absolute/path/jadx-headless-mcp-0.7.1-all.jar
@@ -65,6 +116,7 @@ rea inspect-android-method /path/Example.apk example.MainActivity onCreate
 rea inspect-android-method /path/Example.apk example.MainActivity select --overload-index 1
 rea trace-android-references /path/Example.apk example.MainActivity
 rea trace-android-references /path/Example.apk example.MainActivity --method-name onCreate
+rea trace-android-references /path/Example.apk example.MainActivity --method-name select --overload-index 1
 ```

 Equivalent MCP methods are `inspect_android_package`, `search_android_classes`,
@@ -85,15 +137,22 @@ Every request includes `path`. Class/method operations use `class_name`,
 ```

 Use `inspect_android_class` to choose an overload. REA rejects an ambiguous
-method request rather than silently selecting the first match. The engine's
-method reference operation cannot select overloads, so REA reports that boundary
-as unsupported when multiple same-name methods exist.
+method request rather than silently selecting the first match. Method reference
+queries accept the same `overload_index`; the bridge resolves the selected
+method node and the result reports its index and displayed signature. These
+indices are tied to the current artifact and engine, not stable across builds.

 The engine's smali fallback also joins same-name overloads; REA rejects that
 fallback for an overloaded method instead of attributing all bodies to one
 selected overload. Native/abstract methods with no decompiled body report
 `body_status: "not_available"` without inferring which implementation flag caused it.

+Static initializers (`<clinit>`) return smali with the
+`static_initializer_disassembly` marker. Java decompilation can move their
+instructions into field declarations, leaving an empty or incomplete method
+body. The selected method's disassembly preserves those instructions and retains
+the same source byte limit and explicit truncation reporting.
+
 ## Interpret the result

 - Package inspection retains decoded manifest XML, summary declarations,
@@ -101,19 +160,37 @@ selected overload. Native/abstract methods with no decompiled body report
  upstream engine and should be checked against the retained XML when needed.
  Signature verification is explicitly `not_performed`.
 - Class search consumes all pages and checks their counts. An empty `query`
-  inventories all class names; class names are decompiler representations.
+  inventories all class-name rows; class names are decompiler representations.
+  Repeated names remain repeated in `matches` so the inventory row count and
+  producer evidence stay visible. `duplicate_names` identifies matching names
+  with multiple rows and their occurrence counts; a name alone does not
+  distinguish which provider class node a later lookup selects.
 - Class inventories read parsed metadata without generating source and return
  full reported display types, fields and inner class names. Synthetic members
  that code generation hides can appear in the inventory; it stays consistent
  after method decompilation within the retained session.
-  Exact DEX descriptors are unknown. Overload indices are tied to this artifact
-  and engine, not stable identities across different builds.
+  Exact DEX prototypes are reported from frozen JADX method metadata as
+  `dex_descriptor`, for example `(I)V`. Display names can be JADX aliases;
+  the descriptor alone is not a complete raw DEX method ID. Overload indices
+  are tied to this artifact and engine, not stable identities across builds.
 - Method inspection reports Java or smali, fallback markers and complete/partial
  returned text. Decompiled source is a derived representation. A complete text
  result does not establish complete semantics or a byte-identical reconstruction.
 - Reference results are incoming static relationships between provider nodes.
-  Source lines and DEX instruction offsets remain unknown. They are not observed
-  runtime calls.
+  `reference_unit=using_node` means `total` counts containing JADX nodes, not
+  individual invoke instructions; two calls from one method can appear as one
+  entry. `coverage=complete` covers this node inventory only. For a method
+  reference whose using node has raw code, `direct_invokes` lists matching DEX
+  invoke instructions with their DEX entry, raw source method ID, code-unit
+  offset, method index and opcode. An empty array means no direct invoke was
+  matched in that node; `null` means raw code was unavailable or the using node
+  was not a method. This list does not establish complete call-site coverage
+  outside JADX's using-node inventory or through reflection and dynamic loading.
+  A virtual or interface invoke also does not establish its runtime implementation.
+  Source lines and the containing node's own `dex_offset` remain unknown.
+  Method-reference results also carry the selected `dex_descriptor`;
+  class-reference results report `null`.
+  These relationships are not observed runtime calls.
 - Original requests, APK SHA-256, actual engine JAR SHA-256 and raw producer
  responses remain in Evidence. Source revision is populated only for bytes
  matching the audited release JAR.
@@ -179,9 +256,11 @@ REA never writes to the original APK or launches target code.

 On Windows, stdin writes preserve stream backpressure and run outside the Node
 event loop. Cancellation terminates the retained job, including descendants,
-before workspace cleanup. An abrupt CLI/server process termination closes the
-job and stops Java, but cannot run JavaScript workspace cleanup; temporary
-workspace files can remain after that forced exit.
+before workspace cleanup. On Windows, a detached REA helper owns the private
+workspace's native directory handle. An abrupt CLI/server owner termination
+closes the Java job and the helper's input pipe; the helper then removes the
+workspace by handle. The real Windows lane checks both effects. If the helper
+also dies or the host loses power before it finishes, private files can remain.

 If Java cannot start the bridge, REA reports the selected executable and retained
 startup diagnostics. Select a full JDK through `JAVA_HOME` or `PATH`, and use
@@ -244,12 +323,20 @@ The original APK and host Java configuration remain unchanged.

 The same lane runs on Windows x64 with the bundled native controls and a full
 JDK. It checks CLI/MCP results, real MCP cancellation and disconnect, and Java
-exit after abrupt CLI owner termination. Windows forced termination is checked
-separately from the POSIX CLI SIGTERM handler; it does not claim a structured
-cancellation result or workspace removal after the owner is forcibly killed.
+exit and workspace removal after abrupt CLI owner termination. Windows forced
+termination is checked separately from the POSIX CLI SIGTERM handler. A
+synthetic two-owner test confirms that one owner's exit does not remove the
+other owner's workspace. The lane does not claim a structured cancellation
+result or power-loss recovery.

 The download script verifies fixed SHA-256 values, reuses matching files and
 refuses to overwrite a different existing file. APK/JAR files live under ignored
 `_reference/` and are excluded from package contents. The lane exercises actual
 REA CLI and MCP results with the real engine, separately from synthetic protocol
 regressions. It needs no Gradle build, emulator, Android SDK or Ghidra.
+
+Direct invoke sites retain exact DEX code-unit offsets separately from each
+containing node's unknown `dex_offset`. `direct_invokes_exhaustive: false`
+means instruction rows were bounded by the local bridge frame budget; `null`
+means the producer did not establish instruction coverage. Using-node totals
+and coverage do not imply complete instruction evidence.
@@ -22,6 +22,14 @@ reports the resolved launcher, its selection source, and its version. A
 missing launcher reports `capability_unavailable` with the
 `REA_APKTOOL_COMMAND` remediation rather than attempting an installation.

+On Windows, a `.bat` or `.cmd` launcher (such as Chocolatey's `apktool.bat`)
+runs through an explicit `cmd.exe /d /s /c` call. That call has no AutoRun,
+no delayed expansion, and quotes every argument. A launcher path that
+contains `"`, `%`, `^`, `!`, `&`, `|`, `<` or `>` is reported as
+`not_configured` instead of being passed to `cmd.exe`. If REA's temporary
+decode paths contain those characters, decoding through a batch launcher
+reports `capability_unavailable`.
+
 ## Answer an analyst question

 ```sh
@@ -174,7 +174,7 @@ Resolution follows dyld:
 - `@rpath` tries each `LC_RPATH` of the loading image first, then those of the
  image that loaded it, up to the process root.
 - Dependencies load dependents-first in load-command order, and each image
-  loads once per process. A later request whose install name matches an
+  loads once per process. A later @rpath request whose install name matches an
  already loaded image reuses it.
 - Symlinks such as `Versions/Current` are followed inside the analyzed root.
  An escape through `..` or an absolute link is `escapes-target`.
@@ -6,7 +6,7 @@ graph TD
    end

    subgraph "Production Composition"
-        COMPOSITION["src/composition/<br/>fresh typed provider and session factories<br/>one-shot binding and independent optional loading"]
+        COMPOSITION["src/composition/<br/>fresh typed provider and session factories<br/>one-shot binding and independent optional loading<br/>static WASM/EVM descriptors: named contracts, factories, retryable lifetimes"]
    end

    subgraph "Adapters"
@@ -27,6 +27,7 @@ graph TD

    subgraph "Provider Layers"
        PWNTOOLS["src/native/pwntools/ + src/native/pwndbg/<br/>BYO offline ELF/core adapters<br/>owned snapshots/Python and optional core-only GDB<br/>unchanged upstream parsers, ABI structures and mapping API"]
+        WASM["src/wasm/<br/>caller-supplied WABT; immutable artifact inspection<br/>owned process supervision and retryable workspace cleanup"]
        EVM["src/evm/<br/>unchanged EVMole/WASM in bounded owned worker<br/>explicit carrier parsing and inferred interface adapter"]
        HOPPER["src/hopper/<br/>Hopper launch & Unix-socket protocol<br/>BridgeLauncher, HopperClient, protocol"]
        GHIDRA["src/ghidra/<br/>Ghidra discovery, profile & function analysis<br/>snapshot, Unix/Windows transport, bounded lifecycle"]
@@ -37,6 +38,10 @@ graph TD
        JS_CODEC["src/javascript/sourceMaps/<br/>pinned upstream map codec<br/>owned heap-limited Node process, complete point evidence"]
        INSPECTOR["src/inspector/<br/>passive V8 Inspector discovery and capture<br/>script and target locations"]
        DOTNET["src/dotnet/<br/>Managed static metadata provider<br/>PE/CLI metadata inventory, members, IL"]
+        OBJDUMP["src/objdump/<br/>GNU objdump command profile"]
+        RIZIN["src/rizin/<br/>Static command profile and persistent debugger sessions"]
+        GDB["src/gdb/<br/>Persistent GDB/MI sessions<br/>owned process lifecycle and target state"]
+
        FIRMWARE["src/firmware/<br/>Binwalk / Unblob adapters<br/>owned processes, report normalization, private extraction"]
        ANDROID["src/android/<br/>Static APK adapter<br/>owned SDK stdio, pinned headless JADX, normalization"]
        REFERENCE["src/reference/<br/>Filesystem reference-source reader<br/>bounded traversal, validation, errors"]
@@ -49,7 +54,7 @@ graph TD
    end

    subgraph "Contracts Layer"
-        CONTRACTS["src/contracts/<br/>Caller-visible schemas for analysis, managed, browser, Electron, application, and session tools"]
+        CONTRACTS["src/contracts/<br/>Caller-visible schemas for analysis, reverse engineering, managed, browser, Electron, application, and session tools"]
        ERRORS["src/contracts/errorSchemas.ts<br/>Shared analysis error schemas<br/>CLI diagnostics, generated error catalog"]
    end

@@ -94,6 +99,9 @@ graph TD
    APP -->|record callbacks| RECORDS
    SERVER -->|narrow record ports| RECORDS
    RECORDS --> DOMAIN
+    COMPOSITION --> WASM
+    WASM --> PROCESS
+    WASM --> DOMAIN
    COMPOSITION --> PWNTOOLS
    PWNTOOLS --> PROCESS
    PWNTOOLS --> DOMAIN
@@ -118,6 +126,14 @@ graph TD
    INSPECTOR --> DOMAIN
    APP --> DOTNET
    DOTNET --> DOMAIN
+    COMPOSITION --> OBJDUMP
+    OBJDUMP --> DOMAIN
+    COMPOSITION --> RIZIN
+    RIZIN --> DOMAIN
+    CLI --> GDB
+    SERVER --> GDB
+    GDB --> PROCESS
+    GDB --> DOMAIN
    HOPPER --> HOPPER_BRIDGE
    GHIDRA --> GHIDRA_BRIDGE
    HOPPER --> PROCESS
@@ -98,10 +98,14 @@ section, symbol, mitigation or linkage views. The default `complete` detail and
 CLI output are unchanged.

 Complete results have a 32 MiB input, 64 MiB reply, 1 MiB combined diagnostics
-and 30-second owned command deadline. The reply budget applies to the decoder
-record; the CLI/MCP Evidence envelope adds copies and encoding overhead. The
+and 30-second owned command deadline. `REA_PWNTOOLS_MAX_INPUT_BYTES` raises the
+input bound alone, up to 1 GiB; the reply and deadline bounds are unchanged. The
+reply budget applies to the decoder record; the CLI/MCP Evidence envelope adds
+copies and encoding overhead. The
 pinned MCP SDK defaults to a 10 MiB receive buffer. Large complete results need
-a caller-configured `StdioClientTransport({ maxBufferSize: ... })`; the large
+a caller-configured `StdioClientTransport({ maxBufferSize: ... })` and a matching
+server `REA_MCP_MAX_RESPONSE_BYTES`. See [MCP delivery budgets](mcp-contracts.md#tool-results).
+These settings affect delivery, not decoder coverage or limits. The large
 extended-index verification uses 256 MiB and an explicit five-minute request
 timeout in its separate opt-in MCP lane. REA does not silently truncate the result. Python lowers resource soft limits to at most 3 GiB virtual address space,
 30 CPU seconds and 64 MiB file output. Inherited tighter soft/hard limits are
@@ -181,7 +185,8 @@ payload ranges are supported. The supported tree has three levels: type, name,
 and language. Other resource payload formats remain opaque.

 `max_file_bytes` defaults to 64 MiB (maximum 512 MiB); `max_entries` defaults to
-4096 (maximum 65536). CLI options are `--max-file-bytes` and `--max-entries`.
+16384, the largest stock Windows resource image measured with headroom
+(maximum 65536). CLI options are `--max-file-bytes` and `--max-entries`.
 Metadata is bounded to 8 MiB; exceeding any budget fails without truncation.
 Large Evidence envelopes may require a caller-configured MCP transport receive
 buffer because envelope copies add overhead. Requests honor cancellation during
@@ -8,6 +8,10 @@ exported script's native imports under an explicit URL/import-map context.

 REA can attach to a user-owned Chrome-family browser through the Chrome DevTools Protocol (CDP) and produce bounded Evidence about an existing page. This is a passive reverse-engineering capability, not a general browser automation or remote-control surface.

+For request/response bodies and traffic debugging, see
+[Fiddler Everywhere captures and MCP](web-network-captures.md#reading-fiddler-captures-saz).
+Inspect native SAZ or HAR 1.2 offline with REA, or connect Everywhere's separate MCP server.
+
 ## Shipped surfaces

 - `list_browser_targets` / `rea list-browser-targets` discovers all eligible page targets at the endpoint by default, with an optional exact-origin filter.
@@ -24,6 +28,29 @@ REA can attach to a user-owned Chrome-family browser through the Chrome DevTools

 Target discovery returns the complete in-scope target array in one result.

+## Screenshot comparison memory budget
+
+`compare_web_screenshots` admits both PNGs against a 256 MiB estimated
+working-memory budget before inflating either image. This is a processing
+buffer budget for this operation, not a PNG dimension limit, CDP transport
+limit, or process RSS guarantee. Header-valid dimensions that exceed the
+budget return `resource_constraint` with `details.resource: "memory"` and the
+estimated and maximum working-set bytes. Invalid PNG structure remains
+malformed input; encodings outside supported pixel-comparison coverage remain
+unsupported-target errors.
+
+The conservative estimate includes the two decoder-owned decoded input
+buffers, the current image's concatenated IDAT buffer, Node zlib's default
+16 KiB output chunk backing rounded up for retained chunks, the concatenated
+inflate result while its output chunks are retained, one possible extra output
+chunk before zlib checks `maxOutputLength`, unfiltered scanline storage,
+RGB-to-RGBA expansion, and the first retained RGBA image while the second is
+decoded. RGBA output aliases its unfiltered buffer and is counted once.
+Admission uses checked integer arithmetic before image-sized allocation.
+Caller-owned JSON/base64 strings, input parsing and validation, transport
+buffers, garbage-collector behavior, and other runtime memory are outside this
+estimate; the budget does not promise a bound on total process memory.
+
 Electron `file://` pages use a separate provider and target boundary; see [electron-observation.md](electron-observation.md).

 ## Provider authorities
@@ -53,12 +80,74 @@ google-chrome \

 REA does not launch, own, or terminate this browser. Use a dedicated profile and stop it yourself when the investigation is complete. Do not expose the debugging port on a non-loopback interface.

+Other Chromium-family browsers, such as Chromium or Brave, work the same way
+with the same two flags. On macOS, run the executable inside the application
+bundle, for example
+`"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"` or
+`"/Applications/Brave Browser.app/Contents/MacOS/Brave Browser"`. Fedora installs
+Chromium as `chromium-browser`.
+
+REA needs the browser's HTTP discovery endpoint on a loopback port. A browser
+driven over `--remote-debugging-pipe` or a private channel, for example by an
+automation SDK, exposes no such endpoint, so `list_browser_targets` has nothing
+to attach to. Connection failures report `endpoint_unreachable`; a reachable
+HTTP service that returns a non-200 or non-JSON discovery response reports
+`invalid_endpoint_response`.
+Reproduce that flow in a Chromium-family browser started with
+`--remote-debugging-port`.
+
+In a minimal Linux container, headed Chromium under Xvfb can keep running while
+its DevTools port never opens. Install the GLES, EGL, and Mesa DRI libraries
+(on Fedora: `libglvnd-gles`, `mesa-libEGL`, and `mesa-dri-drivers`), then
+restart the browser.
+
 On Linux, a reachable but locked desktop keyring can stall Chrome before its
 first navigation. For a dedicated investigation profile, add
 `--password-store=basic` to bypass that keyring. Empty or placeholder frame URLs
 then describe an unloaded page; REA continues to require an allowed, committed
 URL. See [the startup investigation](https://github.com/morluto/rea/issues/1384#issuecomment-6093781686).

+### Browsers without HTTP discovery
+
+Some browsers expose a loopback browser WebSocket but no `/json/version` or
+`/json/list`. Chrome 144 and later can enable remote debugging on a running
+browser from `chrome://inspect/#remote-debugging`; this mode is reported to
+serve only the browser socket. Since Chrome 136, `--remote-debugging-port` is
+ignored for the default profile, so this is the remaining way to inspect a page
+in an existing profile.
+
+For such a browser, pass the browser socket instead of the HTTP endpoint:
+
+```bash
+rea list-browser-targets ws://127.0.0.1:9222/devtools/browser/<id>
+```
+
+Chrome writes the port and the `/devtools/browser/<id>` path to the
+`DevToolsActivePort` file in its user data directory. REA does not read that
+file for you.
+
+- Only `list_browser_targets`, `inspect_web_page`, `analyze_web_bundle`,
+  `observe_web_session`, `discover_webmcp_tools`, and `capture_web_screenshot`
+  accept a browser socket. Website runtime attribution, Electron inspection,
+  and scenario attach mode still require HTTP discovery.
+- Each call opens one WebSocket and uses it for discovery
+  (`Browser.getVersion`, `Target.getTargets`) and for the selected page.
+  Discovery must finish within the same 10-second deadline as HTTP discovery.
+  Chrome may ask you to approve each connection.
+- The browser identifier controls access to the socket. Evidence records the
+  endpoint as `ws://127.0.0.1:9222/devtools/browser/[REDACTED]`.
+- `browser.revision` is the socket's `Browser.getVersion` value, such as
+  `@e533e98…`. HTTP discovery reports `WebKit-Version`, such as
+  `537.36 (@e533e98…)`.
+- `Target.getTargets` reports whether each target is attached. Chrome's
+  `/json/list` can omit it, so HTTP discovery may report `attached` as `null`.
+- An existing profile can hold many unrelated pages. Target listing returns
+  each eligible page's URL and title, so pass `--allowed-origins` to list only
+  the origins you are investigating.
+
+A browser driven only through `--remote-debugging-pipe` or a vendor's private
+automation channel has neither endpoint, and REA cannot attach to it.
+
 ## Request boundary

 CDP connections enforce the shared 64 MiB limit per WebSocket message. A larger
@@ -68,9 +157,13 @@ decoded strings or JSON objects.

 Discovery supplies a literal loopback CDP endpoint and returns all eligible
 page targets unless an exact HTTP(S) origin filter narrows the result. Follow-up
-inspection requests select one target; their default scope is its current origin. `localhost`, private-LAN
-addresses, HTTPS CDP endpoints, credentials, paths, queries, fragments, and
-implicit ports are not valid CDP endpoints. Origin filters contain only scheme,
+inspection requests select one target; their default scope is its current origin.
+The endpoint is either `http://127.0.0.1:PORT` or a
+[browser socket](#browsers-without-http-discovery),
+`ws://127.0.0.1:PORT/devtools/browser/ID`; `[::1]` may replace `127.0.0.1`.
+`localhost`, private-LAN addresses, HTTPS and `wss:` endpoints, credentials,
+queries, fragments, implicit ports, and any other path are not valid CDP
+endpoints. Origin filters contain only scheme,
 host, and port; paths, credentials, queries, fragments, and wildcards are not
 valid origins.

@@ -150,6 +243,7 @@ REA retains local observation data and removes only URL userinfo credentials, ex
 - Console observations with a stack source retain call type, argument types, timestamp, and source location. When console text capture is selected, already-delivered primitive values are retained verbatim; objects, getters, and remote properties are never expanded.
 - WebSocket observations retain direction, opcode, and payload byte length. When shape capture is selected, text frames are classified as text or complete value-free JSON shape; binary bytes, hashes, prefixes, and raw frames are never retained.
 - Storage observations always redact values. Key names, IndexedDB names, and cache names require `include_storage_keys`; stable content fingerprints require the additional `include_storage_fingerprints` selection. Cookie fingerprints cover cookies applicable to the current main-frame URL, including its path, rather than every cookie path on that origin. Other storage remains scoped to the origin. Cache bodies above 64 KiB and partial IndexedDB remote objects make fingerprint coverage incomplete, so identical observations remain `unknown`.
+- Unavailable or malformed storage inventories mark `storage_keys` unavailable, even without fingerprints. Comparisons suppress inferred key additions and removals while preserving modifications proven by complete fingerprints on both sides. Previously exported captures are not retroactively corrected; recapture them to obtain these coverage markers.
 - Script metadata is included only when CDP supplies a URL on an allowed origin. Stable keys exclude transient CDP script IDs, and exact transient raw URLs are used only during script/resource reconciliation. URL-less scripts are excluded because their origin cannot be established. When CDP supplies an execution-context association, the accepted script retains its authorized frame ID for later attribution. Script metadata admission bounds retained identity and source-map strings to 8 MiB and 50,000 records. When a source-map URL cannot fit, the script identity remains and source-map coverage reports `resource_budget_exhausted`; rejected script observations are counted without retaining their IDs. Source content is omitted unless explicitly requested and is returned inline as a self-verifying artifact with its SHA-256 digest.

 Frames, resources, scripts, events, and workers outside the selected origin scope are excluded. An explicit `allowed_origins` list can include additional origins; excluded target details are counted without being exposed.
@@ -170,12 +264,31 @@ validation and resource bounds. Omitted script metadata makes its affected
 section explicitly truncated.
 Disallowed-origin entries are filtered before retention.

-CDP discovery, WebSocket connection, and commands continue until the peer responds or the caller cancels the operation; REA has no internal byte, message-count, or pending-command quota for these exchanges. Without caller cancellation, a stalled peer can leave an operation waiting. On cancellation, REA closes only its own WebSocket, waiting up to one second before terminating it. Malformed unsolicited events and malformed replies to pending commands poison the connection so later commands fail closed. When a correlated reply reports a session ID, it must match the command's selected session. An optional method is reported unavailable only when the browser rejects it as unknown; other command rejections retain their reported failure.
+CDP discovery, connection, and commands wait for the peer or caller cancellation;
+a stalled peer can leave an operation waiting. WebSocket messages remain subject
+to the [64 MiB limit](#request-boundary). REA does not impose a message-count or
+pending-command quota. On cancellation, it closes only its own WebSocket,
+waiting up to one second before terminating it. Malformed unsolicited events
+and malformed replies to pending commands poison the connection so later
+commands fail closed. A correlated reply's session ID must match the command's
+selected session. An optional method is unavailable only when the browser
+rejects it as unknown; other command rejections retain their reported failure.

-Network and console coverage starts only after REA attaches and enables the relevant CDP domains. `prior_activity_available` is always `false`; absence from these arrays is not evidence that an event never occurred. If CDP reuses a request ID without supplying redirect metadata, REA retains the predecessor observation and marks network coverage incomplete.
+Network and console coverage starts only after REA attaches and enables the relevant CDP domains. `prior_activity_available` is always `false`; absence from these arrays is not evidence that an event never occurred. Scripts are different: scripts the page parsed before REA attached are still reported under `scripts`. A page that loaded before attach can therefore list its scripts while `network.requests` lacks the requests that fetched them. If CDP reuses a request ID without supplying redirect metadata, REA retains the predecessor observation and marks network coverage incomplete.

 Source maps are not fetched by inspection. Bundle analysis fetches them only when `fetch_source_maps` is selected, with exact-origin redirect checks and no credentials, cookies, or referrer. Each fetch operation has a 30-second deadline and a 64 MiB aggregate streamed-response budget across redirects. A deadline or byte-limit failure remains explicitly unavailable and makes bundle completeness partial. Source-map decoding also bounds raw resolved identities to 64 MiB and aggregate inline evidence to 32 MiB. Failed maps preserve their failure context when it fits; if another context cannot fit, collection stops with the full requested count, the returned processed count, and an explicit unknown-results limitation. Retained mappings and original sources are returned inline. See [source-location limits](web-source-location.md) for the structural decoder bounds. Original source artifacts with `.ts`, `.tsx`, `.mts`, or `.cts` suffixes carry `text/typescript`; other source names retain `text/javascript`.

+source-map items retain the script's `declared_url` separately from the observed
+`effective_url`. successful maps use the latter as their source-resolution base.
+failed fetches retain the last response URL observed within the approved origins;
+`effective_url` is `null` when no response was observed. query parameters are
+preserved, and both URLs count toward the inline evidence budget.
+Current fetches emit a string or null; historical observations may omit
+`effective_url` without changing their stored content or Evidence identity.
+An observed URL must fit before its response body is read. If it cannot fit,
+remaining map results are explicitly unknown; failed cancellation retains its
+cleanup context and resource constraint.
+
 ## Non-goals and threat model

 The browser provider deliberately does not expose generic CDP commands, `Runtime.evaluate`, WebMCP invocation, navigation, input, downloads, page closure, or browser closure. Screenshot capture is the sole pixel surface and returns an image only when requested. REA disables the domains it enabled, detaches the target session, and closes only its own WebSocket.
@@ -262,8 +262,13 @@ Capturing more data alone does not establish a website change.
 states. Captured bodies retain base64 bytes, byte count, media type, and SHA-256
 of the retained bytes. Request bytes are those exposed by Playwright; a
 `not_exposed` request body does not prove absence or complete multipart file
-coverage. Response bytes are decoded by the browser, not the compressed wire
-representation. No later fetch is substituted for the original response.
+coverage. Response bytes are the result of Playwright's `Response.body()` API.
+Playwright may satisfy that API with a credentialed follow-up load of the
+request URL when CDP has no response body. Event and transaction references
+identify the associated Playwright response object, but do not verify that the
+retained bytes came from the original network transaction. Successful selected
+response bodies include this limitation in the capture. These bytes are not
+the compressed wire representation.
 Headers preserve producer order and duplicates. Authorization,
 Proxy-Authorization, Cookie, and Set-Cookie values are null with
 `redacted: true`; ordinary values remain intact. Declared UTF-8 secret values
@@ -307,3 +312,9 @@ artifacts carry their normalized before/after digests and capture states.
 Every changed or unknown artifact record is returned inline for each aligned
 step. Missing, truncated, or mismatched capture coverage remains unknown
 instead of being treated as equality.
+
+`scenario.event_families` records the sorted, de-duplicated `capture.events`
+selection. Captures with different selections report event evidence as unknown
+with a recovery instruction, even when no selected family emitted an event.
+Captures written before this field existed omit it, so their event evidence is
+also unknown.
@@ -37,8 +37,64 @@ module page. See
 [JavaScript application workflows](javascript-application-workflows.md) for
 views, feature traces and comparisons that take the same saved Evidence.

+## Managed assemblies (.NET)
+
+For a .NET PE/CLI assembly, inspect its identity and deployment form before
+inspecting members. These commands read metadata and Common Intermediate
+Language (CIL) without loading or executing the assembly and require no native
+analysis engine:
+
+```bash
+rea inspect-managed-artifact /absolute/path/Assembly.dll --json
+rea inspect-managed-members /absolute/path/Assembly.dll --json
+rea inspect-managed-members /absolute/path/Assembly.dll --type Example.Parser --json
+```
+
+Replace `Example.Parser` with an exact `full_name` from the member inspection,
+or a declaring type name you already know. Omit `--type` to inspect all members,
+or repeat it to select several types.
+Selection limits decoding to the selected types; member references remain an
+assembly-wide inventory, so the output can still be large.
+
+Choose the workflow from the artifact's contents rather than its `.exe` or
+`.dll` extension. The managed parser requires PE/CLI metadata; it does not
+unpack .NET single-file hosts or decode IL2CPP metadata. NativeAOT method bodies
+require native analysis. See the [managed-code guide](managed-code-analysis.md)
+for supported deployment forms, selection semantics, and coverage limits.
+
+## SQLite snapshots
+
+For a local database snapshot, use `rea inspect-sqlite-database PATH --json`.
+Add `--table NAME --row-limit 20` to read a bounded sample from an ordinary
+table. This command does not require a native analysis provider. See
+[SQLite database snapshots](sqlite-databases.md) for database/WAL identity,
+source-file handling and row-value semantics.
+
 ## Native analysis

+Inspect embedded Go build metadata without installing Go or configuring a
+disassembler:
+
+```bash
+rea inspect-go-binary /absolute/path/to/program --json
+```
+
+The shared CLI/MCP workflow reports compiler version, modules, replacements,
+build settings, original bytes and SHA-256. See [Go binaries](go-binaries.md)
+for supported containers and absent or malformed metadata.
+
+Inspect an embedded cargo-auditable Rust dependency list without installing
+Rust or configuring a disassembler:
+
+```bash
+rea inspect-rust-binary /absolute/path/to/program --json
+```
+
+The same workflow reports package names, versions, sources, the root package,
+dependency edges, and the section's file range. A missing `.dep-v0` section is
+null metadata, not proof that the binary has no dependencies. See
+[Rust binaries](rust-binaries.md).
+
 Configure [Hopper or Ghidra](installation.md#hopper), or the
 [IDA adapter](ida-provider.md), before analyzing a native target. Substitute
 your target, search text and function name or address in these examples:
@@ -57,9 +113,10 @@ a function dossier; `decompile` returns pseudocode. For assembly instructions
 alone, use `rea instructions`. A macOS `.app` bundle can be supplied directly.

 Run `rea --help` or a command's `--help` for arguments and output options.
-The [generated catalog](product-catalog.json) lists all CLI commands and MCP
+The [generated catalog](mcp-contracts.md#generated-catalog) lists all CLI commands and MCP
 tools. Provider-specific target support is described in the
-[native guide](native-investigation.md), [DOS guide](ghidra-dos.md), and
+[native guide](native-investigation.md), [DOS guide](ghidra-dos.md),
+[raw binary guide](raw-binary.md), and
 [Windows Ghidra guide](windows-ghidra-p0.md).

 ## Choose a provider
@@ -113,6 +170,10 @@ an exact result when the target bytes, operation, parameters, provider and
 settings match. Mutations and cursor-dependent calls are excluded from the
 cache. Snapshot files are local and use owner-only permissions.

+Dispatching a metadata write invalidates the active session's snapshot cache,
+even if its reply fails: the provider may already have applied the write.
+Close and reopen the target before importing or saving another snapshot.
+
 Snapshots retain eligible target-scoped question histories in full, including
 mutation Evidence recorded while another target was active. Observations and
 related questions must still belong to the saved target. If those dependencies
@@ -125,7 +186,9 @@ rea analyze /absolute/path/to/program --provider ghidra --snapshot /absolute/pat
 rea analyze /absolute/path/to/program --provider ghidra --snapshot /absolute/path/to/analysis/program.json
 ```

-An exact CLI cache hit is read before starting a provider process.
+An exact CLI cache hit is normally read before starting a provider process.
+On Windows, Ghidra first admits the selected source through its native snapshot
+owner, so replay can start that provider.
 In MCP, `open_binary` accepts `snapshot_path` to import a snapshot atomically
 for its matching target; an MCP provider may still start before returning a
 cached result. `close_binary` accepts `snapshot_path` and optional
@@ -163,10 +226,40 @@ Import an older source tree to compare with the current artifact:

 ```bash
 rea import-reference-source /absolute/path/to/source
+rea import-reference-source /absolute/path/to/source --detail summary --outline-depth 2 --json
 ```

+The default (`--detail full`) returns the complete `HistoricalSourceGraph`.
+Choose `--detail summary` for a `HistoricalSourceSummary` of that same committed
+inventory: its `HistoricalSourceManifest` contains the full graph and root
+digests, with provenance, VCS facts, languages, manifest paths, entry and
+relationship counts, parse-failure counts, file/directory states, exclusions
+and limitations. The summary is a presentation view, not a graph accepted by
+source-to-bundle comparison; keep the full output for that workflow.
+
+`--outline-depth` requires `--detail summary` and accepts a nonnegative safe
+integer. Depth 1 shows root children, depth 2 (the default) also shows their
+children, and depth 0 omits all outline entries. Paths retain canonical Unicode
+code-point order. `outline.omitted_entry_count` counts only inventoried entries
+hidden by the chosen depth; it does not count unexamined children of excluded or
+unreadable directories, or change `inventory_state`. Exclusions and source
+limitations remain separate from this presentation omission.
+
+`file_sizes.known_bytes` is the exact sum of known inventoried file sizes as
+decimal text, so large sums cannot lose integer precision. Unknown sizes count
+in `file_sizes.unknown_file_count`; excluded or unexamined paths are not included
+in that total. A zero known-byte sum therefore does not establish an empty source
+tree. Full output retains individual file hashes, relationships and parse-failure
+diagnostics. Summary mode performs the same import and cleanup, then projects
+the committed result without rereading source files. It reduces the selected
+output, not the amount of source examined. A shallow tree or many exclusions can
+still produce a large summary; depth 0 removes the outline but preserves the
+other facts. No fixed output-size or runtime improvement is promised.
+
 The import records hashes and metadata for the supplied files, separately
-from observations of the current app. It honors the source tree's `.gitignore`
+from observations of the current app. For a directory inside a Git checkout,
+it records the HEAD of the nearest enclosing repository; whether the working
+tree matches that commit remains unknown. It honors the source tree's `.gitignore`
 and applies default exclusions for common generated and dependency paths such
 as `node_modules/`, `dist/`, and `*.log`. Pattern-based exclusions record the
 matched pattern and whether it came from project, default, caller, or sensitive
@@ -0,0 +1,372 @@
+# Client configuration and troubleshooting
+
+Use this reference when a client's configuration paths, profiles, trust settings,
+or schema limits affect REA. Start with [installation and setup](installation.md)
+for the reviewed setup transaction and [client IDs](installation.md#supported-agents).
+The versioned verification notes below describe their stated runs; live service
+acceptance is verified only where explicitly recorded.
+
+## Fiddler Everywhere alongside REA
+
+Connect Fiddler Everywhere's MCP server separately for HTTP traffic debugging,
+preserving your REA client entry. See the
+[setup instructions and public skills](web-network-captures.md#fiddler-everywhere-mcp-and-public-skills).
+
+## GitHub Copilot CLI
+
+For GitHub Copilot CLI 1.0.95, `gpt-4.1` model metadata can block ordinary
+chat with REA's full input-schema profile before sending a model request
+(`compaction_static_context_blocked`; [#1554](https://github.com/morluto/rea/issues/1554)).
+For the native offline OpenAI-compatible BYOK adapter, full-profile chat and
+analysis also passed with a distinct `COPILOT_PROVIDER_MODEL_ID` such as
+`rea-gpt-4.1`, `COPILOT_PROVIDER_WIRE_MODEL=gpt-4.1`, and that distinct ID selected
+through `copilot --model rea-gpt-4.1`. The local fixture requested
+`COPILOT_PROVIDER_MAX_PROMPT_TOKENS=1000000`; configure the limit supported by
+your backend. This keeps every full schema and the actual wire model intact.
+It does not establish the effective client budget or fix the built-in-ID path.
+The existing compact profile passed native chat and analysis with the complete
+tool inventory on Linux using a loopback model fixture. After setup, add
+`"REA_MCP_INPUT_SCHEMA_PROFILE": "compact"` to the `env` object of the
+`mcpServers.rea` entry in Copilot's `mcp-config.json`, preserving its other
+settings, then restart the client. Compact advertisements can reduce nested
+schema constraints; REA still validates complete canonical inputs. Full-profile
+`gpt-5.4` metadata is also verified locally. Live model API acceptance and
+Windows remain unverified; see [native client verification](testing.md).
+
+## Claude Desktop
+
+Setup registers `mcpServers.rea` in `claude_desktop_config.json`: under
+`~/Library/Application Support/Claude` on macOS, under
+`$XDG_CONFIG_HOME/Claude` (default `~/.config/Claude`) on Linux, and on Windows
+in the file the installed app reads:
+
+- Without Claude Desktop's MSIX package storage, setup uses
+  `%APPDATA%\Claude\claude_desktop_config.json`.
+- The MSIX build (package family `Claude_pzs8sxrjxfjjc`, installed under
+  `C:\Program Files\WindowsApps\Claude_<version>_x64__pzs8sxrjxfjjc`)
+  virtualizes AppData writes. Once it has saved its configuration, it reads
+  the package-private copy
+  `%LOCALAPPDATA%\Packages\Claude_pzs8sxrjxfjjc\LocalCache\Roaming\Claude\claude_desktop_config.json`
+  and ignores the `%APPDATA%` file. Setup, doctor and uninstall use that
+  private copy when it exists. Before the package has saved one, they use the
+  `%APPDATA%` file it still reads; with neither file, setup creates the
+  private copy.
+- If `%LOCALAPPDATA%\Packages` holds a `Claude_<publisher ID>` package from
+  another publisher, or cannot be read, REA cannot establish which file Claude
+  Desktop reads. Setup, doctor and uninstall report that instead of falling
+  back to `%APPDATA%`. Doctor then reports `claude_desktop` as invalid, and
+  uninstall stops before removing anything for any client.
+
+Quit Claude Desktop fully before approving setup; closing its window is not
+enough. A running Claude Desktop can rewrite its configuration from memory and
+drop the registration. This was observed with the Windows MSIX build. Start
+Claude Desktop again after setup. If the package saves its private copy after
+setup planned to write the `%APPDATA%` file, setup makes no change and asks you
+to rerun it.
+
+A process that the MSIX app started, such as a terminal in Claude Desktop's
+Code tab, sees the package's private files at their `%APPDATA%` paths even
+though it reports no package identity. Setup compares one of those files at
+both paths. When it finds that redirected view, the app is running, so setup
+and uninstall refuse to change Claude Desktop's configuration; rerun them from
+a terminal that Claude Desktop did not start. If the package has no private
+file to compare, setup and uninstall cannot rule that out and also refuse; open
+Claude Desktop once if it has never run, quit it fully, then rerun. Doctor
+reads the same file from either kind of process.
+
+The same process also writes everything under `%APPDATA%` and
+`%LOCALAPPDATA%` into the package's private storage, where other apps never
+read it. From that redirected view, setup and uninstall refuse to change other
+clients' files there (VS Code, Trae, Devin and Hermes) and skill directories
+there. `npm install -g` run from such a process installs into the package's
+`LocalCache\Roaming\npm`, which other processes cannot see at the `%APPDATA%`
+path. Setup refuses to register a launcher stored in any MSIX package's
+private storage. Install REA from an ordinary terminal and rerun setup there.
+
+Uninstall removes REA's entry from both the private copy and the `%APPDATA%`
+file, including registrations that earlier releases wrote to `%APPDATA%`.
+
+REA recognizes only the MSIX package. A non-MSIX Claude Desktop installed
+beside it, which would read the `%APPDATA%` file, is not detected.
+
+The package identity, both file locations, the redirected view, and these
+setup and doctor results were observed on Windows 11 with the MSIX build
+2.31226.1.0, using read-only `doctor` and `setup --dry-run` runs both inside
+and outside the app. On that host, setup run from an ordinary terminal with
+the app quit wrote the package-private copy; after the app restarted, the
+entry was still present, doctor reported it aligned, and Claude Desktop's MCP
+log showed REA starting and answering `tools/list`. Uninstall against the real
+app is unverified.
+
+## Claude Code
+
+Claude Code normally defers MCP tool definitions through tool search. It
+[turns tool search off](https://code.claude.com/docs/en/mcp#configure-tool-search)
+when `ANTHROPIC_BASE_URL` points to a non-first-party host, such as DeepSeek's
+Anthropic-compatible endpoint. It then sends every REA tool's name, description
+and input schema with every request
+([#2019](https://github.com/morluto/rea/issues/2019)). On main at 04089d5be,
+with 172 tools, that payload measured about 496 KB with the default profile and
+about 290 KB with `compact`. By the o200k and cl100k tokenizers, that is roughly
+120k and 68k tokens. Compaction cannot remove tool definitions. If they nearly
+fill the window Claude Code assumes, the context refills after every compaction
+and Claude Code reports that autocompact is thrashing. To avoid this:
+
+- Make Claude Code assume the model's actual context window. DeepSeek's Claude
+  Code guide selects `deepseek-flash[1m]`; the `[1m]` suffix makes Claude Code
+  assume a one-million-token window. For other unrecognized model IDs, set
+  `CLAUDE_CODE_MAX_CONTEXT_TOKENS` to the window the provider documents.
+- If the endpoint forwards `tool_reference` blocks, set `ENABLE_TOOL_SEARCH=true`
+  to restore deferred loading. DeepSeek's Anthropic API guide does not list
+  `tool_reference` among its supported content blocks, so this setting may cause
+  its requests to fail.
+- To shrink the upfront catalog, add
+  `"REA_MCP_INPUT_SCHEMA_PROFILE": "compact"` to the `env` object of the
+  `mcpServers.rea` entry in `~/.claude.json` (`$CLAUDE_CONFIG_DIR/.claude.json`
+  when that variable is set), then restart Claude Code. Setup keeps this
+  override.
+
+The sizes above come from REA's own `tools/list`. A live Claude Code session
+against DeepSeek's API is unverified.
+
+## Kimi Code
+
+Setup registers `rea` in `~/.kimi-code/mcp.json`, or
+`$KIMI_CODE_HOME/mcp.json` when set, while preserving unrelated settings and
+creating a `.rea.backup` before updating an existing file. Kimi Code infers
+stdio from the `command` entry, so setup does not add a `type` field. The
+registration includes `startupTimeoutMs = 30000` and `toolTimeoutMs = 120000`.
+Kimi Code also discovers personal skills in `~/.agents/skills` and project-local
+`.agents/skills`, so no separate skill root is required; setup still installs
+and verifies the matching skill files through the normal shared-skill flow.
+Start a new Kimi Code session to pick up MCP and skill changes.
+
+## Pi
+
+For [Pi](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/docs/mcp.md),
+setup writes a `type: "stdio"` entry under `mcpServers.rea` in the user-level
+`~/.pi/agent/mcp.json` and installs the bundled skill under `~/.agents/skills`.
+Use `rea setup --client pi --dry-run --json` to inspect the plan, then
+`rea setup --client pi --yes --json` to approve it. Check the registration with
+`rea doctor --client pi --json`. Pi reads `mcp.json` as strict JSON, like its
+own loader: comments, trailing commas, a byte-order mark, or an empty existing
+file are invalid, and setup reports them instead of rewriting the file. Repair
+such a file to plain JSON first, then rerun setup.
+
+Setup follows Pi's `PI_CODING_AGENT_DIR`: unset or empty selects the default;
+`~` and `~/` expand to the home directory; absolute paths and file URLs select
+that directory; relative paths stay relative to the command's working directory.
+On Windows, native paths, `~\`, and Git Bash/MSYS, Cygwin, and WSL drive paths
+are normalized as Pi does. Run REA and Pi from the same working directory when
+using a relative override. Pi has no profiles: `OMP_PROFILE`, `PI_PROFILE`,
+`PI_CONFIG_DIR`, and OMP's `enabledServers`/`disabledServers` lists do not affect
+Pi registration. Doctor accepts an omitted `type` for stdio but reports a
+non-stdio type as invalid; `enabled: false` stays disabled.
+
+REA configures only the user scope. Pi also reads `.pi/mcp.json` in trusted
+projects; a same-name project entry can replace the user entry or override its
+enabled state and exposure. Check that file if the user registration is aligned
+but unavailable in a project. An extension that registers `/mcp` can replace
+Pi's built-in MCP support and use its own configuration instead. Run `/reload`
+in an existing Pi session after changing servers outside the session.
+
+## Qwen Code
+
+For Qwen Code, setup registers `rea` in the `mcpServers` table of the
+user-level `~/.qwen/settings.json`, or `$QWEN_HOME/settings.json` when configured.
+Setup and doctor follow Qwen Code's tilde and working-directory-relative home
+overrides. Qwen Code also discovers personal skills in `~/.agents/skills`,
+where setup installs the bundled workflow.
+
+## Antigravity
+
+Antigravity 2.0 discovers personal skills in `~/.gemini/config/skills`.
+Setup installs its selected skill copy there.
+
+## Gemini CLI
+
+Gemini CLI uses `$GEMINI_CLI_HOME/.gemini/settings.json` when that home override
+is nonempty, including its shared personal skills at
+`$GEMINI_CLI_HOME/.agents/skills`. Setup and doctor check MCP allow/exclude
+policies from system defaults, user settings, trusted workspace settings, and
+system overrides. Workspace trust follows `GEMINI_CLI_TRUST_WORKSPACE`,
+`GEMINI_RESTRICTED_MODE`, the effective `security.folderTrust.enabled` setting,
+and the selected `trustedFolders.json` (including
+`GEMINI_CLI_TRUSTED_FOLDERS_PATH`). Settings in an untrusted workspace do not
+participate. A workspace or system `mcpServers.rea` replaces the entire user
+entry; setup stops if the winning definition conflicts with its planned entry.
+Review those files manually: setup never edits administrator/workspace settings
+or grants workspace trust. Trust supplied only by a running IDE is outside this
+standalone CLI check.
+
+## Codex and Devin
+
+Codex treats an unset or empty `CODEX_HOME` as `~/.codex`. Devin honors
+`XDG_CONFIG_HOME` for its `devin/mcp_config.json` on non-Windows systems.
+
+## OMP
+
+For OMP, setup writes a `type: "stdio"` entry to the user-level
+`~/.omp/agent/mcp.json`. It follows `PI_CONFIG_DIR` and
+`PI_CODING_AGENT_DIR` (relative to the setup working directory), and the profile selected by `OMP_PROFILE` or
+`PI_PROFILE` (`~/.omp/profiles/<name>/agent/mcp.json`). Setup also removes
+`rea` from that file's `disabledServers` list, which would otherwise hide the
+registration. Doctor treats an `enabled: false` entry as active when
+`enabledServers` lists `rea`, as OMP does, unless `disabledServers` also lists
+it. Run setup under each profile that should load REA.
+
+## Hermes
+
+For Hermes, setup writes a `mcp_servers.rea` entry to `config.yaml`, preserving
+comments and unrelated settings. It resolves that file from `HERMES_HOME`, else
+from the platform default Hermes itself uses — `%LOCALAPPDATA%\hermes` on
+Windows, `~/.hermes` elsewhere — including any `HERMES_DATA_DIR_SUFFIX`. Hermes
+connects an entry unless `enabled` is false, so setup states `enabled: true`.
+
+Hermes personal skills install under the active Hermes home's `skills` directory.
+Setup, doctor and uninstall follow the root's `active_profile` selection, as
+plain Hermes CLI commands do. An explicit `HERMES_HOME` pointing to
+`profiles/<name>` selects that named profile directly. An invalid, missing or
+deleted sticky profile is reported instead of writing into the default home;
+repair the profile or run `hermes profile use default` before retrying.
+REA follows `HERMES_HOME` whitespace trimming, environment-variable expansion,
+and home expansion before resolving that profile. Setup previews the resolved
+configuration and skill destinations; doctor and uninstall use those same paths.
+
+## Qoder CLI
+
+For [Qoder](https://docs.qoder.com/cli/mcp-reference), setup writes
+`mcpServers.rea` into the user-level `~/.qoder/settings.json` (the
+`QODER_CONFIG_DIR` override is honored). Entries carry `command`, `args`, and
+`env`; stdio is the default transport. Qoder CLI discovers personal skills
+from `~/.qoder/skills` (following `QODER_CONFIG_DIR` when set), so setup
+installs its skill copy there.
+
+## Factory Droid
+
+For [Factory Droid](https://docs.factory.com/harness/mcp), setup writes
+`mcpServers.rea` into the user-level `~/.factory/mcp.json`. Entries carry
+`command`, `args`, and `env`. Droid discovers personal skills from
+`~/.factory/skills` ([skills](https://docs.factory.com/harness/skills)), so
+setup installs its skill copy there.
+
+## Trae
+
+For [Trae](https://docs.trae.ai/ide/add-mcp-servers), setup writes
+`mcpServers.rea` into the user-scope `mcp.json` under Trae's VS Code-style
+user directory (`%APPDATA%\Trae\User` on Windows,
+`~/Library/Application Support/Trae/User` on macOS, and
+`~/.config/Trae/User` elsewhere). Entries carry `command`, `args`, and
+`env`; stdio is the default transport. Trae discovers personal skills from
+`~/.trae/skills` (`%userprofile%\.trae\skills` on Windows), so setup
+installs its skill copy there.
+
+The user-level MCP paths follow the editor's VS Code layout. Trae's official
+documentation describes the raw JSON editor without specifying those paths;
+loading this registration in Trae remains unverified.
+
+## ZCode
+
+For [ZCode](https://zcode.z.ai/en/docs/mcp-services), setup registers `rea` in the nested `mcp.servers` table of the
+user-level `~/.zcode/cli/config.json`. Entries carry `command`, `args`, and
+`env` without a transport discriminator. ZCode discovers personal skills from
+`~/.zcode/skills` rather than the shared directory, so setup installs its
+skill copy there.
+
+ZCode reads the user-scope `~/.agents/mcp.json` fallback only while the native
+`.zcode` config defines no servers at all, with no merging. When setup writes
+the first native server into an empty table, it copies the fallback's
+`mcpServers` entries into the native table so they keep loading, and leaves
+the fallback file untouched. Existing REA environment overrides are preserved
+when refreshing its command. A malformed or unreadable fallback file fails setup instead of
+being silently suppressed; repair or remove it and rerun. ZCode disables a
+server with `enable: false` on its entry; doctor reports such a registration
+as stale, and rerunning setup removes the flag to repair it. Native configuration
+uses strict JSON; setup refuses comments, trailing commas and a BOM rather than
+writing a registration ZCode cannot load.
+
+## DeepSeek Harness
+
+DeepSeek Harness registers MCP clients through a Cordis YAML patch sequence
+rather than a `mcpServers` object. `rea setup --client dsh` appends (or
+replaces) the REA entry in the home patch `$DSH_HOME/cordis.patch.yml`
+(`~/.dsh/cordis.patch.yml` when `DSH_HOME` is unset), preserving unrelated
+patches and comments. `rea doctor --client dsh` checks that same entry id.
+`rea uninstall` removes it by scanning supported clients; uninstall takes no
+`--client` flag. The home patch applies to every Harness profile, so the
+registration is written once and no per-profile edit is needed.
+
+Setup, doctor and uninstall resolve `DSH_HOME` and `DSH_AGENTS_HOME` as Harness
+does: unset, empty or whitespace-only values use `~/.dsh` and `~/.agents`,
+respectively. Bare `~`, `~/` and `~\` expand to the user home; other relative
+values resolve against the command's working directory. Nonblank values retain
+their original spaces, and the agents root is independent of the Harness home.
+Run REA and Harness from the same working directory when using relative overrides.
+The written entry is:
+
+```yaml
+- insert:
+    - id: rea-mcp
+      name: "@deepseek-ai/dsh-mcp-client"
+      config:
+        serverName: rea
+        transport: stdio
+        command: /absolute/path/to/node
+        args: ["/absolute/path/to/rea/scripts/rea.mjs", mcp]
+```
+
+Setup preserves caller-owned Cordis expressions and replaces launcher fields
+with literal values that doctor can verify. It refuses to edit an expression
+that produces the entire REA config, or a group-targeted REA insertion. Use a
+literal config and a root-level insertion as shown above.
+
+For a global npm installation, the launcher is under
+`$(npm root -g)/rea-agents/scripts/rea.mjs`. Restart Harness after setup; it
+exposes tools as `mcp__rea__<tool_name>`. Run `rea setup --skill --dry-run`,
+review the destination, then `rea setup --skill --yes` to install the workflow
+in `~/.agents/skills` (`$DSH_AGENTS_HOME/skills` when configured); Harness
+discovers that directory, and its native `skill` tool can load
+`reverse-engineer-anything`.
+
+REA's optional [native client verification](testing.md#end-to-end-integration-and-golden-evidence)
+checks headless chat, skill loading, all tool input schemas and actual analysis
+using the installed Harness and a local model fixture. This establishes the
+Harness integration, with live DeepSeek API compatibility remaining unverified.
+See the [Harness MCP guide](https://github.com/deepseek-ai/deepseek-harness/blob/639ed015397290b3745d163aafe02ffee4aa3f84/packages/mcp/mcp-client/README.md)
+for its native configuration contract.
+
+## Other client settings
+
+For OpenCode, setup writes the V1 `mcp.rea` entry, which OpenCode V1 and V2
+both load. If the configuration already uses OpenCode V2's native
+`mcp.servers` table, setup registers REA there instead and replaces any earlier
+`mcp.rea` entry from REA. Setup honors `OPENCODE_CONFIG_DIR` and
+`OPENCODE_CONFIG`, preferring the later-loaded JSONC file when both JSON and
+JSONC exist. The ordinary XDG global configuration still participates when a
+custom directory is selected. Doctor checks the merged user registration; uninstall removes
+REA-owned entries from every applicable user file so an older entry cannot
+become active again. Malformed applicable files block setup before changes.
+
+Grok Build loads `[mcp_servers.rea]` from `$GROK_HOME/config.toml`, or from
+`~/.grok/config.toml` when `GROK_HOME` is unset or empty. Setup edits that server
+table, `[mcp_servers.rea.env]`, and a root `disabled_mcp_servers` entry that
+names `rea`. It sets `startup_timeout_sec = 30` and leaves every other name
+in that list. The shared skill installed under `~/.agents/skills` is already
+on Grok Build's skill path.
+
+Grok Bot (`grok_bot`) is detected from `~/.grokbot`, or from `SAND_DATA_ROOT`
+when that value is an absolute path. A relative or empty `SAND_DATA_ROOT`
+stays on `~/.grokbot`. That directory is not the connector store. Grok Bot
+keeps connectors in the signed-in account and runs them on its hosted
+computer. It does not import `mcp.json` from the data directory, and it does
+not attach a stdio server running on this machine. Setup does not call the
+account connector API, does not write a registration file, and does not report
+the data directory as aligned. Ask the Grok Bot chat to add a custom MCP
+server named `rea` that runs on the Bot's computer:
+
+```bash
+npx -y rea-agents@<version> mcp
+```
+
+Do not put credentials in that command or its arguments. `rea doctor --client grok_bot`
+reports this manual step. `rea uninstall` does not remove the account connector.
@@ -0,0 +1,538 @@
+# Client and agent verification
+
+Use these optional lanes to verify an installed client or evaluate an agent's
+REA workflow. [Testing REA](testing.md) provides the normal development commands
+and provider lanes. Each record below states its host, client version, model
+fixture, and remaining coverage; a native-client run with a loopback model does
+not establish live model-service acceptance.
+
+## End-to-end, integration and golden evidence
+
+The optional `verify:qwen-client`, `verify:pi-client`, and `verify:hermes-client`
+lanes require an installed native client. Select its executable with
+`REA_VERIFY_QWEN_COMMAND`, `REA_VERIFY_PI_COMMAND`, or `REA_VERIFY_HERMES_COMMAND`.
+They configure disposable profiles through the public REA CLI, preserve caller
+settings and backups, and exercise real stdio MCP calls and skill loading. The
+`call` mode checks complete catalog discovery and actual JavaScript Evidence;
+pass `-- chat` to check plain chat with REA enabled. Qwen also reads the full
+result that its client offloads, and Pi exercises default codemode execution.
+Use `REA_VERIFY_RUNTIME_ROOT` to select a production-only installed REA package.
+Receipts retain client versions, result digests, host coverage and owned-process
+cleanup. These lanes use a deterministic loopback model and do not prove live
+model-provider compatibility. Pi and Hermes remain POSIX-only. Qwen, Pi and Hermes
+bind both `HOME` and `USERPROFILE` to the disposable account. Qwen and Pi discover
+the shared personal skill through their default search paths without an explicit
+skill setting; Hermes discovers it through the selected `HERMES_HOME`.
+Set `REA_VERIFY_HERMES_STICKY_PROFILE=1` to exercise native Hermes selection of
+a named sticky profile with `HERMES_HOME` still pointing to its root.
+
+The Qwen lane also supports native Windows (verified with Qwen Code 0.25.1 and
+Node 24.19.0 for both `chat` and `call`). With Node 24.18.0, the same `call`
+journey reached full Evidence readback but the client aborted during exit with
+`UV_HANDLE_CLOSING`; that run remains a failure, not a compatibility claim.
+Select the npm install's
+`node_modules/@qwen-code/qwen-code/cli-entry.js` with `REA_VERIFY_QWEN_COMMAND`,
+because `execFile` cannot launch the `qwen.cmd` shim directly. A selected `.js`,
+`.mjs` or `.cjs` entry point runs under the verifier's Node. Windows path checks
+use native separators; persisted output is read through Qwen's real shell tool
+with a shell-independent encoded Node payload. Windows receipts report process
+lineage as unavailable when the verifier has no retained native Job Object lease.
+
+`verify:gemini-client` requires an installed Gemini CLI (verified with
+`@google/gemini-cli@0.63.0` on POSIX and native Windows); select it with
+`REA_VERIFY_GEMINI_COMMAND`. A selected `.js`, `.mjs` or `.cjs` entry point runs
+under the verifier's Node. On Windows, select the npm install's
+`node_modules/@google/gemini-cli/bundle/gemini.js`, because only a `gemini.cmd`
+shim reaches `PATH`.
+The optional lane uses native `GEMINI_CLI_HOME` discovery in an isolated
+Git project, checks setup plans, backups and idempotence, activates the installed
+personal skill, validates all forwarded input JSON Schemas against their declared
+dialect, forwards the complete REA catalog and checks full JavaScript
+Evidence for a Unicode path. Use `-- chat` for ordinary chat with REA enabled,
+or `REA_VERIFY_RUNTIME_ROOT` for a production-only installed REA package.
+The local fixture exercises the real Gemini API adapter, including its
+`untrusted_context` tool-result envelope. Token counts are synthetic; live
+Google API compatibility remains unverified. The lane disables
+the client's memory-based relaunch to preserve the caller's Node heap budget.
+
+`verify:opencode-client` requires an installed OpenCode (verified with
+`opencode-ai@1.18.35`); select it with `REA_VERIFY_OPENCODE_COMMAND`.
+The optional POSIX lane configures isolated XDG roots and `OPENCODE_CONFIG_DIR`,
+checks setup plans, backups, preserved JSONC comments and idempotence, activates
+the installed skill, validates all forwarded input schemas and verifies complete
+JavaScript Evidence for a Unicode path. JSONC is the default fixture; set
+`REA_VERIFY_OPENCODE_CONFIG_FORMAT=json` for JSON. Use `-- chat` for ordinary
+chat or `REA_VERIFY_RUNTIME_ROOT` for a production-only installed package.
+The native core runs with external plugins disabled through `OPENCODE_PURE`;
+the local OpenAI-compatible fixture has a caller-declared one-million-token
+context and synthetic usage. Skills use an explicit isolated `skills.paths`
+directory, so default OS-home discovery remains unverified. Live model APIs and
+native Windows are also unverified.
+
+`verify:claude-client` requires installed Claude Code (verified with
+`@anthropic-ai/claude-code@2.1.296` on POSIX and native Windows); select it
+with `REA_VERIFY_CLAUDE_COMMAND`. On Windows, select
+`claude.exe`: npm installs put only a `claude.cmd` shim on `PATH`, and the real
+executable is under `node_modules/@anthropic-ai/claude-code/bin`.
+The optional lane uses native `CLAUDE_CONFIG_DIR` discovery, an independent
+Git workspace, user settings and default built-in tools. It preserves caller
+preferences and backups, checks idempotent setup, waits through the native
+`WaitForMcpServers` tool when discovery is pending, and validates the complete
+catalog and forwarded input schemas. Call mode loads the personal skill through
+native `Skill`, checks its full body and verifies named-schema JavaScript Evidence
+for a Unicode path. Use `-- chat` for ordinary chat with REA enabled, or
+`REA_VERIFY_RUNTIME_ROOT` for a production-only installed package. The local
+Anthropic Messages/SSE fixture preserves native resource-hint envelopes in its
+request artifacts. Usage is synthetic; live Anthropic API and bare-mode
+personal-skill activation remain unverified by this lane.
+
+`verify:copilot-client` requires installed GitHub Copilot CLI (verified with
+`@github/copilot@1.0.95`); select it with `REA_VERIFY_COPILOT_COMMAND`.
+The optional POSIX lane uses native `COPILOT_HOME` discovery, an independent
+Git workspace, guarded setup plans, preserved registrations, backups and
+idempotence. Native `skill add` registers the isolated installed skill; call
+mode loads its full body, validates all forwarded input schemas and checks
+named-schema JavaScript Evidence for a Unicode path. When Copilot spills a
+large MCP result, the model requests native `view` with `forceReadLargeFiles`
+and validates the complete returned Evidence rather than its preview.
+Use `-- chat` for ordinary
+chat or `REA_VERIFY_RUNTIME_ROOT` for a production-only installed package.
+The native offline BYOK adapter uses a loopback OpenAI completions/SSE fixture
+and `gpt-5.4` model metadata; model inference and token usage are synthetic.
+Set `REA_VERIFY_COPILOT_MODEL` to examine another client model ID and
+`REA_VERIFY_COPILOT_WIRE_MODEL` for its backend model name (defaults to the client
+ID). Every actual HTTP request's model is checked and recorded separately.
+Both `HOME` and `USERPROFILE` point to the disposable account; native `skill add`
+still selects the installed skill directory explicitly.
+
+The same native client passed full-profile chat and analysis with the actual
+wire model `gpt-4.1` under a distinct BYOK model ID:
+
+```bash
+REA_VERIFY_COPILOT_MODEL=rea-gpt-4.1 REA_VERIFY_COPILOT_WIRE_MODEL=gpt-4.1 npm run verify:copilot-client -- chat
+REA_VERIFY_COPILOT_MODEL=rea-gpt-4.1 REA_VERIFY_COPILOT_WIRE_MODEL=gpt-4.1 npm run verify:copilot-client -- call
+```
+
+These runs retain full canonical input/output schemas, all tools, installed-skill
+loading and complete Evidence readback. They use the fixture's requested
+one-million-token BYOK prompt setting; effective capacity and live API acceptance
+remain unknown. With the built-in `gpt-4.1` model ID, the verified client still
+blocks the full input-schema profile before HTTP with
+`compaction_static_context_blocked`, even when requesting a larger BYOK prompt
+capacity. The effective capacity is unknown. The default verifier preserves
+the complete catalog and full schemas. To verify the existing compact profile:
+
+```bash
+REA_VERIFY_COPILOT_MODEL=gpt-4.1 REA_VERIFY_COPILOT_SCHEMA_PROFILE=compact npm run verify:copilot-client -- chat
+REA_VERIFY_COPILOT_MODEL=gpt-4.1 REA_VERIFY_COPILOT_SCHEMA_PROFILE=compact npm run verify:copilot-client -- call
+```
+
+Both native workflows passed with the complete tool inventory. The verifier selects the
+profile only in its disposable Copilot registration and records it in the
+receipt. Compact advertisements can reduce nested validation structure; the
+server still validates complete canonical inputs. This is a tested alternate
+workflow, not a fix for the full-profile client/model limit tracked in #1554.
+Live model APIs, native Windows and default OS-home skill discovery remain
+unverified.
+
+`verify:grok-client` requires installed Grok Build (verified with the official
+Linux x64 1.0.50 binary); select it with `REA_VERIFY_GROK_COMMAND`. This optional
+POSIX lane isolates `GROK_HOME`, an independent Git workspace and additional
+skill roots, disables foreign configuration discovery, guards setup writes,
+preserves unrelated registrations/backups and checks idempotence. Call mode
+loads the full skill through native `read_file`, queries all REA names and input
+schemas through native `search_tool`, and calls analysis through `use_tool`.
+If discovery reports `partial`, the native agent diagnoses REA registration and
+retries discovery. No startup-timeout override or fixed readiness delay is used.
+A larger Unicode-path fixture exercises native result offloading: the full
+retained Evidence is validated against its named schema, then native terminal
+queries recover the selected export facts, subject and full artifact digest in
+the next model request. This verifies useful artifact recovery without claiming
+that every offloaded graph fact enters model context. Native line-number and
+truncation envelopes remain in the request artifacts. Use `-- chat` for ordinary
+chat, or `REA_VERIFY_RUNTIME_ROOT` for a production-only installed package.
+The loopback OpenAI completions/SSE custom model declares a one-million-token
+context and synthetic usage. Live xAI APIs, native Windows and default OS-home
+skill discovery remain unverified.
+
+`verify:commandcode-client` requires installed Command Code (verified with npm
+1.79.2); select it with `REA_VERIFY_COMMANDCODE_COMMAND`. Auth, MCP tokens and
+sessions use the actual OS home, so provision a disposable POSIX account rather
+than overriding `HOME`. Set `REA_VERIFY_COMMANDCODE_ACCOUNT_HOME` to that account's
+actual home and create `.rea-client-verification` there with exactly
+`Disposable REA client verification account` followed by a newline. The lane
+owns that account's `.commandcode` configuration and shared REA skill. Build the
+runtime as the checkout owner, then run `node scripts/verify-commandcode-client.mjs`
+as the disposable account; the npm shortcut also needs a writable checkout.
+The verifier does not create accounts or install clients. It guards setup
+targets, preserves an unrelated registration and backups, checks idempotence,
+and verifies native default shared-skill discovery and complete activation.
+Default deferred-schema delivery stays enabled: the native prompt advertises
+every REA tool, exact-name `search_tools` returns every input schema, and a native
+shell query recovers the saved catalog's count and digest when the client spills
+it. Actual analysis of a Unicode-path fixture delivers full named-schema Evidence
+to the next model request. Use `-- chat` for ordinary chat or
+`REA_VERIFY_RUNTIME_ROOT` for a production-only installed package. The loopback
+OpenAI completions/SSE model is a keyless BYOK endpoint declaring a million-token
+context with synthetic usage. A synthetic account-key value satisfies the
+client's print-mode gate; native local-only mode refuses hosted API calls.
+Updates, telemetry and cron are disabled. Hosted authentication, live models
+and Windows remain unverified.
+
+`verify:omp-client` requires installed OMP (verified with the official Linux x64
+18.8.7 binary); select it with `REA_VERIFY_OMP_COMMAND`. This optional POSIX lane
+isolates the default agent, global config and XDG roots, guards setup targets,
+preserves unrelated registrations and backups, and checks idempotence. It
+compares native rejection of an invalid profile with REA's refusal to plan
+fallback writes. OMP's default `xd://` interface mounts the complete REA catalog
+as devices: call mode reads the complete installed skill and tool documentation,
+then dispatches analysis through native `write`. It checks full named-schema
+JavaScript Evidence with a Unicode path in the next model request. Device
+metadata and documentation are distinct from forwarding all JSON schemas as
+model functions. Native print-mode MCP readiness uses its defaults.
+Use `-- chat` for ordinary chat or `REA_VERIFY_RUNTIME_ROOT` for a production-only
+installed package. The loopback OpenAI completions/SSE model fixture declares a
+one-million-token context and synthetic usage. Live model APIs, native Windows,
+named-profile native execution and default OS-home skill discovery are
+unverified; skill discovery uses an explicit isolated custom directory.
+
+`verify:deepseek-client` requires an installed DeepSeek Harness (`dsh`;
+verified with `@deepseek-ai/dsh@0.2.0-rc.2`) and Git. Set
+`REA_VERIFY_DEEPSEEK_COMMAND` to select its executable. This optional POSIX lane
+seeds caller model settings and a stale Cordis registration in an isolated
+`DSH_HOME`, then runs public `setup --client dsh` dry-run, setup, repeated setup,
+and scoped doctor. It checks the backup and unrelated patches before booting
+Harness with the generated registration. After native execution it uninstalls
+through the public CLI and checks that only caller patches remain. `HOME` and
+`USERPROFILE` point to the disposable account. The native default
+`~/.agents/skills` search path discovers
+the installed skill without a `DSH_AGENTS_HOME` override. A separate Git root prevents inherited project skills from
+masking a missing installation. The default `call` mode checks skill loading,
+complete catalog discovery, the actual forwarded input schemas and regexes,
+and full JavaScript Evidence for a Unicode path. Use `-- chat` for ordinary
+chat with REA enabled, or `REA_VERIFY_RUNTIME_ROOT` for an installed package.
+The fixture uses Harness's custom OpenAI adapter and the one-million-token
+context capacity declared by its default DeepSeek Flash model; it does not
+establish live DeepSeek API or native Windows coverage. Receipts include
+request artifacts, result digests and owned-process lineage.
+
+The DSH acceptance journey also exercises public setup, approval, doctor,
+generated-launcher MCP startup and uninstall without an installed Harness.
+That journey proves REA workflow behavior, not native Cordis expression
+evaluation. Filesystem regressions retain tagged values and unrelated rows
+across alias dependencies and check refusal of unsafe YAML without mutation.
+
+Full E2E tests invoke the production command dispatcher and real providers,
+without fake launchers, runners or responses. `verify:keyed-archive` writes
+Foundation binary, XML, and integer archives using one run-owned Swift module
+cache. It checks CLI and separate stdio MCP parity, pagination, escaping-path
+rejection, an XML graph golden, and numeric meaning for safe controls, unsafe
+integers, boxed numbers, and integral-valued reals. `verify:asset-catalog`
+compiles source-owned colors with
+`actool`, invokes real `assetutil`, then checks CLI/MCP results, exact catalog
+digest, every raw metadata field, pagination and malformed input rejection.
+Neither artifact workflow requires Hopper or launches it. Both run in macOS CI.
+`verify:macos-bundle` needs only macOS with Command Line Tools. It compiles a
+source-owned app with `clang`: a versioned framework, XPC services, an app
+extension, a login item, a privileged helper, launchd plists, and a helper tool,
+signed ad hoc. It packs the app as a directory, a `ditto` ZIP, and an APFS DMG,
+then checks that `inspect-artifact` plus `project-apple-application-graph`
+report the same bundle anatomy for all three through the CLI, with stdio MCP
+parity. It also checks that the DMG is detached afterwards. The same app
+covers `trace-dylib-resolution`:
+
+- each resolution status and finding, with CLI/MCP parity;
+- for every traced image, dependencies, rpaths, and install names equal to
+  `otool -l`;
+- for the main executable and an XPC service, a predicted load order equal to
+  the images dyld actually loads under `DYLD_PRINT_LIBRARIES`.
+
+The lane also compiles executable/library pairs with empty embedded directory,
+versioned-path, and suffix settings. It checks present and removed dependencies
+through CLI and MCP, and compares actual `DYLD_PRINT_SEARCHING` diagnostics for
+root-level candidates. These cases distinguish an empty search directory from
+an empty versioned scan or a suffix that only repeats the original path.
+
+It runs in macOS CI.
+
+Prepare the native inspector before cases that measure producer output or exit
+behavior; keep startup deadlines and cancellation in distinct cases. Run real
+process-capture verification separately from package or build checks. On macOS,
+new npm processes can become token-unreadable after changing their display
+title and prevent verified cleanup during a capture's ownership sweep.
+
+MCP SDK transport tests with recording providers remain integration tests.
+They are useful for schema drift and failure projection but do not prove that
+Hopper, Ghidra or another substituted engine works. `verify:package` proves
+packaging/install behavior and fake-provider integration; use the corresponding
+real-provider lanes for engine claims. Packed-bridge checks verify shipped files and Python syntax.
+Real Apple dispatch and Interface
+Builder verifiers currently prove format integration through production readers.
+
+`verify:hopper` exercises an installed Hopper through the production stdio MCP
+server and CLI. It checks source-owned call chains, CFG edges, references and
+complete large inventories, then probes unnamed bookmarks, annotation clearing,
+batch validation before mutation, malformed addresses and regexes, segment-end
+partial reads, and synthetic file-offset rejection. Advertised schemas are
+validated in their JSON Schema dialect and successful replies are checked against
+their advertised output schemas. Navigation checks cover interior-object cursor
+readback and mapped-memory boundaries. Annotation probes verify invalid native
+text and duplicate destinations/names before mutation, preserve unselected label
+owners, and exercise explicit batch label swaps. Function locals retain observed
+names and stack displacements. Graph probes check symbol/interior-address parity
+and a source-owned recursive cycle; literal tracing preserves complete queries
+and whitespace. Disposable binary copies prove that switching and closing
+actually removes the native document, and that CLI byte, function-dossier,
+literal-trace results and invalid-address diagnostics agree with MCP.
+No provider is mocked in this lane.
+Real search probes cover Unicode names, literal metacharacters, case and regex
+modes, annotation cache invalidation, complete native fragments of long literals
+checked against byte reads, escaped UTF-8/control text, byte-preserving Latin-1
+decoding, and Hopper's UTF-16 symbol-name truncation boundary. It verifies native
+CallReference classifications across reference and dossier results, retains long
+string fragment metadata in dossiers, and exercises pathological regex deadline
+and cancellation followed by successful requests in the same native session. When the macOS
+Objective-C fixture is present, the lane also verifies native UTF-16 string objects
+and their inferred encodings against their actual bytes. Native terminal calls
+are checked across reference, instruction, assembly, block-range and procedure-length
+projections; block endpoints are normalized using actual native membership.
+Exact Objective-C names also exercise named CLI selectors for function, instruction,
+decompilation, reference and search operations. A literal `--help` trace query proves
+that selector data is preserved independently of global CLI flags.
+Unmapped annotation destinations and
+oversized later batch names fail before any earlier edit is applied.
+
+`verify:hopper:deadlines` checks the native client's optional request deadlines
+on macOS with an owned source-built fixture. Zero and elapsed deadlines must
+leave native comments unchanged, including when a synchronous progress observer
+delays timer dispatch. It also observes a short analysis deadline and verifies
+subsequent wire recovery and clean shutdown. Caller timeout does not interrupt
+Hopper's synchronous native operation. Socket boundary tests deterministically
+cover active and queued expiry, late replies and timer cleanup.
+
+The Linux demo lane remains a separate `verify:hopper:linux` command. Hopper
+6.4.2's ELF loader did not recognize the original 4,119-byte literal or three
+non-ASCII byte strings in the measured fixtures. The lane retains complete raw
+byte checks for those literals through source-owned symbols. Recognition of the
+original full-length literal is bound to its exact source address; its decoded
+object coverage remains explicitly unverified on Linux even when recognized.
+Absent non-ASCII recognition also reports decoded-object coverage as unverified.
+Recognized non-ASCII and control-text objects must still match their source
+bytes, encoding, filters and searches.
+A separate 548-byte ASCII literal requires native fragmentation, shortened
+provider display, complete decoded fragments and dossier provenance on Linux;
+the macOS lane retains the original full-length recognition requirement. These
+are observed fixture/build capabilities, not universal Hopper length limits.
+Its lifecycle checks attempt competing CLI launches for the active target and a
+different target, requiring rejection before either can change the owning
+MCP session's documents or procedures. Lease boundary tests also preserve
+unresponsive or malformed live endpoints while allowing confirmed stale
+sockets to recover.
+
+`verify:setup:linux` exercises installation on a clean supported Linux runner as
+root. Use a disposable VM: this lane installs Hopper and its Linux runtime
+dependencies through the system package manager. Set `REA_VERIFY_SERVER_COMMAND`
+to the CLI entrypoint being verified. The command builds and checks source-owned
+ELF fixtures before testing the read-only Hopper setup plan, approved
+installation, scoped doctor, real CLI/MCP analysis, rejection of a modified
+launcher, idempotent setup without requesting a reinstall, and owned child
+cleanup. Before installation and after teardown, a clean-runner process scan
+also rejects surviving Hopper demo helpers or Xvfb groups, including orphans
+whose launcher exited. This scan observes processes without signaling them.
+Hopper detection compares executable file identity through `/proc`, covering
+alternate or configured launcher paths even after an invoked symlink is removed.
+The dedicated runner must have no unrelated Hopper/Xvfb sessions. It requires
+the repository's pinned development toolchain and an
+existing host C compiler, `file`, `nm`, and `strings`. Use `verify:hopper:linux`
+instead when Hopper is already installed.
+
+`verify:hopper:fat` is a separate macOS lane requiring installed Hopper and the
+existing Xcode clang/lipo toolchain. It compiles arm64/x86-64 thin executables and
+one- and two-slice FAT32 containers, verifies exact/interior address mappings
+against bytes in the original files, checks CLI/MCP parity, and checks owned
+runtime cleanup. Source byte changes, removal, permission denial (for non-root
+callers), and nonregular replacement must retain native partial mapping facts,
+reject unverified original-file coordinates,
+and recover after restoration. Single-slice FAT cases also relocate the slice
+without changing its loaded bytes. `verify:hopper:fat64` additionally checks
+FAT64 preparation through Hopper's native Mach-O loader, source-container
+mappings, profile identity,
+malformed and ambiguous slice rejection, and temporary-image lifetime. Both
+lanes have been verified on Hopper 6.1.0-demo; this establishes REA's prepared
+FAT64 workflow, not native FAT64-loader support.
+Cross-architecture fixture compilation is not required by `verify:hopper`.
+
+Golden tests use immutable captured text inputs with producer/source provenance
+under `tests/fixtures/golden/`. Expected results are reviewed for the semantic
+claim; capture commands do not automatically approve new expected outputs.
+Do not call handcrafted utility output or synthetic binary builders real-data
+goldens. Keep unsupported binary layouts and malformed boundaries as targeted
+regressions until a real fixture establishes equivalent coverage.
+
+`verify:browser` exercises source-map failure isolation and expanded-output limits through the compiled CLI and stdio MCP with real Chrome. It also submits five 2 MiB source-map annotations, checks retained script identities and explicit map omissions, and closes its owned fixture target. Replacement and document-reset behavior has separate producer-boundary coverage.
+
+The raw Chrome fixture launchers use `--password-store=basic` so Linux startup
+does not wait for an unlock prompt from a reachable but locked desktop keyring.
+Such a stall leaves the initial page unloaded with empty or placeholder CDP URLs;
+it does not establish permission to capture that page. See
+[the keyring diagnosis](https://github.com/morluto/rea/issues/1384#issuecomment-6093781686).
+
+`verify:browser` also captures a source-owned noise canvas as a real PNG above
+8 MiB through the CLI and stdio MCP, with complete byte/digest parity and real PNG
+decoding. Its SDK client explicitly permits the larger inline JSON response;
+this lane does not establish large image-comparison request transport coverage.
+
+The scenario checks read locale, timezone, device scale and viewport dimensions
+from actual page JavaScript through CLI and stdio MCP DOM captures. They also
+verify that attached-page emulation is released after success, initialization
+failure, action failure and cancellation while the external target stays open.
+
+`verify:browser:network` is a focused real-browser lane for transaction identity,
+selected request/response bytes, binary and compressed responses, duplicate
+headers, credential and declared-secret redaction, redirects, streaming cutoff,
+CLI/MCP parity, and owned-profile cleanup. Set `REA_BROWSER_EXECUTABLE` to an
+installed Chrome-family browser. An optional script argument selects an already
+installed package's `scripts/rea.mjs` entry point for packaged-artifact checks.
+The complete `verify:browser` lane includes these same checks.
+
+After building, `verify:browser:dom` checks empty and HTML-whitespace form
+destinations against a native Chrome DOM-property oracle through CLI and stdio
+MCP. It requires `REA_BROWSER_EXECUTABLE` and accepts an optional installed REA
+entrypoint. The full browser lane includes the same public-adapter assertions
+before other fixtures navigate the selected page.
+
+`verify:browser:scripts` checks active script capture → exact-byte export →
+existing static JavaScript analysis through CLI and stdio MCP, including
+manifest readback, competing query variants, and resolved relative imports.
+It uses an installed browser and accepts an optional installed REA entrypoint.
+The complete `verify:browser` lane also exercises passive script export through
+both public adapters. See [website script export](website-script-export.md).
+
+`verify:browser:modules` compares CLI and stdio MCP traces against an independent
+real Chromium module-loading fixture: import-map scopes, package prefixes,
+null/backtracking rejection, query/fragment identities, repeated module instances,
+lazy and computed unknowns, exact source/map identities and no implicit refetch.
+Set `REA_BROWSER_EXECUTABLE`; the optional script argument selects an installed
+package entrypoint. The complete `verify:browser` lane and existing conditional
+Chrome CI include this verifier. See [module relationships](website-module-trace.md).
+
+## Agent evaluation and conformance records
+
+Evaluate native, JavaScript, managed and browser investigation tasks through a
+real local Codex CLI with:
+
+```bash
+npm run verify:agent
+```
+
+The lane runs scenarios sequentially in standalone Codex (`--no-daemon`) with a
+disposable HOME and CODEX_HOME under `~/.cache/rea-agent-evaluations/`. Set
+`REA_AGENT_EVAL_ROOT` to choose another parent directory outside the OS temporary
+directory: current Codex refuses helper aliases under `/tmp`, preventing shell
+commands and skill reads even if MCP works. Each owned run directory is removed
+after verification unless fixture retention is requested. The lane
+installs the packaged skill and MCP registration through public `rea setup`,
+rather than copying a skill into a project that also inherits the caller's
+skills and configuration. Only authentication is copied, when present, and it
+is removed even with `REA_AGENT_EVAL_KEEP_FIXTURES=true`. No user configuration,
+rules, plugins or history is copied; shell snapshots and additional agents are
+disabled. Existing API-key environment authentication remains available.
+The native missing-provider scenario preapproves only `open_binary` and
+`close_binary` in the disposable client so it reaches REA's actual provider
+boundary instead of stopping at Codex approval. This does not validate a real
+native engine or authorize arbitrary runtime execution.
+
+Select a smaller run with `REA_AGENT_EVAL_SCENARIOS`, a comma-separated list of
+scenario IDs. `asar-zh` exercises the packaged desktop rubric from a Chinese
+request. `javascript-module-view` exercises summary analysis, module paging and
+item inspection in a tree with unrelated modules; its routing and text checks
+remain heuristic, so review the retained transcript for selector choice and
+coverage claims. Set `REA_CODEX_CLI` to an installed Codex executable and
+`REA_AGENT_EVAL_MODEL` to an account-supported model. Use an external resource
+limit for the whole process tree when testing on a constrained host; a Node
+heap limit alone does not bound client and child-process memory.
+
+`missing-target` leaves desktop fixtures visible without selecting one in the
+request. It requires zero REA calls and a structured request for an app name or
+artifact path. Its `expectedFirstTool` is `null`, and
+`targetClarificationPassed` grades that limited routing outcome independently
+of analysis prose heuristics. Analyzing a nearby example cannot pass this case.
+
+The packaged desktop, parser-comparison, and managed scenarios assess a closed set of
+known fixture claims. Each final answer must be a strict JSON manifest containing
+exactly the requested claim IDs, values, Evidence IDs, and Evidence authority and
+confidence. Golden values are not included in the prompt. The evaluator checks
+exact values against both the fixture oracle and authenticated successful REA
+results; an Evidence ID by itself is insufficient. Missing, duplicate, extra,
+contradictory, incorrectly sourced, or unsupported claims fail the scenario.
+
+The desktop rubric covers the exposed bridge API and members, renderer and main
+IPC operations, and the resolved preload. It binds its Evidence to the exact
+packaged artifact path and SHA-256. The parser rubric checks the precise heading
+depth addition, discriminant, complete comparison counts, and the explicit limit
+on runtime semantics. Its comparison must link to the two delivered source
+analyses and use the requested module and export selectors. Boundary tests package
+the same source-owned desktop fixture and compare the same source-owned parser
+files through actual MCP tool results, then verify that fabricated answers fail.
+
+The managed rubric checks the exact public top-level type set and the CLI entry
+point's declaring type and method name. The exact type-set claim requires an
+unfiltered member inventory: `selection` must be null and the Evidence parameters
+must omit `type_names`. A selected inventory can still ground the entry-point
+claim. Both inspections must describe the freshly built DLL's SHA-256. Entry-point
+resolution additionally requires complete artifact and member metadata, matching
+non-null module MVIDs, and a unique method at the header's metadata token. These are
+static metadata claims, not runtime behavior;
+compiler-generated member counts, token values, and MVIDs are not golden answers.
+
+Per-scenario `factualCorrectness` is `passed`, `failed`, or `not_assessed` within
+the `configured_fixture_claims` scope. A pass establishes only the configured
+claims, not unrestricted factual correctness. Native, browser,
+navigation-context, and address-context scenarios have no factual rubric and
+remain `not_assessed`.
+
+The closed factual scenarios request complete producer Evidence on the initial
+inspection or analysis so their source selectors can authenticate all configured facts.
+Summary/page/item workflows are exercised separately by `javascript-module-view`;
+its heuristic gate does not establish full-producer factual correctness.
+
+All scenarios retain routing, workflow, validation, repetition, process-exit,
+and token-use gates. Configured factual scenarios additionally require a factual
+pass. Their answer-text heuristics are diagnostic and do not affect the gate.
+Validation counts include complete text-only REA `invalid_request` envelopes
+and failed SDK calls whose arguments violate the named current REA contract.
+Client approval rejections and provider unavailability remain distinct failures;
+truncated JSON previews are not parsed into invented error codes.
+Repeated `binary_session` observations separated by a successful `open_binary`
+or `close_binary` are treated as lifecycle verification. An unchanged retry,
+including one after a failed lifecycle call, still counts as repetition.
+Scenarios without a rubric retain the legacy text-heuristic gate.
+`answerTermCoverageMet` checks case-insensitive substrings,
+`epistemicCuePresent` checks keywords, and `finalCitesEvidence` checks only an ID's
+presence. These metrics can still accept fabricated prose and must not be read
+as factual assessments. Set `REA_AGENT_EVAL_TRANSCRIPT_DIR` to retain complete
+tool results and final answers for review.
+
+Report schema version 4 adds negative routing scenarios with nullable
+`expectedFirstTool` and a `targetClarificationPassed` outcome. Consumers must
+allow an expected absence of REA calls; this does not make an analysis scenario
+pass without its required tools. Version 3 changed the top-level
+`factualCorrectness` from a constant
+string to an assessment summary with status, scope, and assessed/passed/failed/
+not-assessed scenario counts. Scenario records include the factual assessment and
+configured claim IDs. The evaluation scope is
+`routing_workflow_and_configured_fixture_claims`. Update report consumers for
+these changes. The narrower heuristic field names introduced in version 2 remain:
+`answerHeuristicsMet`, `epistemicCuePresent`, `answerTermCoverageMet`, and
+`requiredAnswerTermGroups`.
+
+Regenerate the managed conformance manifest and Evidence completion ledger from
+live verification results, or check them for drift:
+
+```bash
+npm run evidence:generate
+npm run evidence:check
+```
+
+The records preserve unsupported and unverified coverage as explicit unknowns.
+Run the matching real-tool prerequisites described in this guide.
@@ -35,10 +35,22 @@ license](https://github.com/morluto/rea/blob/main/third_party/evmole/README.md)
  and whitespace. Decoded-byte SHA-256 identifies actual selected EVM bytes.
  Neither digest is the Ethereum Keccak code hash.
 - Full bytecode, upstream function offsets and original producer representation
-  are returned inline. Offsets are byte positions in decoded bytecode.
+  are returned inline by default. Offsets are byte positions in decoded bytecode.
 - Upstream also returns terminal metadata during selector recovery. Its reported
  values remain in `raw_result`; integer precision and non-text-key coverage are
  unknown in the JS binding. Full bytecode retains the original encoded metadata.
+- `bytecode.metadata` independently interprets a plausible terminal CBOR root
+  map from the selected bytes. Integers use decimal strings; byte strings use hex,
+  and scalar keys retain their types. The recorded offset/length addresses the
+  original bytecode, so summary delivery does not copy an opaque CBOR payload.
+  Nested containers have partial scalar coverage; unsupported or malformed
+  candidates report `not-decoded` with a diagnostic. Decoding is bounded to 64
+  container levels and the terminal payload's encoded size.
+- An unavailable trailer does not prove metadata absence or Solidity origin.
+  Compiler-version bytes/text are metadata declarations, not verified compiler,
+  source-file or deployed-contract identity. Historical omission parses as a null
+  unknown facet without changing the original Evidence record. Metadata/source
+  URLs are not fetched.
 - Selectors, argument strings and state mutability are **inferences**. Function
  names/signatures are not looked up. Optimized code may produce different
  mutability candidates; recovered types are not a verified source ABI.
@@ -47,6 +59,49 @@ license](https://github.com/morluto/rea/blob/main/third_party/evmole/README.md)
 - Analysis is offline: no target runtime execution, implicit chain/RPC/explorer
  lookup, wallet action, transaction broadcast or network request.

+## Selected interface delivery and retained views
+
+Existing CLI calls and MCP calls with omitted or `complete` detail keep the
+complete inline result. On a retaining MCP server, select `detail: "summary"`
+when the question needs every inferred function and its identity/uncertainty,
+without bytecode text or duplicated raw stdout:
+
+```json
+{
+  "name": "inspect_evm_interface",
+  "arguments": {
+    "path": "/artifacts/runtime.hex",
+    "encoding": "hex",
+    "detail": "summary"
+  }
+}
+```
+
+The result is distinct derived Evidence linked to its complete parent. Its
+`normalized_result.summary` preserves every producer-returned selector, decoded
+byte offset, dispatch, argument/mutability string including null, carrier/decoded
+identities, original limitations and explicit unknowns. Payload `evidence_kind`
+remains `inferred`. It preserves stderr/truncation and records stdout text length.
+Complete-within-view coverage covers returned candidates; discovery completeness
+remains unknown even for an empty or exhausted collection. Analysis still runs;
+delivery selection does not save interpreter work or change source parameters.
+
+Use `normalized_result.parent_evidence_id` with `inspect_analysis_view` source
+kind `retained-evidence`. EVM view kind `evm` pages `functions` by original row
+order, or `bytecode_hex`, `stdout`, `stderr` by UTF-16 code units. Hex pages preserve
+the original lowercase text without a prefix; two characters represent one decoded
+byte. A text-page offset differs from a function's decoded-byte offset. Complete
+parent Evidence remains accessible through the normal ledger/export workflow;
+reading views does not rerun analysis or change its identity. Existing record
+retention and close semantics are unchanged.
+
+Portable historical complete EVM Evidence can use source kind `inline` without
+a provider. A fixed server without Evidence retention rejects selected summary
+delivery before starting analysis and continues to support complete delivery.
+Discover the actual advertised schema before sending `detail` to an older server;
+use its complete operation when an equivalent workflow is required. The one-shot
+CLI keeps its existing complete result.
+
 ## Resource and lifecycle limits

 Inputs are limited to 4 MiB, complete replies to 16 MiB and combined diagnostic
@@ -82,3 +137,15 @@ The real verification lane traces representative raw/hex CLI workflows and check
 that only declared Node/prlimit launchers and process ownership inspection run.
 Socket traces permit local stdio socket pairs and reject attempted network requests.
 These observations verify those fixtures; they are not a general network sandbox.
+
+For the independent metadata parser, `npm run verify:solidity:metadata` uses an
+existing caller-selected solc 0.8.30 module:
+
+```sh
+REA_VERIFY_SOLC_MODULE=/absolute/path/to/solc/index.js npm run verify:solidity:metadata
+```
+
+It compiles the owned fixture with unoptimized, viaIR, content-hash-free and
+CBOR-free settings, verifies the observed trailer facts and reports local decode
+timings. No contract is executed. This lane does not establish the complete
+Linux EVM provider's CLI/MCP, resource-limit or cleanup behavior.
@@ -123,8 +123,10 @@ Select a verified native child for the existing tools:
 ```

 Then use `analyze_function`, `inspect_native_api` or the other native primitives.
-Native architecture and executable validation still apply. This milestone does
-not add MIPS/PPC/RISC-V decompilation, raw flash load profiles, bootloader/device
+Native architecture and executable validation still apply. Generic
+[MIPS ELF](ghidra-mips.md) and [RISC-V ELF](ghidra-riscv.md) lanes are separate
+documents with their own admission boundaries. This milestone does
+not add PPC decompilation, raw flash load profiles, bootloader/device
 tree semantics, vulnerability scanning, emulation or hardware runtime capture.

 ## Resource and lifecycle policy
@@ -0,0 +1,189 @@
+# Flutter build identification
+
+REA identifies a Flutter app's build facts by **pure parsing** of its APK:
+the Dart VM snapshot hash, per-ABI library digests, the engine's GNU
+build-id, and any labeled version strings. No engine, no tool installation,
+no execution, and no external lookups — this family answers "which Dart SDK
+build is this, and what exactly is in the payload?" before any deeper
+native analysis.
+
+This is the triage step for Flutter reversing: the snapshot hash is the
+artifact that maps to a Dart SDK release through an external dataset (such
+as the one reFlutter maintains), and knowing the SDK version is what makes
+SDK-layout-specific snapshot parsers applicable.
+
+## Answer an analyst question
+
+```sh
+rea identify-flutter-build /targets/Example.apk
+rea inspect-dart-aot /targets/Example.apk
+rea inspect-dart-aot /targets/Example.apk --abi armeabi-v7a
+```
+
+The equivalent MCP methods are `identify_flutter_build` (with `path`) and
+`inspect_dart_aot` (with `path` and an optional `abi`; the default prefers
+arm64-v8a, then x86_64, then x86, then armeabi-v7a, then the remaining
+ABIs alphabetically).
+
+## Optional retained AOT delivery
+
+MCP `inspect_dart_aot` keeps complete output when `detail` is omitted or
+`"complete"`. Explicit `detail: "summary"` returns the selected target/ABI/library
+identity, all snapshot headers, structural coverage, string-pool coverage and
+original lower-bound counts without the string lists. It retains complete source
+Evidence and returns a distinct linked `inspect_analysis_view` summary. Analysis
+and the caller's ABI choice are unchanged; one-shot CLI output stays complete.
+
+Use `normalized_result.parent_evidence_id` as the retained source for
+`inspect_analysis_view`, with `view: {kind: "dart-aot", collection: "package_uris",
+offset: 0, limit: 64}`. Collections are `package_uris`, `dart_uris` and
+`class_like_tokens`. Pages preserve the original strings and order, selected ABI,
+library identity and source coverage. `returned_counts` describe producer-returned
+rows; counts ending in `_at_least` remain lower bounds. Exhausting a page sequence
+does not make a partial pool complete or turn byte-pattern tokens into typed
+classes/functions. Views do not reopen the APK. Original Evidence remains
+exportable under its existing session lifetime. Summary requires a retaining
+server; older servers keep the complete request after schema discovery.
+
+## What the AOT inspection returns
+
+`inspect_dart_aot` reads one ABI's `libapp.so` in full and reports:
+
+- **Snapshot sections** — the four `_kDart*Snapshot*` symbols located
+  through the ELF **dynamic symbol table** (ELF32 and ELF64, little-endian,
+  every read bounds-checked), each with file offset, size, and — for data
+  sections, which carry the `f5f5dcdc` magic — the parsed header kind and
+  length (the stored length excludes the four magic bytes). Data headers
+  whose declared snapshot length does not fit their symbol retain their
+  observed fields and report partial coverage. Complete structural coverage
+  currently requires kind 3, the FullAOT value in the supported
+  [Dart 2.14 layout](https://github.com/dart-lang/sdk/blob/2.14.0/runtime/vm/snapshot.h).
+  Kind numbers vary across Dart releases; other values are preserved with
+  partial coverage, without guessing their interpretation.
+  Address resolution examines every overlapping `PT_LOAD` interpretation,
+  including partial overlaps. Equivalent mappings are coalesced; conflicting
+  mappings leave affected symbols unresolved and qualify coverage. Dynamic
+  hash, symbol and string tables use the same mapping checks.
+  A payload without exported snapshot symbols is refused rather
+  than guessed at.
+- **The snapshot hash** from the header at the start of each exported data
+  symbol. Complete hash coverage requires a valid header from both the VM
+  and isolate; duplicate records inside one symbol cannot stand in for the
+  other's header. Aliased or overlapping data symbols report partial coverage.
+- **A string-pool projection from the isolate data section**: distinct
+  `package:` source URIs (the app's own and its dependencies' Dart files —
+  a dependency inventory), `dart:` SDK URIs (which SDK libraries the app
+  uses), and identifier-shaped tokens, each bounded with counts and
+  `coverage: "partial"` when truncated. On the reference target this counts
+  756 package URIs and 201 dart URIs on arm64-v8a and x86_64, and 212 / 48
+  on armeabi-v7a, retaining up to 500 of each in the response.
+  `libapp.coverage` describes structural snapshot checks;
+  `string_pool.coverage` describes the bounded projection. Overall coverage
+  is partial when either facet is partial.
+
+## Interpret the result
+
+- **`flutter_detected`** is true when any `lib/<abi>/libapp.so` exists. A
+  negative answer reports complete coverage: absence is an answer.
+- **`snapshot_hash`** is the Dart build hash observed 20 bytes after the
+  snapshot magic (`f5 f5 dc dc`) inside `libapp.so`. Identification scans
+  byte patterns; AOT inspection reads the exported VM and isolate headers.
+  `snapshot_hash_sources` counts the valid records observed by the chosen
+  operation, and when they **disagree** the distinct values are
+  kept in `snapshot_hash_candidates` with `snapshot_hash: null` and
+  `coverage: "partial"` rather than silently picking one.
+- **`elf_header_valid`** checks the supported little-endian ELF32/ELF64
+  header, table entry widths, and full table extents against the library's
+  actual byte length. Truncated, malformed, or unsupported headers report
+  false and partial coverage while retaining byte-scanned hashes. This
+  header check does not establish valid table contents or exported symbols.
+- **`libflutter` facts** are per-ABI observations: byte size, SHA-256
+  digest, GNU build IDs from ELF-declared `PT_NOTE` / `SHT_NOTE` ranges,
+  and any _labeled_ strings the build
+  carries. Release engine builds often strip the `Dart VM version` string —
+  absence is reported as `null`, not guessed. Android clang/LLVM toolchain
+  lines are reported verbatim as build provenance.
+  `build_id_candidates` preserves distinct IDs. `build_id` is populated only
+  when all declared note ranges were examined and one ID was found;
+  conflicting IDs or malformed/bounded note metadata make coverage partial.
+  Matching bytes outside declared note ranges cannot supply a build ID.
+- Labeled strings must fit their complete grammar. `dart_version_capped`
+  and `toolchain_capped` independently report labels omitted by grammar or
+  projection bounds; either makes identification coverage partial. Values
+  split across read chunks are retained without publishing truncated prefixes.
+- **`coverage: "partial"`** also appears when an ABI carries `libapp.so`
+  without `libflutter.so` (or vice versa).
+- The APK itself is digested before parsing, so `target.sha256` binds the
+  observation to the exact input.
+
+## Limits
+
+- **Owned snapshots.** Both operations copy the admitted APK into a
+  provider-owned, read-only snapshot before digesting and parsing, so a
+  concurrently modified original cannot mix into one observation; failed
+  reader closure and snapshot cleanup are retained and retried on `close()`.
+  The snapshot is removed only after all owned readers and file descriptors
+  close successfully. Copying pins one admitted descriptor and compares its
+  identity, size, mode and timestamps after the copy. Source permission
+  failures remain `access_denied`, and source identity changes remain
+  `artifact_changed`. Provider-owned storage failures report a
+  `resource_constraint` with their native I/O cause. Cancellation is checked
+  during copying, metadata/symbol walks, string scanning, and after cleanup.
+- **Bounded reads.** APK snapshots are bounded at 2 GiB, including while
+  copying. Identification permits 768 MiB of cumulative library extraction;
+  declared sizes are checked before extraction. Library reads are bounded at 256 MiB each and ZIP
+  metadata at 8 MiB; the AOT image buffer is allocated at exactly the
+  declared size before reading. Exceeded budgets report
+  `resource_constraint`.
+- **Engine notes.** Engine libraries are spooled into the owned root. ELF
+  headers, note tables and note payloads share an 8 MiB metadata budget;
+  up to 64 distinct build IDs and 1024 distinct nonempty note ranges are retained.
+  The range cap bounds bookkeeping and separate positional reads even for tiny
+  overlapping declarations. Exact duplicates share one read and empty ranges
+  consume no inventory slots. Exceeding a bound preserves the candidates examined
+  from retained ranges and reports partial coverage with `build_id: null`.
+- **Metadata and candidate capacity.** Both operations refuse duplicate
+  relevant library paths. Inventory is limited to 32 ABIs and 1024 bytes per
+  relevant library path, independently of compressed size. Identification
+  refuses more than 64 distinct snapshot hashes with `resource_constraint`
+  rather than silently dropping conflicting evidence.
+- **Selected ABI.** AOT inspection applies the library budget to the selected
+  ABI after selection; an oversized unselected library does not block it.
+  Undefined dynamic-symbol imports are not treated as snapshot exports.
+  Section-stripped images require a bounded symbol count from `DT_HASH` or
+  `DT_GNU_HASH`; layouts without either are refused. Symbol names and their
+  terminators must fit their linked string table.
+- REA performs **no hash-to-SDK lookup**: mapping the snapshot hash to a
+  Dart SDK release requires an external dataset and is deliberately out of
+  scope for this family.
+- The AOT string pool is a **byte-pattern projection, not deserialized
+  cluster semantics**: typed class and function recovery requires an
+  SDK-specific parser (the snapshot layout changes per release — exactly
+  why the hash is the key), and `class_like_tokens` include incidental
+  matches. JEB's GUI Dart unit and native analysis of `libapp.so` cover
+  some ground today.
+
+## Real verification
+
+`npm run verify:flutter -- --apk PATH [--plain-apk PATH]` runs both tools
+through compiled CLI and stdio MCP, compares Evidence, and checks every ABI.
+Use `--expectations scripts/fixtures/flutter-gallery-expectations.json` with
+the Gallery 2.9.2 release APK to check exact per-ABI engine labels,
+URI counts and the snapshot hash against its pinned SHA-256. The verifier
+requires complete structural snapshot coverage on every ABI and explicitly
+reports the bounded pool projection as partial/unverified when truncated.
+Other APKs need their own expectations file to verify exact engine labels
+and URI counts; otherwise those facets are reported unverified.
+The optional plain APK checks negative detection. Record: Flutter Gallery 2.9.2 (112 MB, three
+ABIs) — every ABI reports the same agreeing hash
+`65817c30a78bb44c3dc3771876b6010a` from two snapshot sections, per-ABI
+engine build-ids, no labeled Dart version (as expected for a release
+engine), and the Android clang toolchain line; a plain aapt2-built APK
+reports not-detected with complete coverage. The suite uses self-contained
+synthetic ELF images and scanner payloads; it does not read private local files.
+
+The Gallery workflow and all 147 focused Flutter tests also pass on macOS Sonoma
+14.8.9 x86-64 and Windows 11 x64 build 26300.9457 guests with Node 24.18.0,
+including the conflicting-load and note-range-limit CLI/MCP regressions. Windows
+verification runs under a non-administrator account. Verifiers use a shared
+temporary parent so each CLI/MCP process can own and release its own snapshots.
@@ -0,0 +1,13 @@
+# Frida instrumentation
+
+REA exposes Frida device and process discovery, attach or spawn sessions, and caller-authored JavaScript instrumentation through the CLI and MCP. The Node bindings are an optional package dependency and load only when a Frida operation runs. REA does not install Frida server on a target.
+
+For remote devices, pass a host with an optional port (`host[:port]`; bracket IPv6 addresses) and any token, certificate, origin, or keepalive options on that operation. The keepalive interval is measured in seconds, matching Frida Core 17.23.2. Authentication values are held only for the active call or session and are excluded from Evidence. Do not embed credentials in the endpoint address.
+
+MCP exposes `list_frida_devices`, `list_frida_processes`, `start_frida_session`, `load_frida_script`, `resume_frida_session`, `unload_frida_script`, `frida_session_status`, and `close_frida_session`. A spawned process starts paused so scripts can load first. Closing unloads session scripts and detaches. If a spawn remains paused, REA resumes it only after script cleanup or session detach confirms the instrumentation is released; if both cleanup paths fail, the target stays paused and the cleanup error is retained. REA never kills the target. MCP sessions end with the server process. The one-shot CLI command `instrument-with-frida` accepts one inline script or script file, observes messages for a bounded duration, and returns Evidence with the source hash and captured messages.
+
+Failed startup cleanup retains acquired Frida handles, including a paused spawn that has not attached, for later cleanup retries. Its remote connection stays leased until session detach and target resume complete. CLI commands await provider cleanup before returning; MCP retains these resources for cleanup when its server closes. Server shutdown stops admitting new Frida operations and waits for admitted startup, discovery, and queued work before releasing their resources. A failed cleanup remains reported and can be retried by closing the server again, rather than treating an unregistered session as fully released.
+
+Frida scripts execute inside the selected target with that process's memory, filesystem, and network authority. Message capture is bounded to 16 MiB and reports truncation. Script source is not stored in Evidence; its selected path and SHA-256 digest are. Evidence retains non-authentication target selection such as device or endpoint, mode, program or PID, and spawn arguments. Authentication values remain excluded. Results report the observed session and messages, while unsupported platform metadata remains unknown. A request cancelled before a Frida operation starts has no provider effects. If cancellation arrives during a native Frida call, REA waits for that call to return, then closes any session it created and returns the observed session as a partial observation. A cancelled script load on an existing session releases only its new script before queued operations can run; the session and previously loaded scripts remain available. Once the provider commits a successful load, later request cancellation does not unload that script; use `unload_frida_script` to release it. If script cleanup fails, the script remains owned and a paused target cannot resume until the pending instrumentation is released. Queued script loads, resumes, and unloads check cancellation before dispatching native effects. Cancellation Evidence separately reports whether a target resume was observed, whether instrumentation cleanup completed, and that final target liveness remains unknown. REA does not kill the target. Remote discovery retains its observations and reports a separate cleanup error if the temporary connection cannot be removed.
+
+The [Frida Core API](https://frida.re/docs/frida-core-api/) documents device management, sessions, and script loading; the [JavaScript API](https://frida.re/docs/javascript-api/) documents target-side instrumentation. [Luma](https://luma.frida.re/) is Frida's standalone GUI; REA does not automate or control it. `npm run verify:frida` requires a Linux host with the optional Frida binding and uses a short-lived local Node.js fixture created by the verifier. Add `-- --address HOST:PORT --pid PID` to run the separate remote attach lane against a caller-selected target. Mock tests do not establish real device/server compatibility.
@@ -130,8 +130,9 @@ range's file offset and analysis address; it does not locate the first differing
 byte inside that range. Mapping ends are **inclusive**, unlike the exclusive ends
 from `list_segments`. Appended overlays remain in source-file identity and are
 excluded from initialized module coverage. Uninitialized allocations are reported;
-their contents and DOS runtime semantics are not verified. Formats other than MZ and explicit COM return
-`unsupported` with measured observations, not a successful MZ verification.
+their contents and DOS runtime semantics are not verified. Independent verification
+also covers explicit [raw-binary imports](raw-binary.md); other formats return
+`unsupported` with measured observations.

 `read_bytes` returns initialized provider memory, including loader fixups; an
 unmapped or uninitialized byte ends the read with an explicit returned length and
@@ -150,8 +151,10 @@ For a packed executable, importing the original can analyze its unpacking
 stub. Analyze a separately prepared unpacked copy to inspect the original
 program's recovered code. Retain both artifact digests and the transformation
 evidence; REA never silently replaces the admitted target. Appended overlays
-are separate from the initialized MZ module. Raw COM files and DOS extenders
-without an admitted MZ real-mode entry are outside this import boundary.
+are separate from the initialized MZ module. Headerless COM requires the
+[explicit COM interpretation](#headerless-com-images); REA does not guess it
+from the filename or unrecognized bytes. DOS extenders without an admitted
+MZ real-mode entry remain outside this import boundary.

 DOS MZ is available through Ghidra, with no Hopper support claim. It is outside
 the Windows native PE P0 boundary. macOS x64 DOS verification remains unverified.
@@ -3,7 +3,7 @@
 This lane extends native ELF analysis; it does not execute MIPS code or rehost
 firmware. It advances the generic-MIPS portion of
 [#718](https://github.com/morluto/rea/issues/718), not the whole firmware roadmap.
-PSP/Allegrex remains a separate specialization tracked in
+PSP/Allegrex has a [separate static-ELF profile](ghidra-psp.md) tracked in
 [#1330](https://github.com/morluto/rea/issues/1330).

 ## Admission boundary
@@ -13,13 +13,36 @@ retains class, byte order, `e_type`, raw `e_flags`, and inspected ABI declaratio
 Recognition does not mean every provider or MIPS variant is supported. Existing
 PE, managed and Mach-O architecture sets are unchanged.

-The first Ghidra lane admits ELF32 `ET_EXEC`, little or big endian, with explicit
-o32 and `EF_MIPS_ARCH_32R2` flags and no machine-specific or ASE encoding flags.
-These are header declarations, not proof that every instruction conforms to
-that ISA. ELF64, n32/n64, unspecified ABI, MIPS16/microMIPS, machine-specific
-variants, shared/relocatable objects and PSP PRX are not covered by this lane.
-Their family may be identified, but provider selection returns the specific
-unsupported constraint rather than guessing a compatible profile.
+The Ghidra lane admits ELF32 `ET_EXEC` and `ET_REL`, little or big endian, with
+explicit o32 and `EF_MIPS_ARCH_32R2` flags and no machine-specific or ASE
+encoding flags. These are header declarations, not proof that every instruction
+conforms to that ISA. ELF64, n32/n64, unspecified ABI, microMIPS,
+machine-specific variants and PSP PRX are not covered by this standard-instruction
+lane. MIPS16 shared images have the separate bounded profile below; other shared
+or compressed images remain unsupported. Their family may be identified, but provider selection returns the
+specific unsupported constraint rather than guessing a compatible profile.
+
+A second admission covers recordless legacy firmware images
+([#1620](https://github.com/morluto/rea/issues/1620)): ELF32 `ET_EXEC` headers
+declaring `EF_MIPS_ARCH_32` or `EF_MIPS_ARCH_32R2` with an unspecified ABI
+field, `EF_MIPS_ABI2` clear, and an _inspected_ absence of any ABI flags
+record (a target whose record was merely not observed stays refused). For
+these images o32 is an analysis assumption derived from the ELF32/ET_EXEC
+structure, recorded as `mips_abi_basis` in the profile and stated as a
+limitation in Evidence, never a declared fact; floating-point ABI semantics
+stay unverified. Ghidra's own header auto-selection picks a 64-bit language
+for exactly these files, so the lane pins `-processor MIPS:<LE|BE>:32:default`
+through a `provider-pinned-v1` profile selection and checks the
+bridge-reported `language_id` after import. A caller-selected
+`REA_GHIDRA_LANGUAGE_ID`/`REA_GHIDRA_COMPILER_SPEC_ID` still wins over the
+pin; combining the pinned profile with an environment override is a typed
+refusal. A header that declares an ABI field but carries no record is still
+refused. Two further boundaries apply: an inspection that could not examine
+every ABI declaration source (for example a carved ELF whose section table
+extends past the captured bytes without a program-header record) does not
+establish the recordless admission, and an `EF_MIPS_ARCH_32` header that
+_declares_ an ABI is refused rather than relabeled, because the declared-ABI
+lane is verified for `EF_MIPS_ARCH_32R2` headers only.

 Explicit `EF_MIPS_FP64` and `EF_MIPS_NAN2008` declarations are refused. Their
 absence is insufficient: Clang can emit identical `e_flags` for FP32, FP64 and
@@ -39,19 +62,125 @@ absent or supply a default ABI.
 The field layout follows
 [`Elf_MIPS_ABIFlags_v0`](https://github.com/bminor/glibc/blob/master/elf/elf.h).

-Provider admission requires record version 0, MIPS32r2, 32-bit GPR/CPR1, no CPR2,
-and a double-precision or FPXX ABI declaration with no ISA extension, ASE or
-unknown general flags. Missing records remain unknown rather than defaulting
-to a compatible ABI, and this initial lane refuses them. Unknown record values
-are retained for diagnosis but do not gain provider support. These checks bind
-the declared interpretation; the integer-only fixture does not establish full
+Provider admission on the declared-ABI lane above requires record version 0,
+MIPS32r2, 32-bit GPRs, no CPR2, no ISA extension, ASE or unknown general
+flags, and one of two floating-point declarations: 32-bit CPR1 with a
+double-precision or FPXX ABI, or no CPR1 with the soft-float ABI. Other
+combinations, including single-float, FP64 and unspecified ABIs, are refused.
+Missing records remain unknown rather than defaulting to a compatible ABI, and
+the declared-ABI lane refuses them; the recordless legacy admission above is
+the only path that admits a recordless image. Unknown record
+values are retained for diagnosis but do not gain provider support. These checks
+bind the declared interpretation; the fixtures do not establish full
 floating-point behavioral conformance.

+### MIPS16 shared images
+
+A separate profile admits ELF32 `ET_DYN`, either byte order, declaring explicit
+o32, MIPS32r2 and exactly the MIPS16 ASE. Machine variants, n32, FP64 and NAN2008
+remain refused. Source inspection must examine the complete dynamic relocation
+declarations through uniquely mapped file-backed `PT_LOAD` ranges. This profile
+verifies `R_MIPS_NONE` and symbol-index-zero `R_MIPS_REL32`; section declarations
+must describe that same dynamic table. Unavailable section headers and additional
+section relocation tables (including `--emit-relocs` output) remain unsupported.
+Symbol-bearing REL32,
+other relocation types, RELA and PLT declarations remain unsupported. A real
+negative control shows that Ghidra can normalize a symbol-bearing MIPS16 function
+pointer to an even address, losing the source mode tag. REA refuses that shape
+rather than rewriting the source or silently repairing the analysis result.
+
+The profile pins the endian-matched `MIPS:<LE|BE>:32:default` processor and default
+compiler specification, and checks the bridge-reported language after import.
+The generic processor preserves mixed MIPS32/MIPS16 code through per-address
+`ISA_MODE` context; a global `16e` processor would impose one mode across the
+image. A caller-selected incompatible processor or compiler specification is
+explicitly refused without replacing the requested choice.
+
+An observed ABI flags record must agree with the header and MIPS16 ASE, and
+must declare the same verified floating-point combinations as the standard
+lane. With no observed record, every ABI declaration table must have been
+completely examined. The header still declares o32, but floating-point ABI
+semantics remain unknown and are stated in Evidence; selecting the default
+compiler specification does not establish floating-point storage. Partial
+section tables do not prove record absence. The existing soft-float prototype
+policy applies when the source explicitly declares soft-float.
+
+Returned addresses are analysis coordinates, not runtime load addresses.
+`inspect_native_load_image` preserves original relocation bytes, current loaded
+memory bytes and provider status. External GOT bindings remain unverified
+without their defining images. The provider's `modified_sha256` describes its
+ephemeral relocated image and can differ from `original_sha256`; it does not
+mean that the caller's artifact was written. `inspect_native_instruction` retains the program
+language variant in `mode` and reports actual per-address `ISA_MODE` separately
+in `decoder_context`, including null when unavailable. Artifact bytes and digests
+remain unchanged; this workflow does not execute the shared library.
+
+The optional real-provider check builds eight source-owned GNU controls, covering
+both byte orders, mixed MIPS32/MIPS16 code, declared and absent ABI records, and
+the rejected symbol-bearing and additional section relocations. It compares source instruction bytes,
+function symbols and the local function-pointer tag with actual CLI/MCP results,
+checks context through a snapshot round trip, and verifies process cleanup.
+It requires caller-supplied GNU MIPS cross-binutils and Ghidra; no fixture target
+is executed or patched after linking. Floating-point behavior of compressed
+functions is not established by these integer controls.
+
+```bash
+REA_MIPS_AS=/absolute/path/to/mips-linux-gnu-as \
+REA_MIPS_LD=/absolute/path/to/mips-linux-gnu-ld \
+GHIDRA_INSTALL_DIR=/absolute/path/to/ghidra \
+npm run verify:ghidra:mips-shared
+```
+
 MIPS header and ABI record facts are committed in the analysis profile so
 interpretation changes invalidate profile-bound snapshots. The support-lane
 identity is versioned independently of file identity. Evidence, lifecycle
 outputs and saved snapshots preserve the `mips` family. Hopper and IDA adapters
-do not implicitly gain MIPS support. Windows Ghidra P0 remains PE x86/x86-64 only.
+do not implicitly gain MIPS support. Windows Ghidra P0 does not admit MIPS targets.
+
+### Soft-float prototype
+
+Ghidra's MIPS32 compiler specifications model hard-float o32 only: a typed
+`float` or `double` parameter is bound to `$f12`/`$f14` and returned in `$f0`.
+Soft-float o32 passes those values in integer registers and stack slots, so a
+typed prototype (for example from DWARF) would make the decompiler read FPU
+registers the code never writes and drop the real dataflow. That applies to
+functions, external functions and indirect calls through typed function
+pointers alike.
+
+For soft-float targets REA commits
+`mips_soft_float_storage: o32-soft-float-prototype-v1` in the analysis profile
+and runs `ReaGhidraMipsSoftFloat.java` after auto-analysis. The script encodes
+the program's own default o32 prototype, removes its floating-point parameter
+and return entries and the rules that consume floating-point storage, and
+installs the result as the program-specific `__rea_soft_float` prototype through
+a Ghidra specification extension. Deriving it keeps the byte-order-specific
+aggregate rules and the hidden struct-return pointer. The script then assigns
+that prototype, with dynamic storage, to every non-thunk function and external
+function, and to every function-pointer definition, whose declared types use
+`float` or `double`. Thunks follow their targets. Types are kept.
+
+The script fails, and writes no report, if the derived prototype still places a
+float in an FPU register or any reassigned function does. It records the
+function and definition counts last; REA requires that report exactly when the
+profile commits the policy and refuses the session otherwise, before serving any
+operation. Session Evidence carries the policy and counts as limitations.
+Callables without typed float signatures keep Ghidra's default prototype, under
+which Ghidra already infers integer-register parameters. Hard-float profiles do
+not gain the policy, so their digests are unchanged.
+
+### Relocatable objects
+
+`ET_REL` objects, such as Linux kernel modules (`.ko`), are imported by Ghidra's
+ELF loader, which assigns each allocated section an analysis address and applies
+the object's REL relocations there. Undefined symbols are placed in Ghidra's
+`EXTERNAL` block: a thunk for code, a label for data. These addresses are
+analysis coordinates, not the addresses a kernel or linker would choose, and
+session Evidence says so. `inspect_native_load_image` reports each relocation
+with Ghidra's status; HI16 entries report `PARTIAL` because Ghidra applies them
+together with their paired LO16. Relocatable targets commit
+`mips_support_lane: elf32-rel-o32-arch32r2-standard-v1`, separate from the
+executable lane, whose identity executable targets keep. Kernel modules are
+usually soft-float and therefore also use the soft-float storage policy.

 ## Use

@@ -70,7 +199,10 @@ REA's ELF parser.

 ## Verification

-The source-owned freestanding fixture is `tests/conformance/c/mips.c`. The
+### Executables
+
+The source-owned freestanding fixture is `tests/conformance/c/mips.c`; the
+soft-float variant also links `tests/conformance/c/mips-soft-float.c`. The
 optional cross-target lane requires a caller-supplied Clang with MIPS targets,
 LLD, GNU `readelf`, Ghidra and its compatible JDK; ordinary host-native checks
 do not acquire these tools. `REA_MIPS_CLANG` selects Clang and
@@ -81,7 +213,8 @@ These are test-time prerequisites, not REA runtime providers.
 GHIDRA_INSTALL_DIR=/absolute/path/to/ghidra npm run verify:ghidra:mips
 ```

-The lane compiles both byte orders and never executes the targets. GNU readelf
+The lane compiles both byte orders with hard-float and soft-float ABIs and never
+executes the targets. GNU readelf
 independently supplies header/ABI facts and symbol addresses, compared with the
 production REA target resolver before provider startup. Its version, bounded
 command output and selected tool path are retained in the verification report.
@@ -104,9 +237,44 @@ all program-header payloads. It checks the actual Ghidra language, entry-address
 decompilation, distinct artifact digest and missing-section limitations through
 CLI Evidence and an MCP-produced snapshot. Caller-selected Ghidra heap limits
 are passed to these child workflows.
+For soft-float fixtures the lane also checks, through `inspect_native_api`, the
+exact parameter and return types and storage of float, double, mixed (pair
+alignment and stack spill) and struct-returning functions; that a caller's
+decompilation receives every source float constant without FPU inputs,
+including an indirect call through a typed function pointer; and that the
+snapshot profile and Evidence retain the policy and outcome counts.
 Pseudocode presence is a liveness check, not proof of semantic equivalence.
 Instruction decoding and reader agreement likewise do not establish runtime,
 delay-slot execution or complete floating-point semantics. Assertion regressions
 reject incorrect byte order, constants, destinations and loaded identities.
 Parser/profile tests are narrower checks, not substitutes for this real-provider
 run. Report the actual lanes run and leave unexecuted verification pending.
+
+### Relocatable objects
+
+This lane has its own command and fixture:
+
+```bash
+GHIDRA_INSTALL_DIR=/absolute/path/to/ghidra npm run verify:ghidra:mips-relocatable
+```
+
+It compiles `tests/conformance/c/mips-relocatable.c` with `-c -G0` in both byte
+orders, plus a little-endian soft-float `-mlong-calls` variant matching Linux
+module options. GNU readelf independently supplies sections, symbols and REL
+entries. For every relocation in an allocated section, the lane reads the
+loaded word through `read_bytes` and requires it to encode `S + A`: `S` from
+Ghidra's own section placement (`list_segments`) or `EXTERNAL` placement, `A`
+from the original file bytes. HI16/LO16 are checked as the effective address
+each LO16 forms with each HI16 it resolves, including several HI16s before one
+LO16, so a missing signed-low carry fails and a HI16 without a following LO16
+fails. The lane also requires matching Ghidra relocation entries and statuses,
+section file offsets and sizes, coverage of a shared HI16, a LO16 resolving
+several pending HI16s (from an assembly block, since Clang does not emit that
+from C), a LO16 carry, external code and data and `R_MIPS_32` words, direct
+calls from the entry
+function, the relocatable support lane in the snapshot profile, the
+relocatable limitation in Evidence and an unchanged source object. The fixture
+also calls `powf`, an external whose float prototype comes from Ghidra's C
+library archive; every variant must decompile `powf(1.5,2.5)` without FPU
+inputs, and the soft-float variant must report that the soft-float prototype
+covered it.
@@ -17,10 +17,14 @@ submodule and is a separate, more capable workflow.
 The parser currently handles x64 PE32+ and RTR 9.1, including bounded
 DEHYDRATED_DATA processing, MethodTable relationships, virtual slots, and
 frozen strings. Other RTR versions and PE architectures are reported as
-unsupported or partial. The matching .NET 8.0.22 fixture is source-built and
-the parser's Windows-host Ghidra workflow has not yet had a real Windows-host
-run. This does not establish Windows Ghidra import or decompiler behavior.
-Windows Ghidra P0 has no database mutation authority. A PE/CLI or ReadyToRun
+unsupported or partial. The matching .NET 8.0.22 fixture is source-built.
+
+On a Windows host, Windows Ghidra P0 runs the parser without a JAR. Metadata,
+MethodTable relationships, virtual slots and frozen strings come from the
+parser report, and `analyze_function` decompiles slot targets. Tools that read
+the Ghidra database do not see rehydrated bytes (see below), because only the
+optional adapter writes them into the database and Windows Ghidra P0 has no
+database mutation authority. A PE/CLI or ReadyToRun
 assembly should use `inspect_managed_artifact`; it is not NativeAOT.

 ## Build and configure
@@ -57,7 +61,9 @@ returns `observations.metadata_recovery` inline from the captured source snapsho
 including the RTR header, MethodTable addresses, frozen strings, derived overlay
 identity, and examined/recovered counts. Its enclosing `unsupported` load-image
 status means independent PE load-image verification was not performed; it does
-not mean the metadata parser failed or the executable failed to load.
+not mean the metadata parser failed or the executable failed to load. Without
+the optional adapter, each recovered type's `type` is `null`, because no Ghidra
+DataType exists; select it by `address`.

 Pass a returned MethodTable address to `inspect_native_data_type`. A direct PE
 query also derives the snapshot metadata from the authenticated Ghidra session
@@ -67,8 +73,12 @@ The database result remains `unavailable` when Ghidra has no DataType at that
 address; its `metadata_recovery.source` is `read-only-derived-overlay` and
 carries related type, interface, slot, and target observations. Generated
 address identities do not claim original managed names. The parser does not
-merge frozen literals into Ghidra's database string inventory. Follow slot
-targets with `analyze_function` and ordinary cross-references.
+merge frozen literals into Ghidra's database string inventory: read them from
+`metadata_recovery.frozen_strings`, not `search_strings`. Slot targets come
+from the parser report; on Windows P0 without the adapter, a `read_bytes` of
+a slot in the rehydrated region returns no bytes. The Linux adapter route
+verifies a complete slot read and its target. Follow slot targets with `analyze_function` and ordinary
+cross-references.

 The built-in CLI path uses a MethodTable address returned by the parser or
 another observation:
@@ -129,3 +139,11 @@ compiler/linker symbols and raw directory bytes provide the oracle. Stripped
 fixtures exercise heuristic detection. Tests do not execute the resulting
 NativeAOT binaries. Fixture generation and adapter compilation are optional and
 never prerequisites of ordinary native analysis or setup.
+
+The real MCP lane runs only on a Linux x64 host with the adapter. The
+Windows-host behavior above was checked manually over MCP on Windows 11 x64
+(non-elevated, Ghidra 12.1.4, JDK 21, no JAR) with the source-built win-x64
+symbols fixture. The format, RTR version and frozen strings matched the fixture
+oracle; the derived MethodTable, its base type, its interface and a virtual
+slot reaching the overridden method matched the linker map; and pseudocode was
+readable. No CI lane repeats this check.
@@ -0,0 +1,90 @@
+# PowerPC ELF analysis through Ghidra
+
+This lane extends native ELF analysis to 32-bit big-endian PowerPC; it does
+not execute PowerPC code. It is the first milestone of
+[#1741](https://github.com/morluto/rea/issues/1741). GameCube/Wii DOL
+executables, PPC64 and other PowerPC variants are separate follow-ups.
+
+## Admission boundary
+
+ELF `EM_PPC` (20) is retained as the provider-neutral `powerpc` family. The
+target also retains class, byte order, `e_type` and raw `e_flags`. Recognition
+does not mean every provider or PowerPC variant is supported. `EM_PPC64` (21)
+is not recognized as this family.
+
+The Ghidra lane admits ELF32 `ET_EXEC`, big endian. ELF64, little-endian,
+relocatable and shared objects are refused with a specific unsupported
+constraint. `EF_PPC_EMB` declares an embedded variant and is refused; the
+remaining `e_flags` bits only declare relocation conventions and do not change
+instruction decoding.
+
+## Bound language
+
+Ghidra's ELF loader selects `PowerPC:BE:32:e500:default` for a plain `EM_PPC`
+header with `e_flags` 0. e500 is the embedded Book-E variant whose SPE unit
+replaces the classic floating-point unit, so it misinterprets ordinary
+750-family code. Nothing in the header distinguishes the two.
+
+The lane therefore imports with `PowerPC:BE:32:default` and compiler spec
+`default`, commits that choice in the analysis profile as
+`language_selection: "provider-pinned-v1"`, and checks the language Ghidra
+reports after import. A different language set through
+`REA_GHIDRA_LANGUAGE_ID` (and optionally `REA_GHIDRA_COMPILER_SPEC_ID`)
+replaces the lane default and is admitted against the installed language
+catalog as usual; for example `PowerPC:BE:32:e500` for genuine e500 code.
+
+PowerPC header facts are committed in the analysis profile so interpretation
+changes invalidate profile-bound snapshots. The support-lane identity is
+versioned independently of file identity. Evidence, lifecycle outputs and saved
+snapshots preserve the `powerpc` family. Hopper and IDA adapters do not
+implicitly gain PowerPC support. Windows Ghidra P0 remains PE x86/x86-64 only.
+
+## Use
+
+With a caller-installed Ghidra and compatible JDK on a supported Linux/macOS host:
+
+```bash
+rea analyze /absolute/path/to/ppc.elf --provider ghidra --json
+rea function /absolute/path/to/ppc.elf function_name --provider ghidra --json
+```
+
+MCP uses the existing tools, including `open_binary` with an absolute path and
+`provider_id: "ghidra"`. No PowerPC-specific tool catalog or emulator is
+introduced.
+
+## Verification
+
+The source-owned freestanding fixture is `tests/conformance/c/ppc.c`. The
+optional cross-target lane requires a caller-supplied Clang with the PowerPC
+target, LLD, GNU `readelf`, Ghidra and its compatible JDK; ordinary host-native
+checks do not acquire these tools. `REA_PPC_CLANG` selects Clang and
+`REA_PPC_READELF` selects GNU readelf (for example
+`/opt/homebrew/opt/binutils/bin/readelf` on macOS). These are test-time
+prerequisites, not REA runtime providers.
+
+```bash
+GHIDRA_INSTALL_DIR=/absolute/path/to/ghidra npm run verify:ghidra:ppc
+```
+
+The lane compiles the fixture for a 750-family core (`-mcpu=750`) and never
+executes it. GNU readelf independently supplies header facts and symbol
+addresses, compared with the production REA target resolver before provider
+startup. Its version, bounded command output and selected tool path are
+retained in the verification report. The parser deliberately recognizes only
+this fixture's GNU output, rejecting missing, ambiguous or unexpected fields
+rather than substituting defaults.
+
+Existing public MCP operations check Ghidra's **observed** language and
+compiler specification and loaded-file hashes; an e500 import fails the lane.
+The source-owned `ppc-probe.S` supplies fixed `addi` and conditional-branch
+encodings: verification checks bytes, length, immediate and direct destination,
+independent of decompiler variable names. Global and string bytes are checked at
+independently reported symbol addresses.
+
+CLI/MCP discovery and decompilation, direct calls, target/profile/Evidence
+identity, snapshot round-trip and final process cleanup remain covered.
+Pseudocode presence is a liveness check, not proof of semantic equivalence.
+Instruction decoding and reader agreement do not establish runtime or complete
+floating-point semantics. Parser/profile tests are narrower checks, not
+substitutes for this real-provider run. Report the actual lanes run and leave
+unexecuted verification pending.
@@ -41,6 +41,100 @@ does not prove text is absent. Both list and search Evidence carry this
 limitation. When an address is known, use `read_bytes` and `xrefs` to inspect
 bytes and references.

+## PE debug-symbol coverage
+
+The post-analysis handshake reports Ghidra's PDB Program Information attributes:
+the referenced file, GUID, hexadecimal age, legacy signature, version, and
+`PDB Loaded` state. Absent attributes remain null. PE Evidence carries a
+limitation when a PDB reference is present but Ghidra reports it as not loaded,
+or its load state is unknown. A bridge without this report leaves PDB coverage
+unknown. A reported loaded state is a Ghidra observation, not an independent
+PDB identity or complete symbol-coverage attestation, and retains an inline
+coverage caveat. Handshake PDB strings share a 4096 UTF-16-unit serialization
+budget (at most 24 KiB after JSON escaping); whole attributes that do not fit
+are omitted with an explicit resource limitation, rather than truncated.
+
+Original names and types may be missing even when auto-analysis completes.
+An empty name search does not establish that a function is absent; use
+`list_procedures` and address-based analysis to inspect recovered functions.
+REA imports a private executable copy and does not admit a caller-selected
+PDB companion. This reporting does not search symbol servers, copy a sibling
+PDB, or change analyzer settings.
+
+## Loaded addresses and file offsets
+
+`rea analyze <path> --provider ghidra --json` and MCP `binary_overview` report
+the observed loaded `image_base` in `normalized_result`. They retain each
+segment's reported `image_base` and `address_space`. A provider that does not
+report a base, or whose segments disagree, omits the overview field, following
+the existing contract on main. REA does not substitute zero or the first
+segment's address. All reported bases and address spaces remain visible in the
+complete segment inventory, including conflicting observations.
+
+`function` and most MCP procedure selectors use procedure names or
+provider analysis addresses. Ghidra `decompile` / MCP `procedure_pseudo_code`
+also accept source-file byte offsets. Existing symbols and analysis addresses
+take precedence; a `0x`-prefixed number that resolves to no procedure falls back
+to its observed source mapping. Use `--file-offset=0x...` on the CLI or
+`procedure: "file-offset:0x..."` through MCP to select file coordinates explicitly.
+Multiple mapping candidates fail as ambiguous rather than selecting a bank or
+address space. A successful conversion retains the file offset, mapped address
+and procedure entry in raw Evidence and `locations`.
+
+If an offset maps to loaded target bytes but no procedure starts at or contains
+the mapped address, the failure preserves both the original file offset and
+analysis address. This differs from an unmapped offset and does not recommend
+retrying the same offset syntax.
+If the function resolves but the backend fails or cancels decompilation, the
+failure retains the selector, mapped address and procedure entry without
+changing the backend failure code.
+
+Both mapping directions require a complete original `FileBytes` source whose
+SHA-256 matches the admitted target. Auxiliary files and partial source objects
+remain visible in the load-image inventory but do not establish offsets in the
+original target. Ghidra's object-local offsets are not silently treated as
+original-file coordinates. Original-source identity is checked once per source
+object per session; analysis-memory edits do not change the original bytes.
+
+ELF virtual addresses and file offsets are
+different coordinates. Ghidra can rebase a PIE ELF during import: in the
+source-owned verification fixture, ELF entry `0x1013` becomes Ghidra procedure
+`0x101013` with image base `0x100000`. That example is not a general conversion
+rule.
+
+Use `list_procedures` to select an observed procedure address. To relate file
+bytes to loaded memory, call MCP `inspect_native_load_image` or:
+
+```sh
+rea search <path> . --kind procedures --mode regex --provider ghidra --json
+rea inspect-native-load-image <path> --provider ghidra --json
+rea decompile <path> <observed-procedure-address> --provider ghidra --json
+rea decompile <path> --file-offset=0x1013 --provider ghidra --json
+rea address-to-file-offset <path> <observed-procedure-address> --provider ghidra --json
+```
+
+The load-image result preserves source-file indexes, file offsets, address
+spaces, and inclusive mapping ranges. Follow the mapping for the selected
+source file and region; adding `image_base` to an arbitrary file offset is not
+valid. For a linear byte-addressed, file-backed mapping, a file offset `F` in
+`[mapping.file_offset, mapping.file_offset + mapping.length)` maps to
+`mapping.start + (F - mapping.file_offset)` in the reported `address_space`.
+If several mappings cover the same file offset, keep those candidates distinct
+and select the procedure from the observed inventory. Uninitialized regions may
+have no file backing. ELF mappings are observed
+without independent format-specific load-image verification, as the result's
+status and limitations state.
+
+An unsuccessful numeric procedure lookup reports the current Ghidra image base
+and gives the CLI commands above alongside the corresponding MCP operations.
+Repeat any target-format, architecture, and entry-point options used to open the
+target. Other procedure operations keep their analysis-address semantics;
+file-offset fallback applies only to decompilation.
+
+Overview workflow revision 3 binds this output to snapshot identity. Older
+serialized overview Evidence remains readable; it is not replayed as a current
+overview merely because its source artifact is unchanged.
+
 ## Admitted function-analysis semantics

 | Concern                 | Ghidra contract                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
@@ -57,9 +151,10 @@ bytes and references.

 Function extents and instruction completeness refer to Ghidra's analyzed body.
 An inferred no-return call can terminate that body before subsequent bytes.
-Pseudocode, assembly, procedure info, instruction windows, and dossiers report
-observed terminal call sites in Evidence limitations, including resolved callee
-names, addresses, and FunctionManager flags when available. Instruction windows
+Pseudocode, assembly, procedure info, instruction windows, callee and caller
+reads, and dossiers report observed terminal call sites in Evidence limitations,
+including resolved callee names, addresses, and FunctionManager flags when
+available. Instruction windows
 and dossiers also retain these warnings in their result limitations. An
 unresolved callee or unestablished terminal-flow basis remains explicit. These
 observations qualify the analysis; they do not prove that a no-return flag is
@@ -103,3 +198,32 @@ It accepts the same architecture and DLL flags and verifies the selected PE role
 in the analysis profile.
 All lanes require matching native controls; package startup alone establishes
 only package compatibility.
+
+## Masked byte search
+
+`search_bytes` (`rea search-bytes PATH '48 8B ?? ?? 89' --provider ghidra`)
+searches initialized analyzed memory, including loader fixups. Supply hexadecimal
+byte pairs, with `??` for any byte. An optional equal-length hexadecimal `mask`
+selects bits: `pattern: "a0"`, `mask: "f0"` matches bytes from `a0` through `af`.
+Wildcards remain unconstrained even with an explicit mask. At least one bit
+must be constrained.
+
+Optional `start` and `end` addresses are inclusive and select one address space.
+A match must end at or before `end`, so `end` set inside a match excludes it.
+Without bounds, loaded memory blocks are eligible, in Ghidra address order.
+Blocks that are not loaded (file parts such as `.symtab`) and overlay blocks are
+listed in `searched.skipped_not_loaded` and are not searched; pass `start` or
+`end` in their address space to search them.
+Overlapping matches are included; matches never cross block boundaries.
+Each match reports its address, segment, unique source file offset (or null),
+and containing procedure (or null).
+
+Pages default to 100 matches and accept 1 to 1000. Pass `next_cursor` as `cursor`
+with the same target, profile, pattern, mask and bounds to continue. `complete`
+means no further matches remain. `searched.blocks` and `searched.bytes` describe
+the inclusive memory ranges examined for this page, including lookahead for the
+next match. Coverage can overlap between pages; do not sum it as unique memory.
+`searched.skipped_uninitialized` lists eligible uninitialized blocks on every
+page. These blocks are never searched. Cancellation returns a typed error,
+discards the active Ghidra database, and does not return a successful partial page.
+Hopper byte search remains unavailable until its implementation is verified.
@@ -0,0 +1,105 @@
+# PSP static ELF analysis through Ghidra
+
+This specialization builds on [generic MIPS support](ghidra-mips.md) for
+[#1330](https://github.com/morluto/rea/issues/1330). It reuses REA's existing
+CLI/MCP, Ghidra bridge, Evidence and snapshot workflows. Allegrex instructions,
+PSP loading and calling conventions remain owned by
+[ghidra-allegrex](https://github.com/kotcrab/ghidra-allegrex).
+
+## Initial boundary
+
+The initial profile is limited to Linux x64, little-endian ELF32 `ET_EXEC`,
+`EM_MIPS` and the PSPSDK declaration `e_flags=0x10a23001`. Those bits declare
+MIPS II, Allegrex, EABI32 and NOREORDER; they are not proof of every instruction's
+semantics. The profile also requires an inspected version-0 MIPS ABI flags
+record declaring ISA level 2/revision 0, 32-bit GPR and CPR1, no CPR2,
+single-precision hard float (`fpAbi=2`), no ISA extension or ASEs,
+`flags1=1` and `flags2=0`. These are the declarations independently reported by
+PSPDEV v20261001's GNU readelf 2.44 and its 24-byte section dump. Missing,
+uninspected, contradictory and other producer declarations are refused.
+The PSP profile is not an exception to the generic MIPS32r2/o32 policy; that
+separate gate is unchanged. Profile revision v2 records this corrected boundary.
+
+PRX, relocatable/shared images, PBP/ISO/CSO containers, runtime emulation and live
+code overlays are outside this first profile. Other hosts and flag combinations
+are not silently admitted. File names do not identify a PSP target.
+
+## Bring your own compatible extension
+
+Install the extension in the selected Ghidra distribution at
+`$GHIDRA_INSTALL_DIR/Ghidra/Extensions/ghidra-allegrex`, with its
+`Module.manifest`, `extension.properties`, language files and loader JARs.
+User-home installations are not searched because REA uses an isolated analysis
+home. REA never downloads, installs, rebuilds or vendors the extension.
+
+The published compatibility experiment used Ghidra **12.1.3**, extension
+**v21.4**, PSPDEV **v20261001** and JDK 21. New public-provider results must be
+reported separately from that historical bridge experiment. The extension's
+`version` property declares its compatible **Ghidra version**, not its release
+tag; the property must match the selected installation. Other same-version
+builds are not described as independently verified merely because they pass
+this metadata check. A malformed or unloadable extension still fails real import.
+
+REA fingerprints installed extension files before resolving the analysis
+profile and checks that identity again immediately before launch. Symlinks,
+nonregular files, missing components and incompatible metadata are refused.
+Inspection is bounded to 1,024 entries, 16 directory levels and 64 MiB, with
+streamed file hashing and 64 KiB properties. Rebuilds or content changes yield a
+different profile; an old snapshot is not silently reused with the new profile.
+This is content identity, not a signature or a sandbox for installed code.
+Do not modify the Ghidra installation while a session is running.
+
+## Ordinary public interfaces
+
+```bash
+rea analyze /absolute/path/to/psp.elf --provider ghidra --json
+rea function /absolute/path/to/psp.elf function_name --provider ghidra --json
+```
+
+MCP uses `open_binary` with an absolute path and `provider_id: "ghidra"`, then
+ordinary inventory and function operations. The loader is explicitly
+`PspElfLoader`, language `Allegrex:LE:32:default`, compiler `default`. The
+handshake must confirm the language and compiler before serving observations.
+Missing prerequisites return `PSP_EXTENSION_UNAVAILABLE` with an installation
+remedy; no provider fallback or automatic installation occurs.
+
+Evidence retains the `mips` architecture family and its PSP-specific profile,
+including extension content identity. It also carries the VFPU and static-only
+limitations. The extension does not model VFPU prefix effects completely;
+nonempty decompiled code is not proof of correct vector behavior. Existing
+session-local metadata editing does not change executable bytes or turn a
+snapshot into a persistent Ghidra project.
+
+## Verification
+
+Use the existing real MIPS verifier with its PSP option. Required tools are
+caller-supplied `make`, `psp-config`, `psp-gcc`, `psp-readelf`, `psp-objdump`,
+`PSPDEV`, `GHIDRA_INSTALL_DIR` and a compatible `JAVA_HOME`.
+
+```bash
+npm run build:cached
+node scripts/verify-real-ghidra-mips.mjs --psp
+# In a disposable installation without Allegrex, test actual public refusals:
+node scripts/verify-real-ghidra-mips.mjs --psp-missing-extension
+# Generic MIPS must remain generic with Allegrex installed:
+npm run verify:ghidra:mips
+```
+
+The generic verifier also exercises a section-truncated firmware fixture. That
+check uses generic MIPS symbols and language expectations and runs only in generic
+mode; the PSP report explicitly records carved-PSP coverage as not run.
+
+The maintained PSP fixture builds from `tests/conformance/psp/`, never executes,
+and is independent of commercial game data. PSP GNU tools supply ELF/ABI,
+raw ABI-record, symbol and encoding observations; REA's parsed target and saved
+snapshot must agree with the independent ABI declaration. Public CLI/MCP must
+recover known functions, a direct
+call, the immediate `0x1234`, BITREV, a global word and a marker. The verifier
+checks observed loader/language, original bytes, Evidence, snapshot reopening,
+and an A-B-A target switch with different global contents at the same address.
+A snapshot for A must not open B. Cleanup must finish before success is reported.
+
+These checks are selected static facts, not full behavioral equivalence, VFPU
+semantics or arbitrary PSP compatibility. Independently published PSP binaries
+are supplemental calibration targets after payload classification and license
+review; unsupported PRX/PBP must not count as a successful static-ELF case.
@@ -0,0 +1,105 @@
+# RISC-V ELF analysis through Ghidra
+
+This lane extends native ELF analysis for `EM_RISCV` targets. It does not
+execute RISC-V code, emulate a boot sequence, or rehost firmware. It addresses
+[#2063](https://github.com/morluto/rea/issues/2063) and follows the same
+lane shape as [generic MIPS support](ghidra-mips.md).
+
+## Admission boundary
+
+ELF `EM_RISCV` (243) is retained as the provider-neutral `riscv` family. The
+target also retains the raw ELF declarations that decide a lane: class, byte
+order, `e_type`, and `e_flags`. Recognition does not mean every provider or
+RISC-V variant is supported. Existing PE, managed and Mach-O architecture sets
+are unchanged. Headerless inputs use the separate [raw-binary admission
+boundary](raw-binary.md); raw RISC-V images are not supported.
+
+The Ghidra lane admits little-endian ELF32 and ELF64 `ET_EXEC` images that
+declare `EF_RISCV_FLOAT_ABI_DOUBLE` in `e_flags`, with no `EF_RISCV_RVE`,
+`EF_RISCV_TSO`, or any bit outside `EF_RISCV_RVC | EF_RISCV_FLOAT_ABI`. These are
+header declarations, not proof that every instruction conforms to that ISA.
+
+Provider selection returns the specific unsupported constraint rather than
+guessing a compatible profile:
+
+| Declaration                                     | Outcome                                                                                  |
+| ----------------------------------------------- | ---------------------------------------------------------------------------------------- |
+| Little-endian ELF64 `ET_EXEC`, double-float ABI | admitted, `RISCV:LE:64:default` + `gcc`                                                  |
+| Little-endian ELF32 `ET_EXEC`, double-float ABI | admitted, `RISCV:LE:32:default` + `gcc`                                                  |
+| Big-endian                                      | refused; Ghidra 12.1.4 ships no big-endian RISC-V language                               |
+| `ET_REL`, `ET_DYN`, other `e_type`              | refused with the declared value                                                          |
+| `EF_RISCV_RVE`                                  | refused; the embedded 16-register profile (x0–x15) needs its own processor specification |
+| `EF_RISCV_TSO`                                  | refused; the total-store-order memory model is unverified here                           |
+| Soft, single, or quad float ABI                 | refused; the pinned generic compiler specification models double-float                   |
+| Reserved or newer `e_flags` bits                | refused with the offending bits named                                                    |
+
+`e_type` is the only load-shape declaration this lane reads. A fixed-address
+`ET_EXEC` declaration does **not** establish static linkage: `PT_INTERP` and
+dynamic dependencies are not inspected, and the session records that boundary as
+a limitation rather than inferring a load model.
+
+## Why the language is pinned
+
+Ghidra's header-based selection is not reliable for RV32. The 12.1.4
+`riscv.opinion` pins `variant="default"` for the 64-bit constraint and leaves the
+32-bit constraint unpinned with the comment that the loader "must choose between
+'default' and 'AndeStar' variants". Measured against the installed release tree:
+
+| Fixture                                         | Ghidra auto-selection         |
+| ----------------------------------------------- | ----------------------------- |
+| RV64 ELF, `e_flags 0x5` (RVC, double-float ABI) | `RISCV:LE:64:default:gcc`     |
+| RV32 ELF, `e_flags 0x5`, no Andes declaration   | `RISCV:LE:32:AndeStar_v5:gcc` |
+
+So an ordinary RV32 image lands on an Andes processor model unless the language
+is stated. REA therefore pins `RISCV:LE:32:default` or `RISCV:LE:64:default`
+with the `gcc` compiler specification per class, records
+`language_selection: provider-pinned-v1`, and refuses a caller-selected
+`REA_GHIDRA_LANGUAGE_ID` or `REA_GHIDRA_COMPILER_SPEC_ID` that disagrees with the
+pin without changing the requested override. An override that matches the pin is
+accepted and recorded as `configured-v1`.
+
+## Evidence
+
+The lane identity is `elf32-exec-le-double-float-generic-v1` or
+`elf64-exec-le-double-float-generic-v1`, and `riscv_elf` retains `elf_class`,
+`byte_order`, `type`, and the raw `flags` alongside the read-only booleans
+`float_abi_double`, `rvc_declared`, `rve_declared`, and `tso_declared`. Those
+fields report declarations the header makes; none of them is evidence of the
+original compiler, of instruction conformance, or of a floating-point calling
+convention beyond the ABI the pinned language models.
+
+## Verification and limits
+
+Ghidra 12.1.4 PUBLIC with a 64-bit JDK 21+ is required, and on macOS the release
+tree ships no host decompiler binary, so `Decompiler/os/<platform>/decompile` and
+`sleigh` must come from your own build; see
+[installation](installation.md#ghidra).
+
+Run `npm run verify:ghidra:riscv` with an existing Ghidra 12.1.4 installation,
+JDK 21+, and a RISC-V-capable Clang, LLD and LLVM readelf. The verifier acquires
+no tools and never executes target code. Select installed tools explicitly when
+they are not on `PATH`:
+
+```sh
+GHIDRA_INSTALL_DIR=/path/to/ghidra_12.1.4_PUBLIC \
+JAVA_HOME=/path/to/jdk-21 \
+REA_RISCV_CLANG=/path/to/llvm/bin/clang \
+REA_RISCV_LLD=/path/to/lld/bin/ld.lld \
+REA_RISCV_READELF=/path/to/llvm/bin/llvm-readelf \
+npm run verify:ghidra:riscv
+```
+
+The lane builds the committed freestanding C and assembly sources into RV32 and
+RV64 `ET_EXEC` fixtures. Independent LLVM readelf observations establish ELF
+declarations and symbol addresses. Public CLI and stdio MCP workflows verify
+function enumeration, decompilation, pinned language/compiler identity, decoded
+probe bytes, global data, and direct calls. It validates advertised MCP output
+schemas and both JSON representations, saves and reopens snapshots, rejects
+header-mutated soft-float and unsupported high-bit declarations, and checks
+unchanged source bytes and owned cleanup. Failures retain the diagnostic
+workspace; successful runs remove it and report the observed process lineage.
+
+Module tests cover admission and malformed headers. Call-graph completeness,
+string inventories, dynamic linkage, Andes or RVE/TSO images, and big-endian
+targets are **not** established by this lane. `.riscv.attributes` ISA extensions
+are not inspected; the generic language pin remains an analysis convention.
@@ -0,0 +1,99 @@
+# Go binary build metadata
+
+`inspect_go_binary` / `inspect-go-binary` answers which Go compiler version,
+modules, replacements and build settings are embedded in one selected local
+executable or library. It reads the file without running it and requires no
+Go installation, active binary target, Hopper, Ghidra or IDA.
+
+```bash
+rea inspect-go-binary /absolute/path/to/application --json
+```
+
+The CLI resolves relative paths against the caller's working directory. MCP
+requires an absolute filesystem path and preserves that exact selection:
+
+```json
+{
+  "name": "inspect_go_binary",
+  "arguments": { "path": "/artifacts/application" }
+}
+```
+
+## Observations
+
+The inline Evidence result contains the artifact path, SHA-256, byte length,
+container format, architecture, word size and byte order. A recognized
+`build_info` record contains:
+
+- `go_version`, preserved as embedded UTF-8 text rather than forced into a
+  semantic version, or `null` when the embedded bytes are not UTF-8;
+- `go_version_bytes_base64`, preserving the exact compiler-string bytes;
+- exact file offsets and byte lengths for the header, compiler string and module
+  string;
+- `module_text`, the decoded module-information body, or `null` when the complete
+  body cannot be represented as UTF-8;
+- `module_bytes_base64`, the original module-string bytes including Go framing;
+- `module.path`, the embedded package path;
+- `module.main`, the main module's path, version, checksum and optional
+  replacement;
+- `module.dependencies`, with the same fields and their original order;
+- `module.settings`, retaining ordered key/value entries and duplicate keys;
+- `module.unparsed_lines`, `module.unparsed_line_bytes_base64` and
+  `module.complete`, so unfamiliar, malformed, conflicting or undecodable records
+  do not disappear behind a complete result.
+
+Missing checksums are `null`; empty fields and development/local replacement
+versions remain unchanged. Build settings such as `GOOS`, `GOARCH` or `vcs.revision`
+are reported when present. Their absence does not establish a build choice.
+Go strings may contain arbitrary bytes. Non-UTF-8 records remain available as
+base64 while other decodable module records are still reported; the reader does
+not insert replacement characters or reject the entire artifact for those
+strings.
+
+## Supported boundary
+
+The built-in TypeScript reader supports ELF, PE and thin Mach-O images with Go's
+inline or historical pointer build-info encoding. Source ranges refer to the
+original stable file snapshot, including pointer-backed strings. It does not
+scan arbitrary byte strings as compiler metadata or interpret a fat Mach-O
+container as one selected architecture.
+
+Historical moduleless Go binaries can store their empty module-string header in
+declared zero-initialized virtual memory rather than file-backed bytes. The reader
+recognizes this only when the complete optional module header lies in that memory
+without a conflicting file mapping. Its module contents are empty and
+`module_location` has zero bytes, anchored at the build-info field containing the
+module-header pointer. Compiler strings and other unmapped pointers remain strict
+file-backed reads.
+
+`build_info: null` means no recognized record was found in the inspected image.
+It does not prove that the binary is not Go. Stripping, old builds, deliberate
+removal and obfuscation can limit metadata. A malformed supported container or
+recognized build-info record returns an input-format failure. Unsupported
+containers receive a separate unsupported-target failure.
+
+Symlinks and nonregular files are rejected. The shared stable reader detects
+replacement and observed in-place changes during acquisition. Missing files,
+host permission denial, changed artifacts and cancellation retain distinct
+errors. The selected file is never modified; this inspection creates no
+temporary files, starts no subprocesses and accesses no network endpoint.
+
+Resource guards bound the selected file to 256 MiB, aggregate embedded strings
+to 1 MiB, structural table decoding to 16 MiB and the complete normalized metadata
+report to 8 MiB. Oversized results fail without partial success. MCP also applies
+its [response transport budget](mcp-contracts.md), which counts both text and
+structured content; a report within the metadata limit can still require export
+from retained Evidence. These limits are not operating-system memory or CPU
+isolation.
+
+Embedded metadata is evidence of file content. It does not authenticate a
+compiler, dependency, commit or supply chain, and it does not establish runtime
+behavior. Function names, source mappings, types and decompiled code are outside
+this tool's coverage. Use the [native workflow](native-investigation.md) when
+those questions require a disassembler.
+
+## Real producer verification
+
+`npm run verify:go:binary` exercises compiler-built fixtures through the public
+CLI and MCP, comparing their metadata with Go's standard `debug/buildinfo`
+reader. The verification lane requires Go; normal inspection does not.
@@ -0,0 +1,79 @@
+# Static HarmonyOS package inventory
+
+REA inventories HarmonyOS application packages through the shared hardened
+ZIP artifact reader in `src/artifacts/`, then projects the authenticated
+inventory into an execution-free application graph. This is the first
+increment of the HarmonyOS roadmap (#1431): package identification and
+component inventory only.
+
+Recognized root formats:
+
+- `.hap` — installable application module
+- `.hsp` — runtime shared package
+- `.app` — distribution App Pack (HAP/HSP children are listed by path). The
+  App Pack classification applies to verified-ZIP-magic root inputs only; a
+  nested ZIP member ending in `.app` inside another package, such as an IPA,
+  keeps the generic `zip` classification.
+
+HarmonyOS `.har` shared libraries are intentionally not classified by suffix:
+`.har` also names HTTP Archive JSON, so a Harmony shared archive requires
+verified ZIP bytes and is deferred with an explicit limitation.
+
+## Usage
+
+No external SDK, toolchain, or engine is required. Inventory the package and
+project the graph; CLI and MCP use the same workflow and return
+artifact-bound Evidence inline:
+
+```sh
+rea inspect-artifact /path/to/Example.hap --format json > inspection.json
+node --input-type=module -e 'import { readFileSync } from "node:fs"; console.log(JSON.stringify({ inventory_evidence: [JSON.parse(readFileSync("inspection.json", "utf8"))] }));' > projection-input.json
+rea project-harmony-application-graph projection-input.json --format json
+```
+
+The MCP tool `project_harmony_application_graph` takes the same
+`inventory_evidence` input. The projection reports:
+
+- exact component paths, artifact IDs, and hashes (manifests, Ark bytecode,
+  resources, native libraries, JavaScript, signing)
+- a packaging-model hint: Stage is claimed only from a `module.json` entry,
+  FA only from `config.json`, otherwise `unknown`
+- FA resources under `assets/<module>/resources/` as well as Stage `resources/`
+- unavailable component paths with unknown IDs and hashes, including encrypted entries; these make projection coverage partial
+- App Pack child package paths without recursive inventory
+- runtime-family hints (`ark`, `native`, `javascript`) from paths and formats
+- path-based N-API bridge hypotheses between bytecode and `libs/<abi>` native
+  libraries
+
+Every claim is path-presence only. Manifest semantics (module identity,
+abilities, entry points, permissions, dependencies), Ark bytecode contents and
+versions, resource resolution, signing verification, and observed native
+boundaries are out of scope for this increment and stay in the projection's
+`limitations` array until a dedicated HarmonyOS provider exists. REA fails
+closed on packages whose inventory cannot be authenticated.
+
+## Verification lane
+
+`npm run fixtures:harmony` downloads the SHA-256-pinned 1.0.1 HAP release
+from the MIT-licensed [harmony-next-pipeline source repository](https://github.com/ohosvscode/harmony-next-pipeline)
+into ignored `_reference/harmony-integration/`. The pin records the source
+revision, release asset ID, size, and digest. The downloaded artifact is not
+redistributed; bundled component licenses have not been independently audited.
+
+```sh
+npm run fixtures:harmony
+REA_HARMONY_TEST_HAP=_reference/harmony-integration/entry-default-unsigned.hap \
+  npm run verify:harmony
+```
+
+The lane exercises the compiled CLI and actual SDK-connected stdio MCP:
+`inspect_artifact` authenticity, projection determinism, Stage-model path hints,
+exact manifest/bytecode/resource component sets, empty native and bridge sets,
+complete-inventory coverage, and CLI/MCP parity. It checks that this fixture has
+no META-INF signing entries; that does not establish whether the package is signed.
+
+The release artifact's compiler/SDK versions and reproducible source build have
+not been independently established. This lane does not complete the pinned
+toolchain acceptance in #1431 and does not verify Ark decoding or signatures.
+Set `REA_HARMONY_ENTRYPOINT` to an installed package's `scripts/rea.mjs` to run
+the same checks against that package; the default lane uses the repository build.
@@ -0,0 +1,183 @@
+# Verify real HTTP/3 MITM captures
+
+This optional Linux laboratory checks native mitmproxy captures through REA's
+public CLI and MCP. It changes no runtime contracts or adapters and establishes
+only the specific producer, mode and clients exercised here.
+
+## Prepare a disposable guest
+
+The exercised image is [Omarchy 4.0.4, build 440](https://nightly.omarchy.org/stable/4.0.4/440/omarchy-4.0.4.qcow2),
+SHA-256 `43412b896fe3d730e373a51a8acf99343d7c3e6dee78aa9f9bd7fca181127264`.
+Download it without installing guest packages on the host, check its digest,
+and use a **new qcow2 overlay** with that file as a read-only backing image.
+The matching [boot artifacts](https://nightly.omarchy.org/stable/4.0.4/440/unpacked/)
+include the kernel, initrd, cmdline and upstream vm-boot recipe. Use a fresh SSH
+key, loopback forwarding and guest disk; do not reuse another investigation's
+writable overlay or host trust configuration. Shut the guest down before
+switching overlays. Preserve the image and boot artifact hashes with evidence.
+
+This sealed guest already supplies cURL 8.22.0 with OpenSSL 3.6.4, ngtcp2 1.25.0,
+nghttp3 1.18.0 and nghttp2 1.70.0. **No guest rolling package upgrade is needed.**
+Run `curl --version` and verify HTTP2 and HTTP3 features before starting. A stock
+cURL from another distribution may omit HTTP/3. For a different guest, building
+cURL requires a compatible TLS library, ngtcp2 and nghttp3; follow the
+[official build instructions](https://curl.se/docs/http3.html), install to a guest
+private prefix, and record exact source revisions and build inputs. That alternative
+has not been exercised by this laboratory; it is not the pinned reproduction path.
+
+Use a fresh VM overlay over a sealed Linux image. Use user-mode networking with
+SSH forwarding bound to host loopback; keep host directories unmounted. Install
+and build dependencies inside the guest, then remove its default route before
+generating traffic. Host trust, routes and installed packages stay unchanged.
+
+Record the image digest, REA source revision and these dependency versions:
+
+- REA's `.nvmrc`, `package.json#packageManager` and locked npm dependencies.
+- Official Linux x86_64 mitmproxy **12.2.3** standalone archive. SHA-256:
+  `2e95286b618fa6fd33e5e62a78c2e5112571d85f42ec2bac29b97ee242bdb5c5`.
+- A private Python virtual environment using the complete pinned graph in
+  `scripts/fixtures/http3-lab-requirements.txt`. Save `pip freeze` and build/download logs.
+- OpenSSL and cURL advertising HTTP3 and HTTP2 in `curl --version`.
+
+The verified guest used Python 3.14.7, cURL 8.22.0, ngtcp2 1.25.0 and nghttp3
+1.18.0. Other dependency profiles need verification. Back up the guest's
+`/etc/hosts` before adding `127.0.0.1 lab.test`. Reserve guest TCP/UDP loopback
+ports 18443 and 19443. These changes belong only to the disposable guest.
+
+From the guest REA checkout, use its pinned Node/npm toolchain:
+
+```sh
+npm ci
+npm run build:cached
+python3 -m venv /absolute/lab/venv
+/absolute/lab/venv/bin/pip install --no-deps -r scripts/fixtures/http3-lab-requirements.txt
+/absolute/lab/venv/bin/pip check
+/absolute/lab/venv/bin/pip freeze > /absolute/lab/python-requirements.txt
+```
+
+Before removing the guest default route, download the official mitmproxy archive
+and verify it, then extract to the guest's private tools directory:
+
+```sh
+mkdir -p /absolute/lab/tools
+curl --fail --location https://downloads.mitmproxy.org/12.2.3/mitmproxy-12.2.3-linux-x86_64.tar.gz \
+  --output /absolute/lab/mitmproxy.tar.gz
+printf '%s  %s\n' 2e95286b618fa6fd33e5e62a78c2e5112571d85f42ec2bac29b97ee242bdb5c5 \
+  /absolute/lab/mitmproxy.tar.gz | sha256sum --check --strict
+tar -xzf /absolute/lab/mitmproxy.tar.gz -C /absolute/lab/tools
+/absolute/lab/tools/mitmdump --version
+```
+
+Pin the REA revision with `git rev-parse HEAD` or archive SHA-256. Build using
+the exact `.nvmrc` and npm version; do not copy a different checkout's `dist`.
+For stronger offline dependency reproduction, preserve downloaded Python
+archives, their digests and built wheels along with the image and toolchain.
+The requirement versions pin the runtime graph, not every transient build tool;
+source builds and alternate interpreters are not guaranteed byte-identical.
+
+## Run the narrow profile
+
+Select absolute executables and a new evidence directory. The producer refuses
+to reuse a directory. No certificates are installed in a system trust store;
+clients and proxy receive explicit paths to freshly generated test CAs.
+
+```sh
+REA_MITMDUMP_COMMAND=/absolute/lab/tools/mitmdump \
+REA_HTTP3_PYTHON_COMMAND=/absolute/lab/venv/bin/python \
+npm run verify:web:http3 -- /absolute/evidence/run-1
+```
+
+The producer sends three concurrent binary POSTs on one HTTP/3 connection to the
+same URL through mitmproxy reverse mode to an aioquic HTTP/3 origin. The origin
+waits for all complete requests, then dispatches responses in reverse order of
+completion at the origin. QUIC stream send/receive orders need not coincide. A cURL HTTP/3-only
+request checks a second client implementation. Mitmproxy's real save addon
+writes `normal.mitm`; no native flows are synthesized.
+
+A second proxy run disables HTTP/3. cURL `--http3` falls back to HTTP/2, reaching
+an HTTP/1.1 TLS test origin. `fallback.mitm` must retain those lower protocols,
+so a successful HTTP request cannot masquerade as an HTTP/3 success.
+
+## Acceptance and evidence
+
+Each execution creates `verifications/<run-UUID>/` with its own diagnostics.
+`verification-latest.json` belongs to the most recently registered attempt and
+changes atomically from `running` to `passed` or `failed`. Registration and
+the ownership check/replacement use the same SQLite transaction lock. An older
+attempt finishing later keeps its own reports without replacing the marker.
+The `.verification-lock.sqlite` pathname must remain in place while attempts
+may run; SQLite releases the lock on connection close or process death.
+Lock acquisition waits for the current owner without a timeout or lease
+stealing. The verifier has per-command deadlines, but no overall run deadline;
+a live stalled lock owner can therefore delay publication until it exits.
+Only a successful attempt writes its own `verification.json`, after
+all cleanup and input-integrity checks. `result.json` also records failures and
+the original error separately from cleanup errors. Older successes remain in
+identified history; a legacy root `verification.json` is archived before a rerun.
+A killed verifier may leave `running`, which must never be interpreted as success.
+
+Capture inputs live in `captures/`. To inspect them again without new traffic:
+
+```sh
+REA_MITMDUMP_COMMAND=/absolute/lab/tools/mitmdump \
+npm run verify:web:http3 -- /absolute/evidence/run-1 --inspect-existing
+```
+
+Acceptance for the current run:
+
+- Every positive native record reports HTTP/3, UDP, ALPN h3 and successful TLS
+  timestamps on both legs. Connection IDs, producer TLS labels, exact TLS setup
+  timestamps and connection errors match the UUID-matched producer observation
+  on each leg. Endpoint handshakes independently observe QUIC v1.
+- Native flow UUIDs remain distinct. Three endpoint streams share a connection
+  on each leg. Origin dispatch reverses completed receipt. Client and capture
+  ordering is observed, while association is verified independent of their order.
+- CLI and MCP agree. Request/response bytes, lengths, SHA-256, ordered duplicate
+  headers, method, scheme, authority and path match independent observations.
+  Readable header names and values, request URL fields and ALPN also match;
+  non-UTF-8 header bytes retain a null display value rather than replacement text.
+- Actual MCP advertised schemas validate, Evidence preserves source digests and
+  provider provenance, and the complete returned Evidence matches the retained MCP session records.
+- The fallback capture reports client HTTP/2 over TCP and origin HTTP/1.1, with
+  cURL independently reporting version 2. Input files remain unchanged.
+- Every cleanup action is attempted on success and failure, and final lineage
+  contains no live descendants. Recorded producer groups are inspected separately
+  after token-verified cleanup, including after their launcher has exited.
+
+Retain captures, endpoint/producer JSONL, proxy diagnostics, cURL output, manifest,
+CLI/MCP output, advertised schema and session bundle. Export public certificates
+and dependency/source provenance separately. Do not commit evidence, private
+keys, dependencies or VM disks. Shut down the guest after exporting evidence.
+
+The reverse proxy can preserve authority `lab.test:18443` while recording origin
+port 19443. Mitmproxy's serialized flow/connection UUIDs are not stream IDs.
+Endpoint stream IDs are scoped observations; this test invents no cross-leg
+mapping. Mitmproxy 12.2.3's `QUICv1` label alone does not prove negotiation: the
+fixture checks the negotiated `_quic._version` in pinned aioquic 1.2.0.
+
+Coverage is native capture, Linux reverse mode, QUIC v1 and these clients.
+Empty/absent/incomplete/redacted body semantics, cancellations, resets,
+certificate failures, malformed captures, HAR and qlog are separate increments.
+Local/WireGuard, browsers, QUIC v2, migration, 0-RTT, mTLS, replay, sustained load
+and loss tolerance remain unverified by this profile. Passing it does not
+establish general HTTP/3 support in REA.
+
+Official references: [mitmproxy protocol support](https://github.com/mitmproxy/mitmproxy/blob/v12.2.3/docs/src/content/concepts/protocols.md),
+[HTTP/3 implementation](https://github.com/mitmproxy/mitmproxy/blob/v12.2.3/mitmproxy/proxy/layers/http/_http3.py),
+[serialized HTTP state](https://github.com/mitmproxy/mitmproxy/blob/v12.2.3/mitmproxy/http.py),
+[aioquic configuration](https://github.com/aiortc/aioquic/blob/1.2.0/src/aioquic/quic/configuration.py).
+
+## Verifier regressions
+
+After building REA, run these inside the same isolated guest:
+
+```sh
+node --test scripts/lib/http3-verifier-lifecycle.test.mjs scripts/lib/http3-capture-assertions.test.mjs
+/absolute/lab/venv/bin/python scripts/fixtures/http3-lab.test.py
+```
+
+These cover different QUIC stream completion orders, stale success reports,
+multiple failed closes, separately owned process groups and partially acquired
+resources, plus producer TLS fact mismatches on both legs and altered readable
+fields with unchanged binary evidence. They test verifier
+semantics; they do not add HTTP/3 provider coverage.
@@ -1,9 +1,18 @@
-# IDA Pro provider
+# IDA provider

-REA adapts an existing [mrexodia/ida-pro-mcp](https://github.com/mrexodia/ida-pro-mcp) server into its native analysis tools. Install and configure IDA, Hex-Rays, Python, and upstream MCP using that repository's instructions. REA does not install, upgrade, activate, or license them, and does not bundle another IDA server.
+REA adapts an existing IDA MCP server into its native analysis tools. Existing IDA Pro profiles use [mrexodia/ida-pro-mcp](https://github.com/mrexodia/ida-pro-mcp); an optional native bridge for IDA Pro and IDA Free is described below. Install and configure your selected engine and upstream MCP using their instructions. REA does not install, upgrade, activate, or license them, and does not bundle another IDA server.

 Choose one lifecycle mode in an upstream MCP JSON registration. Set `REA_IDA_MCP_CONFIG` to that file, then select `ida` once through `REA_ANALYSIS_PROVIDER`, the CLI's `--provider ida`, or MCP's `open_binary.provider_id`. No provider startup occurs during discovery.

+## Supported upstream versions
+
+Each mode accepts one upstream tool profile, checked against `tools/list` before any analysis runs. Upstream's `main` branch has since moved to 2.0.0, which renamed the GUI tools and dropped `idb_close`, so it serves neither profile. Installing from that branch fails at startup with the missing tool names in the error diagnostics.
+
+| Mode     | Upstream profile    | Known working                                          |
+| -------- | ------------------- | ------------------------------------------------------ |
+| attached | legacy GUI tools    | 1.4.0                                                  |
+| headless | database supervisor | builds serving `idb_open`, `idb_list`, and `idb_close` |
+
 ## Analyze an already-open GUI target

 The attached profile supports upstream **1.4.0 legacy tools**, including `get_metadata`, `list_functions`, and `decompile_function`. Copy the upstream registration's `command`, `args`, and optional `env` into a JSON file. An existing `mcpServers` object with an `ida-pro-mcp` entry is also accepted. For example:
@@ -32,6 +41,34 @@ Open the **original input binary** in REA, not an `.idb` or `.i64` database. Its

 REA's close operation releases its MCP connection or proxy. It never saves or closes the existing GUI database. External edits remain visible because results are live and are not replayed from snapshots. These checks do not lock the GUI or prove that the database is unpatched.

+## Optional native IDA MCP bridge
+
+IDA Pro and IDA Free can use the separate [ida-free-mcp native plugin](https://github.com/0xshlomil/ida-free-mcp) through an explicit attached registration. Install and start that plugin using its upstream instructions, open the original binary in the IDA GUI, and wait for analysis. The native MCP 1.1.2 release binds `0.0.0.0` without a listen-address setting. Restrict inbound access with the host firewall or use an upstream build that binds only to loopback before starting it. REA's loopback URL restriction does not change the plugin's listener. REA does not install IDA, build the plugin, activate a license, or launch the GUI.
+
+```json
+{
+  "mode": "attached",
+  "protocol": "native",
+  "url": "http://127.0.0.1:13337/mcp"
+}
+```
+
+Set `REA_IDA_MCP_CONFIG` to this file and select `ida` using the same CLI/MCP commands above. Omitting `protocol` preserves existing IDA Pro registrations. Native mode requires `resources/read` for `ida://idb/metadata` and its recorded input SHA-256; it never substitutes a digest of an unrelated local file. The native resource's `path` describes the IDB in the upstream protocol, not the original input path. REA retains it and checks database identity and image base before and after requests.
+
+This experimental profile provides function/string inventory and search, function lookup, assembly, GUI pseudocode, callees, address xrefs, instruction windows, and function dossiers. Direct callers remain unavailable: native code/data xref labels do not prove call edges. It requires neither IDAPython nor idalib. Native headless registrations are rejected.
+
+Native string observations use upstream C-string rendering; other representations remain unknown. Literal searches request conservative candidates and apply REA's original case and Unicode matching rules. Regex searches retain the full scan because native byte-oriented regex semantics cannot replace REA's Unicode regex semantics. All candidate pages are examined, and upstream search errors remain failures.
+
+Native callee resolution batches unique queries and checks each reply's query identity. Results remain live: a recorded input hash does not establish database revision, so cross-request analysis caches cannot prove freshness. For focused questions, use pseudocode, callees, or instruction-window tools; a function dossier requests additional facets and returns more Evidence.
+
+IDA Free's cloud decompiler needs network access and may request consent. Follow its licensing and supported architecture limits; see [IDA Free](https://hex-rays.com/ida-free). Decompilation moves GUI focus and the upstream plugin closes existing pseudocode tabs before requesting fresh output. REA advertises UI/network effects and closes only its MCP connection, preserving the open database. Rendered pseudocode has upstream traversal limits and does not prove complete source recovery. Cloud/widget failures remain explicit failures; retry after the GUI finishes or dismiss its prompt.
+
+Native MCP 1.1.2 can navigate in the previously focused pseudocode tab when switching functions. Use upstream 1.1.3 or a source build including [the navigation correction](https://github.com/0xshlomil/ida-free-mcp/commit/6adf2c615a2a03661059673e9c136ef28d0dc778), which focuses the disassembly view **before** jumping to the requested entry. This REA profile does not repair stale GUI output. Verification must include two different functions before claiming reliable function switching. The initial Linux verification used a locally corrected 1.1.2 plugin; upstream 1.1.3 has not been verified by that run.
+
+Native tool observations retain actual tool names and complete JSON text. Upstream `structuredContent` can be a shortened preview even while text contains the full result. REA reads the full native representation and rejects missing/malformed text instead of treating previews as complete observations.
+
+The legacy producer moves the active GUI as a side effect: assembly and raw function instructions jump to the function, and successful pseudocode opens or reuses a pseudocode view. A function dossier does both. Native GUI decompilation also changes the active view. `binary_session` reports these operations with `effects.may_show_ui: true`; headless operations and the other attached operations report `false`.
+
 ## Give REA a target for headless analysis

 The headless profile supports upstream's **database supervisor** API: `idb_open`, `idb_list`, `idb_close`, explicit `database` arguments, and modern analysis tools. Legacy 1.4.0 `idalib-mcp`, which accepts one initial filename and serves SSE, does not implement this profile. Follow the [upstream headless instructions](https://github.com/mrexodia/ida-pro-mcp#headless-idalib-session-model) for an installation with the supervisor API.
@@ -56,10 +93,12 @@ rea function 'D:\samples\program.exe' main --json

 Headless REA and the configured supervisor must share the host filesystem and native path syntax. Run REA on Windows when the supervisor runs on Windows. The attached profile can span WSL and a Windows GUI because it binds recorded input hashes instead of asking the GUI to open a WSL path.

-REA copies the admitted binary into a private temporary directory, checks its digest, and requests an owned worker with `force_headless`. POSIX workspaces use mode `0700`; Windows uses the matching package's native private-DACL boundary on a local NTFS temporary directory. It injects the returned database ID into analysis requests. It verifies the active owned worker and input path rather than adopting a user database. On close or failed startup, it releases its session with `save: false`, verifies that it is absent from upstream session discovery, and removes its private input and sidecars. An unverified release retains the workspace and reports incomplete cleanup.
+REA copies the admitted binary into a private temporary directory, checks its digest, and requests an owned worker with `force_headless`. POSIX workspaces are created atomically with owner-only permissions, respecting the process umask; Windows uses the matching package's native private-DACL boundary on a local NTFS temporary directory. It injects the returned database ID into analysis requests. It verifies the active owned worker and input path rather than adopting a user database. On close or failed startup, it releases its session with `save: false`, verifies that it is absent from upstream session discovery, and removes its private input and sidecars. An unverified release retains the workspace and reports incomplete cleanup.

 One REA session owns one database. Stdio registrations default `IDA_MCP_MAX_WORKERS` to `1`; the explicit upstream `--max-workers 1` also bounds a supervisor. Database size and engine memory depend on the target and upstream; REA does not impose an engine memory limit. The original binary is never executed or overwritten. Optional `workspaceRoot` selects an existing absolute temporary-directory parent on the REA host. Ensure it has space for the input copy and IDA working files.

+When a completed `idb_open` response reports success or is otherwise ambiguous, but no active owned private worker can be verified, REA retains the private workspace even if `idb_list` appears empty. A later close can retry verification if the original owned worker becomes discoverable. A completed explicit `idb_open` rejection retains the existing no-worker cleanup path. This is a local admission/ownership safeguard and **does not** provide a process-termination receipt or change the legacy reaped-worker protocol.
+
 Headless cleanup uses the upstream database lifecycle. Abruptly terminating REA, its proxy, or the supervisor does not establish worker release. Prefer `close_binary` for MCP sessions; after an interrupted process, inspect upstream session inventory and retained workspace diagnostics before removing working files. This adapter does not claim Windows Job Object containment for the upstream detached workers.

 ## Connect an agent once
@@ -90,17 +129,19 @@ Use a package version that actually includes this adapter; repository main can b
 {"name":"close_binary","arguments":{}}
 ```

-Opening an MCP target starts its selected provider session; headless opening waits for auto-analysis. Set the agent's MCP request deadline long enough for the target. The registration's optional `timeoutMs` defaults to 300000 for upstream requests. Cancellation drains the current upstream request before cleanup; it does not force-stop GUI analysis. An open request that times out without a completed response retains the private workspace because worker inventory alone cannot prove the request has stopped. `binary_session` reports actual tool availability. Repeated doctor/setup calls are unnecessary when the registration already works.
+Opening an MCP target starts its selected provider session; headless opening waits for auto-analysis. Set the agent's MCP request deadline long enough for the target. The registration's optional `timeoutMs` defaults to 300000 for upstream requests. A freshly opened headless worker can briefly report `status: "busy"` from `server_health`; REA re-reads the health every 100 ms until the worker answers, the caller cancels, or `timeoutMs` has elapsed since the first read. Only expiry while the worker is still busy produces a retryable `provider_timeout`; no further read is started once the deadline has passed. That error names the busy tool of the last observation, or `unknown` when the worker reported none or a value that is not tool-name-shaped. Cancellation during the wait produces `cancelled`. Any other unexpected health payload, including a busy payload with an invalid shape, remains `unreadable_output`. Cancellation drains the current upstream request before cleanup; it does not force-stop GUI analysis. An open request that times out without a completed response retains the private workspace because worker inventory alone cannot prove the request has stopped. `binary_session` reports actual tool availability. Repeated doctor/setup calls are unnecessary when the registration already works.

 For guided registration, set `REA_IDA_MCP_CONFIG` before `rea setup --client <client>`. Setup includes that file reference in its reviewed plan and preserves it after approval; it does not copy upstream credentials into the agent configuration. `rea doctor --provider ida --json` validates the registration without starting IDA and explicitly leaves runtime compatibility and target binding for analysis.

 ## Coverage and evidence

-Both profiles support procedure/string inventory and literal or regex search, procedure resolution, pseudocode, assembly, raw function instructions, resolved callees, incoming address references, and a function dossier. Inventories and modern pseudocode/assembly follow upstream pagination. Upstream limits on modern xrefs and callees are retained as explicit truncation limitations. A transport-level truncated preview is rejected rather than reported as complete evidence; the error preserves upstream retrieval guidance.
+All profiles support procedure/string inventory and literal or regex search, procedure resolution, pseudocode, assembly, raw function instructions, resolved callees, incoming address references, and a function dossier. Inventories and modern pseudocode/assembly follow upstream pagination. Upstream limits on modern xrefs and callees are retained as explicit truncation limitations. In the modern headless profile, targets that are not proven x86-only use processor-neutral `callgraph` edges instead of the x86 instruction checks in upstream `callees` (mrexodia/ida-pro-mcp#549). REA requests depth zero: the pinned upstream emits the root's outgoing edges without scanning child functions, then REA batches function-name lookup for those addresses. The graph covers recognized function destinations only; imports and other non-function targets remain unknown. A server without `callgraph` returns an empty relationship list with an explicit coverage limitation. Graph and path workflows retain those limitations alongside observed edges. A failed or inconsistent graph reply remains an error. A transport-level truncated preview is rejected rather than reported as complete evidence; the error preserves upstream retrieval guidance.

-The legacy profile additionally supports direct callers. Its producer may report call-site or interior addresses; REA resolves the reported function symbol to its canonical entry. The modern profile does not prove direct callers from its code/data xref labels, so that capability is unavailable and the dossier marks it unknown.
+The modern headless profile also supports `resolve_containing_procedure` and `list_documents`. An address resolves through upstream `lookup_funcs` to the canonical entry of the function that contains it; the body stays unavailable like other IDA identities. When no function contains it, REA reads the database-scoped minimal `survey_binary` and reports `not_in_procedure` for an address inside one of its segments (ends are exclusive) and `outside_segments` otherwise. A missing, truncated or inconsistent survey fails the read instead of guessing a reason. `list_documents` returns the single admitted program path, so `inspect_address_context` and `trace_feature` need no explicit document. `trace_feature` marks a trace truncated, with a residual unknown for that address, when upstream caps the references it returned. The legacy and native profiles do not declare these operations.

-Complete body ranges, typed reference edges, CFG, referenced data, source locals, comment kinds, and database revision are unavailable. Empty unsupported dossier facets mean **unknown**, not observed absence. `binary_overview`, whole-symbol inventory, mutation, debugging, arbitrary Python forwarding, and GUI navigation are outside this adapter's coverage. Check session availability before composing a workflow that needs those facets.
+The legacy profile additionally supports direct callers. Its producer may report call-site or interior addresses; REA resolves the reported function symbol to its canonical entry. The modern and native profiles do not prove direct callers from its code/data xref labels, so that capability is unavailable and the dossier marks it unknown.
+
+Complete body ranges, typed reference edges, CFG, referenced data, source locals, comment kinds, and database revision are unavailable. Empty unsupported dossier facets mean **unknown**, not observed absence. `binary_overview`, whole-symbol inventory, mutation, debugging, arbitrary Python forwarding, and requested GUI navigation are outside this adapter's coverage. Check session availability before composing a workflow that needs those facets.

 `provider.version` identifies the REA adapter; the profile explicitly records `version_scope: "rea-ida-adapter"`. The MCP handshake's `serverInfo` is separate and does not establish IDA or Python distribution versions. Raw observations preserve producer metadata and output. Attached hashes identify the recorded original input; headless identity uses the digest-verified private copy and observed worker input path. Neither proves source-code recovery or an unmodified analysis database.

@@ -108,4 +149,10 @@ Complete body ranges, typed reference edges, CFG, referenced data, source locals

 Run `npm run verify:ida -- --target /absolute/path/to/program --procedure main` with a configured upstream server. It exercises production CLI and MCP tools, validates their output contracts and Evidence, checks original-input preservation and session cleanup, and keeps local target data out of its console summary. Supply `--package-root` to verify an extracted or installed package. See [testing](testing.md) for the distinction between adapter tests and real-provider evidence.

-The initial real workflows use stdio registrations for the legacy Windows GUI profile and a Windows x64 headless supervisor with IDA 9.3. Streamable HTTP is tested against a loopback MCP fixture with the pinned client SDK; a real IDA HTTP-supervisor workflow remains unverified. Linux/macOS headless engine operation, other IDA releases, architectures, and modern attached GUI tools require their corresponding real workflows before support is claimed.
+Native verification requires `--second-procedure other_function --second-marker unique_second_body_text`; these options remain optional for the other profiles. The marker must occur in the second function's pseudocode and be absent from the first; the verifier also checks that returning to the first restores its text.
+
+The verifier requires the resolved procedure in both inventory and procedure search, then compares address, assembly, pseudocode, instruction windows and relationship results with the CLI/MCP dossier. Supply `--expected-string known_literal` to check a known positive string inventory/search result and `--expected-xref 0xSOURCE` to check a known reference source address. Choose those oracles independently from the operation under test. Without them, string and xref correctness are listed in `unverified`; valid output shapes alone do not establish those behaviors. Callee parity does not independently establish semantic completeness. Native callee successes must report an explicit truncation marker; missing markers remain protocol failures.
+
+The initial real workflows use stdio registrations for the legacy Windows GUI profile and a Windows x64 headless supervisor with IDA 9.3. The headless workflow has also passed on macOS arm64 with IDA Professional 9.1 (build 9.1.250226) against the same upstream commit `c133c3853faa111a9b00ee615c013b720d0c4acd`, using a source-built arm64 Mach-O fixture; Mach-O C symbols carry a leading underscore, so the procedure name is `_rea_fixture_add`. Streamable HTTP is tested against a loopback MCP fixture with the pinned client SDK; a real IDA HTTP-supervisor workflow remains unverified. Linux headless engine operation, other IDA releases, architectures, and modern attached GUI tools require their corresponding real workflows before support is claimed.
+
+Native Windows verification uses IDA Pro 9.3 with the public two-function fixture and Windows `whoami.exe`, using upstream `c9d15d98e8a143fc74f6e14037c48f63889b6d6c` with a loopback-listener change. It covers production CLI/MCP parity, distinct-function switching, identity and connection cleanup. See the [correctness and performance measurements](https://github.com/morluto/rea/blob/main/experiments/analysis-efficiency/ida-native-windows.md). This follow-up did not run Windows IDA Free or native macOS.
--- a/Show More
+++ b/Show More