chore: combine trace and function evidence prototypes

* fix(ghidra): mark empty xrefs results as unresolved, not unreferenced

Ghidra 12.2 resolves no references to data addresses on some Mach-O
imports, and xrefs returned a bare empty list with nothing separating
'no references exist' from 'this provider resolved no references', so
callers concluded addresses were unreferenced.

projectExecution now attaches ghidraEmptyXrefsLimitation to an empty
xrefs result: the list reflects what this analysis resolved, not proof
the address is unreferenced. Non-empty results carry no caveat.

Covered by behavior tests for both branches; validated against
FM9-Edit-arm64 (expect: limitation) once the daemon session reopens.

* fix(ghidra): drop circular trace_feature cross-check from empty xrefs caveat

trace_feature calls xrefs internally, so recommending it as an independent
cross-check repeats the same lookup the caveat qualifies. Keep disassembly
as the single independent check.

* fix(ghidra): report actionable-reference absence and invalidate stale snapshots

Review follow-up on the empty xrefs caveat:
- The bridge filters synthetic entry-point references before observation, so
  "resolved no references" could be false; the caveat now reports that no
  actionable references were returned and names the omission.
- Drop the trace_feature cross-check: traceReferences observes xrefs per
  matched string, repeating the same lookup instead of independent evidence.
- Bump instruction_references semantics to v3 so analysis snapshots captured
  before the caveat replay no bare empty results; the profile commitment in
  the snapshot binding no longer matches.

* feat(ghidra): mark empty function-reference facets as unresolved too

Extends the empty-xrefs caveat to the sibling ReferenceManager consumers
review follow-up identified: procedure_references and analyze_function
reference facets built on the same observation path could still present
empty arrays as collected facts.

- ghidraEmptyFunctionReferencesLimitation attaches when
  procedure_references returns no references for the requested direction
  and when a dossier observes no references in either direction;
  callers/callees (function-manager derived, observed working on the
  reported target) stay caveat-free.
- One emptyReferenceCaveat owner replaces the inline xrefs branch.
- function_references semantics bump to v3 so snapshots captured before
  the caveat do not replay bare empty facets; raw profile digests
  recomputed.

* fix(ghidra): cover annotation and instruction reference facets in the caveat

Review follow-up: annotate_native_function returns a dossier built by the
same analyzeFunction collection and inspect_native_instruction builds
incoming_references with visitObservedReferencesTo; both could present
empty arrays as collected facts under the reported ReferenceManager
failure.

- annotate_native_function attaches the caveat when its dossier observes
  no references in either direction.
- inspect_native_instruction attaches it when incoming_references is an
  explicit empty array (null or omission stays unavailable, matching the
  field contract).
- Split the diagnostics describe to keep each block under the lint bound.
- Behavior tests cover attach and non-attach for both operations.

* fix(ghidra): name the honest cross-check limits in the reference caveats

Review follow-up: the motivating failure was a string address, so
'cross-check the disassembly at this address' pointed at nothing — data
and string addresses have no instruction to disassemble — and
inspect_native_instruction's incoming_references is built through the
same visitObservedReferencesTo path, so it was not independent either.

- xrefs caveat now states that no independent in-product cross-check
  exists for the empty result, that surrounding disassembly only helps
  at a code address, and that data or string addresses need an external
  disassembly scan (otool -tV adrp/add pairing).
- Function-facet caveat separates outgoing references (own disassembly
  is independent) from incoming references (no in-product check; scan
  other code externally).

* fix(ghidra): use the address-scoped caveat for inspect_native_instruction

Greptile review follow-up: inspect_native_instruction checks incoming
references to one exact address, so the function-scoped caveat misdescribed
the scope; an instruction can be unreferenced while its function has
references elsewhere, and the address may have no containing function.
Attach the address-scoped xrefs caveat instead.

* fix(ghidra): qualify each empty reference direction independently

Analyze-function and annotate-native-function results attached the
unresolved-references caveat only when BOTH reference directions were
empty, so a dossier with one populated direction presented its empty
direction as a collected fact.

Split the combined function-references caveat into direction-scoped
limitations and attach one per empty direction (analyze_function,
annotate_native_function dossier readback, procedure_references by its
requested direction).

* fix(ghidra): qualify empty outgoing instruction references

inspect_native_instruction results carry two independent
ReferenceManager facets: the incoming list and the outgoing references
built from instruction.getReferencesFrom(). Only the incoming list was
checked, so an instruction with a populated incoming list and no
outgoing references presented the empty outgoing facet as a collected
fact, and any populated incoming list suppressed the caveat entirely.

Each facet is now qualified independently: the empty incoming list
keeps the address-scoped caveat; the empty outgoing list attaches the
new instruction-scoped outgoing limitation (this stream is not
entry-point filtered, so it claims no entry-point omission). Truth
table covered in describe("Ghidra empty instruction reference facets").

* fix(ghidra): gate the instruction outgoing caveat on decoded status

Non-decoded inspect_native_instruction results (data, not-in-listing,
not-instruction-boundary, outside-memory) return before the bridge ever
queries instruction.getReferencesFrom(), so their empty references list
is an unobserved facet, not an empty ReferenceManager observation. The
unconditional outgoing caveat claimed the ReferenceManager returned no
references from an instruction that does not exist at that address.

The outgoing limitation now attaches only when status is decoded; the
address-scoped incoming caveat still applies per the empty incoming
list, matching the original motivating data-address case.

* fix(ghidra): dossier incoming omission and external outgoing recovery

Two unqualified cases remained in the function-reference caveats:

- The analyze_function/annotation dossier excludes incoming references
  whose source is inside the function's own body (a self-recursive call
  is enough), so the projected empty list can mean 'no outside source
  observed' or 'only in-body sources observed'. The dossier now uses a
  dedicated incoming limitation naming the exclusion, pointing at the
  outgoing edges/disassembly where in-body edges remain visible, and at
  procedure_references(incoming) - the same ReferenceManager stream
  without the in-body exclusion - as the in-product discriminator.
  procedure_references keeps the generic incoming caveat: its incoming
  stream is not in-body filtered.

- An external function has no local body, so the outgoing caveat's
  'function's own disassembly' recommendation was impossible. Records
  classified external now get an external-specific outgoing limitation
  naming the defining artifact and this program's call sites instead
  (procedure_references outgoing, analyze_function, annotation readback).

* fix(ghidra): qualify dossier incoming caveat observation

The dossier-specific incoming caveat claimed the ReferenceManager
observed no incoming references from outside the body, but
visitObservedReference (ReaGhidraBridge.java:3973-3974) drops
entry-point references before the body filter (1619-1621), so the
manager may have returned references the dossier never reports.
Reframe as 'returned no actionable' and name both omissions; correct
the recovery claim: procedure_references(incoming) lacks the in-body
filter but omits synthetic entry-point references as well.

.agents/skills/reverse-engineer-anything/references/evidence-workflows.md

@@ -73,6 +73,14 @@ Parent Evidence ID plus source index preserves same-address
 alternatives. Follow `coverage.next_offset` only as far as the question needs;
 exhausting retained rows does not establish complete program discovery.
 
+`batch_decompile` returns every body unless you select `detail: "summary"` when
+the connected tool advertises it. Use its parent Evidence ID with
+`view: {kind: "batch-decompile", collection: "items"}` for body-free rows and
+exact `where.indices`, `addresses`, `procedure_addresses` or `statuses`. Add
+`include_pseudocode: true` for several chosen bodies, or use
+`collection: "pseudocode"` with one row's `index` to page a large body. Batch
+limitations aggregate every item, so do not attribute them to one successful row.
+
 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.

.github/pull_request_template.md

@@ -106,7 +106,8 @@ limits, security/privacy/containment implications, and meaningful proof gaps. --
 
 ## Author readiness
 
-<!-- The human owner completes these declarations before requesting review.
+<!-- The human owner completes these declarations before requesting review,
+unless the PR author is exempt as listed in CONTRIBUTING.md.
 The PR author is the owner by default. A different owner must explicitly accept
 ownership on this PR.
 AI agents must leave these declarations to a human to dictate to them or there could be catastrophic consequences for many users of REA.

.github/workflows/ci.yml

@@ -807,6 +807,8 @@ jobs:
           tests/boundary/mcp/nativeDemangleMcp.test.ts
           tests/boundary/mcp/nativeMachoSummaryMcp.test.ts
           tests/boundary/providers/native/nativeSignatureTargetBinding.test.ts
+          tests/boundary/providers/native/nativeUiOffscreenWindow.test.ts
+          tests/boundary/process/nativeUiVerifierLane.test.ts
           tests/boundary/cli/nativeSignatureCli.test.ts
           tests/boundary/cli/nativeMachoArm64_32Cli.test.ts
           tests/boundary/cli/nativeMachoEncryptionCli.test.ts

.github/workflows/real-binary-layout.yml

@@ -55,12 +55,12 @@ jobs:
       - uses: ./.github/actions/portable-build-cache
         with:
           build_cache_artifact: ${{ inputs.build_cache_artifact }}
-      - name: Install ELF32 cross toolchains for source-owned fixtures
+      - name: Install AArch64 and ELF32 cross toolchains for source-owned fixtures
         run: |
           sudo apt-get update
           sudo apt-get install --yes --no-install-recommends \
-            gcc-i686-linux-gnu gcc-arm-linux-gnueabi gcc-mips-linux-gnu gcc-mipsel-linux-gnu \
-            libc6-dev-i386-cross libc6-dev-armel-cross libc6-dev-mips-cross libc6-dev-mipsel-cross
+            gcc-aarch64-linux-gnu gcc-i686-linux-gnu gcc-arm-linux-gnueabi gcc-mips-linux-gnu gcc-mipsel-linux-gnu \
+            libc6-dev-arm64-cross libc6-dev-i386-cross libc6-dev-armel-cross libc6-dev-mips-cross libc6-dev-mipsel-cross
       - name: Acquire exact upstream libraries in an isolated runner environment
         shell: bash
         run: |

.github/workflows/website-check.yml

@@ -21,6 +21,10 @@ jobs:
           python-version: "3.12"
           cache: pip
           cache-dependency-path: website/requirements.txt
+      - name: Set up Node.js for the website Worker checks
+        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
       - name: Install sharing-image dependencies
         run: |
           sudo apt-get update

.github/workflows/website-pages.yml

@@ -54,6 +54,10 @@ jobs:
           python-version: "3.12"
           cache: pip
           cache-dependency-path: website/requirements.txt
+      - name: Set up Node.js for the website Worker checks
+        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: .nvmrc
       - name: Install sharing-image dependencies
         run: |
           sudo apt-get update
@@ -75,10 +79,21 @@ jobs:
           PY
       - name: Configure Pages
         uses: actions/configure-pages@983d7736d9b0ae728b81ab479565c72886d7745b # v5.0.0
+      # upload-pages-artifact excludes dot directories, including .well-known.
+      - name: Package the website including agent discovery
+        run: |
+          tar --dereference --hard-dereference --directory website/_site \
+            --exclude=.git --exclude=.github -cf "$RUNNER_TEMP/artifact.tar" .
+          tar -tf "$RUNNER_TEMP/artifact.tar" \
+            ./.well-known/ard.json ./.well-known/ai-catalog.json \
+            ./.well-known/agent-skills/index.json > /dev/null
       - name: Upload the website
-        uses: actions/upload-pages-artifact@7b1f4a764d45c48632c6b24a0339c27f5614fb0b # v4.0.0
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
         with:
-          path: website/_site
+          name: github-pages
+          path: ${{ runner.temp }}/artifact.tar
+          retention-days: 1
+          if-no-files-found: error
 
   deploy:
     needs: prepare
@@ -130,14 +145,24 @@ jobs:
           PUBLISH_TARGET: ${{ inputs.target }}
         run: |
           python3 -u - <<'PY'
+          import hashlib
           import json
           import os
           import sys
           import time
           from urllib.error import URLError
-          from urllib.parse import urljoin
+          from urllib.parse import urljoin, urlsplit
           from urllib.request import Request, urlopen
 
+          def open_public(url, accept=None):
+              headers = {
+                  "Cache-Control": "no-cache",
+                  "User-Agent": "REA-Website-Verifier/1.0 (+https://github.com/morluto/rea)",
+              }
+              if accept:
+                  headers["Accept"] = accept
+              return urlopen(Request(url, headers=headers), timeout=10)
+
           expected = os.environ["GITHUB_SHA"]
           pages_url = os.environ["PAGES_URL"]
           if not pages_url:
@@ -153,11 +178,7 @@ jobs:
               for name, base in list(pending.items()):
                   url = urljoin(base, "deployment.json") + "?commit=" + expected
                   try:
-                      request = Request(url, headers={
-                          "Cache-Control": "no-cache",
-                          "User-Agent": "REA-Website-Verifier/1.0 (+https://github.com/morluto/rea)",
-                      })
-                      with urlopen(request, timeout=10) as response:
+                      with open_public(url) as response:
                           actual = json.load(response).get("commit")
                       if actual == expected:
                           print(f"{name} serves {expected}: {base}")
@@ -172,6 +193,27 @@ jobs:
                   time.sleep(5)
           if pending:
               sys.exit("Could not verify published revision on: " + ", ".join(pending))
+          for name, base in hosts.items():
+              with open_public(urljoin(base, ".well-known/agent-skills/index.json")) as response:
+                  skill, = json.load(response)["skills"]
+              # Fetch from each selected host, including the /rea/ mirror.
+              artifact_url = urljoin(base, urlsplit(skill["url"]).path.lstrip("/"))
+              with open_public(artifact_url) as response:
+                  digest = "sha256:" + hashlib.sha256(response.read()).hexdigest()
+              if digest != skill["digest"]:
+                  sys.exit(name + " serves a skill archive with the wrong digest")
+              for manifest in ("ard.json", "ai-catalog.json"):
+                  with open_public(urljoin(base, ".well-known/" + manifest)) as response:
+                      if not json.load(response).get("entries"):
+                          sys.exit(name + " serves an empty discovery manifest")
+              if name == "Cloudflare":
+                  with open_public(base, accept="text/markdown") as response:
+                      vary = {value.strip().lower() for value in response.headers.get("Vary", "").split(",")}
+                      if response.headers.get_content_type() != "text/markdown" or "accept" not in vary:
+                          sys.exit("Cloudflare did not negotiate Markdown with Vary: Accept")
+                      if 'rel="ard"' not in response.headers.get("Link", ""):
+                          sys.exit("Cloudflare is missing its discovery Link header")
+              print(name + " serves the discovery manifests and verified skill archive")
           with open(os.environ["GITHUB_STEP_SUMMARY"], "a", encoding="utf-8") as summary:
               links = " and ".join(f"[{name}]({url})" for name, url in hosts.items())
               summary.write(f"Published commit `{expected}` to {links}.\n")

.oxlintrc.json

@@ -76,14 +76,12 @@
     {
       "files": ["src/**/*.ts"],
       "rules": {
-        "rea/no-bulk-append-spread": "error",
         "typescript/no-misused-promises": "error"
       }
     },
     {
       "files": ["tests/**/*.ts", "src/**/*.test.ts", "src/**/*.fixture.ts"],
       "rules": {
-        "rea/no-bulk-append-spread": "off",
         "max-lines": "off",
         "max-lines-per-function": [
           "error",

AGENTS.md

@@ -128,8 +128,8 @@ owner to complete. Never attest to human review on their behalf. Draft status
 is optional; preserve the PR's current draft status unless the user asks to change
 it. Later pushes do not invalidate the readiness declaration or require another
 human sign-off. Keep the PR body and affected validation current as work changes.
-The readiness check validates
-the filled declaration; CI and automated review results do not establish that
+The readiness check validates the filled declaration or the author exemption
+listed in CONTRIBUTING.md; CI and automated review results do not establish that
 human review occurred, maintainer scope approval, or merge permission.
 
 <!-- BEGIN:turborepo-agent-rules -->

CONTRIBUTING.md

@@ -93,7 +93,8 @@ Before requesting maintainer review, the human owner must:
   passing. Follow the [verification requirements](#development-feedback-and-pr-verification)
   for required checks and real-provider coverage.
 - Complete the **Author readiness** declaration in the PR template, explicitly
-  stating that the PR is ready for initial review.
+  stating that the PR is ready for initial review, unless the PR author is
+  exempt from the declaration as listed below.
 
 Keep the PR description concise and current. Explain the problem, observable
 outcome, approach, and useful evidence; link long logs or put them in optional
@@ -114,9 +115,12 @@ validation current as work changes. AI agents must leave human declarations to
 the human owner and must not attest to a human review on their behalf.
 
 The `PR readiness` check reports pending for drafts and requires a named human
-owner and all three template declarations for non-drafts. It checks that the
-declaration is filled, not whether the code or evidence is correct. Existing PRs
-can copy the **Author readiness** section from the current template.
+owner and all three template declarations for non-drafts. PRs authored by
+@morluto, @kaoru0822-kitauji, @N0zoM1z0, or @gensokyo-dev are exempt from
+the declaration; their human ownership, review, and evidence responsibilities
+above still apply. The check identifies authors by GitHub account ID. It checks
+that the declaration is filled, not whether the code or evidence is correct.
+Existing PRs can copy the **Author readiness** section from the current template.
 Open PRs must have distinct head commits because GitHub shares commit statuses.
 
 Maintainers enable merge enforcement after the workflow is merged: run

README.md

@@ -20,6 +20,8 @@
 
 **[Website](https://rea.tools/) · [Guides](https://rea.tools/guides/) · [Showcases](https://rea.tools/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Join the REA community on Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Quick start](#quick-start) · [How REA works](#how-rea-works) · [What you can analyze](#what-you-can-analyze) · [Showcases](#showcases) · [Build RL environments](#build-rl-environments) · [FAQ](#faq) · [Documentation](#documentation)
 
 <code>npx rea-agents setup</code>

THIRD_PARTY_NOTICES.md

@@ -60,7 +60,7 @@ pin inspection sources; they are not included in the npm package.
 | 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.                                                                                                                                                                                                                                    |
-| Kuna                           | [Apache-2.0](https://github.com/Noelo-Lab/kuna/blob/v1.988/LICENSE), with its [NOTICE](https://github.com/Noelo-Lab/kuna/blob/v1.988/NOTICE)                                                                                                             | Bring your own `kuna`; REA verified 1.988 and never downloads, installs or bundles it. REA invokes it as a separate process.                                                                                                                                                       |
+| Kuna                           | [Apache-2.0](https://github.com/Noelo-Lab/kuna/blob/v1.990/LICENSE), with its [NOTICE](https://github.com/Noelo-Lab/kuna/blob/v1.990/NOTICE)                                                                                                             | Bring your own `kuna`; REA verified 1.990 and never downloads, installs or bundles it. REA invokes it as a separate process.                                                                                                                                                       |
 | 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.                                                                                                                                                                                                                        |

bridge/native/ReaNativeUI.swift

@@ -65,7 +65,10 @@ func observe(_ request: Request) async throws -> [String: Any] {
   guard SHA256.hash(data: executableBytes).map({ String(format: "%02x", $0) }).joined() == request.sha256 else {
     try fail("target-mismatch", "Running executable bytes differ from the active target digest")
   }
-  guard let windows = CGWindowListCopyWindowInfo([.optionIncludingWindow], request.window_id) as? [[String: Any]],
+  // .optionIncludingWindow omits a selected window while it is off screen, which
+  // is not absence: select from every window, and let capture report what it can.
+  guard let windows = (CGWindowListCopyWindowInfo([.optionAll], kCGNullWindowID) as? [[String: Any]])?
+    .filter({ ($0[kCGWindowNumber as String] as? NSNumber)?.uint32Value == request.window_id }),
     windows.count == 1, let window = windows.first,
     (window[kCGWindowOwnerPID as String] as? NSNumber)?.int32Value == request.pid,
     (window[kCGWindowLayer as String] as? NSNumber)?.intValue == 0,

bridge/process/ProcessRunTokenReader.swift

@@ -15,6 +15,8 @@ struct ProcessIdentityObservation: Encodable {
   let effective_uid: UInt32?
   let lifecycle: String?
   let relation: ProcessRelationObservation?
+  /** Display-only kernel accounting name (p_comm, at most MAXCOMLEN bytes). */
+  let name: String?
   let reason: String?
 }
 
@@ -31,6 +33,7 @@ struct ProcessMetadata {
   let parentPid: Int32
   let processGroupId: Int32
   let pointerSize: Int
+  let name: String
 }
 
 struct ProcessSnapshot {
@@ -92,7 +95,8 @@ func readProcessSnapshot(
             lifecycle: process.kp_proc.p_stat == SZOMB ? "zombie" : "non-zombie",
             parentPid: process.kp_eproc.e_ppid,
             processGroupId: process.kp_eproc.e_pgid,
-            pointerSize: process.kp_proc.p_flag & P_LP64 != 0 ? 8 : 4
+            pointerSize: process.kp_proc.p_flag & P_LP64 != 0 ? 8 : 4,
+            name: processName(process)
           )
         )
       },
@@ -103,6 +107,13 @@ func readProcessSnapshot(
   return ProcessSnapshot(metadata: [:], failure: "process_table_failed_\(ENOMEM)")
 }
 
+/** Decode the NUL-terminated accounting name; invalid UTF-8 is replaced for display. */
+private func processName(_ process: kinfo_proc) -> String {
+  withUnsafeBytes(of: process.kp_proc.p_comm) { bytes in
+    String(decoding: bytes.prefix { $0 != 0 }, as: UTF8.self)
+  }
+}
+
 func readProcessIdentities(pids: [Int32]) -> [ProcessIdentityObservation] {
   readProcessIdentities(pids: pids, snapshot: readProcessSnapshot())
 }
@@ -130,6 +141,7 @@ func readProcessIdentities(
         effective_uid: nil,
         lifecycle: nil,
         relation: nil,
+        name: nil,
         reason: "not-in-snapshot"
       )
     }
@@ -144,6 +156,7 @@ func readProcessIdentities(
         parent_pid: metadata.parentPid,
         process_group_id: metadata.processGroupId
       ),
+      name: metadata.name,
       reason: nil
     )
   }
@@ -157,6 +170,7 @@ private func unavailableIdentity(pid: Int32, reason: String) -> ProcessIdentityO
     effective_uid: nil,
     lifecycle: nil,
     relation: nil,
+    name: nil,
     reason: reason
   )
 }
@@ -214,6 +228,22 @@ func processRunTokenObservation(
   parse: Result<ParsedRunToken, RunTokenReadError>,
   codeSigningFlags: () -> UInt32?
 ) -> RunTokenObservation {
+  // Native inspection restrictions apply even when argv is incomplete or
+  // overwritten, before parsing reaches the environment. Keep this fact at
+  // the native boundary; cleanup still decides whether the process is related.
+  func withheldEnvironment() -> RunTokenObservation? {
+    guard let flags = codeSigningFlags() else { return nil }
+    if (flags & csPlatformBinary) != 0 {
+      return unavailable(pid: pid, reason: "platform_binary_environment_withheld")
+    }
+    if (flags & csRestrict) != 0 {
+      return unavailable(pid: pid, reason: "restricted_process_environment_withheld")
+    }
+    return nil
+  }
+  func unavailableEnvironment(_ reason: String) -> RunTokenObservation {
+    withheldEnvironment() ?? unavailable(pid: pid, reason: reason)
+  }
   switch parse {
   case .success(let parsed):
     if let runId = parsed.runId {
@@ -224,40 +254,25 @@ func processRunTokenObservation(
     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")
-    }
+    if let withheld = withheldEnvironment() { return 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")
+      return unavailableEnvironment("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")
+      return unavailableEnvironment("ambiguous_environment_boundary")
+    // These failures already establish that token bytes were available.
     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")
+      return unavailableEnvironment("apple_vector_unavailable")
     case .invalidBuffer:
-      return unavailable(pid: pid, reason: "malformed_procargs")
+      return unavailableEnvironment("malformed_procargs")
     case .invalidBufferAt(let stage):
-      return unavailable(pid: pid, reason: "malformed_procargs_\(stage.rawValue)")
+      return unavailableEnvironment("malformed_procargs_\(stage.rawValue)")
     }
   }
 }

bridge/pwntools/layout_impl.py

@@ -16,10 +16,10 @@ PROFILE = "[email protected];[email protected];[email protected]"
 PACKAGES = {"pwntools": "4.15.0", "pyelftools": "0.33", "unicorn": "2.1.2"}
 OUTPUT_BYTES = 64 * 1024 * 1024
 # (e_machine, ELF class bits, byte order) -> architecture pwntools must also report.
-PROFILES = {(62, 64, "little"): "amd64", (3, 32, "little"): "i386", (40, 32, "little"): "arm",
-            (8, 32, "little"): "mips", (8, 32, "big"): "mips"}
-PROFILE_SCOPE = ("Layout profile requires ELF64 x86-64, ELF32 i386, ELF32 ARM little-endian "
-                 "or ELF32 MIPS little- or big-endian objects.")
+PROFILES = {(62, 64, "little"): "amd64", (183, 64, "little"): "aarch64", (3, 32, "little"): "i386",
+            (40, 32, "little"): "arm", (8, 32, "little"): "mips", (8, 32, "big"): "mips"}
+PROFILE_SCOPE = ("Layout profile requires ELF64 x86-64, ELF64 AArch64 little-endian, ELF32 i386, "
+                 "ELF32 ARM little-endian or ELF32 MIPS little- or big-endian objects.")
 
 
 class LayoutFailure(Exception):

docs/adb-device-analysis.md

@@ -257,12 +257,12 @@ beyond already-connected devices (`adb connect`) are not exposed yet.
 
 ## Real verification
 
-`npm run verify:adb` exercises every operation against a live device. It
+`npm run verify:adb` exercises observation and transfer operations against a live device. It
 requires `adb` (or `REA_ADB_PATH`) and any attached emulator or phone;
 `--serial` selects one when several are attached, `--pull PACKAGE` pulls a
-real package and verifies the returned digests against the files on disk,
-and `--install-apk PATH` drives a full install → resolve → uninstall
-lifecycle with a lane-owned probe.
+real package and verifies the returned digests against the files on disk.
+The optional `--install-apk` lifecycle is unavailable until the selected APK
+identity and device cleanup ownership can be established safely.
 
 Record: verified against adb 34.0.5 on Linux against an Android 14
 (API 34) x86_64 emulator — all operations above, a 224-package inventory,

docs/agent-environments.md

@@ -1149,7 +1149,12 @@ global conflict involving an equal pair is not used to normalize either pair; th
 
 A step agrees when the statuses and the normalised bodies are equal. The report gives agreement overall, per probe and
 per tool, `agreement_ignoring_order` for steps that differ only in item order, and each step's differences, at most 20,
-with a JSON pointer and both values cut short:
+with a JSON pointer and both values cut short. Each step reports `differences_total` before that cap,
+`differences_omitted`, and `differences_truncated` (true when entries were omitted). Agreement and categories use the
+complete comparison. Omitted entries are not retained, so `--full-output` cannot recover them; inspect the original
+capture and probe Evidence to investigate beyond the returned pointers.
+
+The difference categories are:
 
 | Category        | Meaning                                                                                                      |
 | --------------- | ------------------------------------------------------------------------------------------------------------ |

docs/binary-diagnostics.md

@@ -9,9 +9,11 @@ ordinary binary inspection use the same modular capability.
 ## Bring your own engine
 
 Real verification covers **Linux x64** hosts and **EXEC/DYN/REL** objects in
-these profiles: ELF64 x86-64, ELF32 i386, ELF32 little-endian ARM and ELF32
-little- or big-endian MIPS. Other machines, classes and byte orders, including
-AArch64 and big-endian ARM, return `unsupported_target`. Provide an absolute
+these profiles: ELF64 x86-64, ELF64 little-endian AArch64, ELF32 i386, ELF32
+little-endian ARM and ELF32 little- or big-endian MIPS. Other machines, classes
+and byte orders, including big-endian ARM, return `unsupported_target`. Other
+hosts, including Linux arm64, return `capability_unavailable` naming the
+detected host before any target is read. Provide an absolute
 Python executable whose environment already contains unchanged pwntools 4.15.0,
 pyelftools 0.33 and Unicorn 2.1.2:
 
@@ -424,8 +426,10 @@ rea inspect-native-exports ./libexample.so --detail summary --evidence_output ./
   without an export directory is a complete empty inventory.
 - ELF rows list defined `.dynsym` symbols with global, weak or GNU-unique
   binding and default or protected visibility. Each row keeps `value`
-  (`st_value` as stored; bit 0 marks Thumb code on 32-bit ARM), `size`, `type`
-  and `binding`. ELF string tables do not declare a text encoding. Names that
+  (`st_value` as stored; bit 0 marks Thumb code on 32-bit ARM), `size`, `type`,
+  `binding` and `visibility` (`"default"` or `"protected"`; `"unknown"` in older
+  historical Evidence where visibility was not recorded). ELF string tables do
+  not declare a text encoding. Names that
   contain non-ASCII bytes are decoded as UTF-8 when those bytes form valid
   UTF-8; `name_utf8_decodable` records whether the non-ASCII name bytes are
   valid UTF-8 and `name_bytes_base64` retains those exact bytes (`null` for

docs/browser-observation.md

@@ -500,7 +500,7 @@ REA retains local observation data and removes only URL userinfo credentials, ex
   connection/frame inventories; matching direction/opcode sequences alone do
   not prove alignment after drops. Redirect resource types use only shared known metadata:
   an omitted type leaves equality unknown rather than proving a modification.
-- Console observations with a stack source retain call type, argument types, the CDP `Runtime.Timestamp` in milliseconds since Unix epoch, and source location. An absent or invalid timestamp is `null` and makes `console_events` incomplete; an observed zero remains zero. Argument types are listed by CDP argument position. An absent or non-array argument inventory is `null`, and a malformed argument keeps its position as `unknown`; either makes `console_events` incomplete, while an observed empty inventory remains `[]`. A malformed stack or top stack frame is not replaced by its caller: the console event is excluded, and a request initiator location or WebMCP `registration_source` is `null`; each is reported as `invalid_protocol_value`. This epoch-millisecond clock is separate from the monotonic-second timestamps in `observe-web-session`. When console text capture is selected, already-delivered primitive values are retained verbatim with their original `argument_index`; objects, getters, and remote properties are never expanded. Text capture is `unavailable` when the argument inventory is unknown.
+- Console observations with a stack source retain call type, argument types, the CDP `Runtime.Timestamp` in milliseconds since Unix epoch, and source location. An absent or invalid timestamp is `null` and makes `console_events` incomplete; an observed zero remains zero. Argument types are listed by CDP argument position. An absent or non-array argument inventory is `null`, and a malformed argument keeps its position as `unknown`; either makes `console_events` incomplete, while an observed empty inventory remains `[]`. A malformed stack or top stack frame is not replaced by its caller: the console event is excluded, and a request initiator location or WebMCP `registration_source` is `null`; each is reported as `invalid_protocol_value`. Source locations are CDP's zero-based lines and UTF-16 columns. A negative column, which Chrome reports on the first line of some inline scripts, is kept as reported and does not locate a column in the source URL; an invalid line or a non-integer column is `null` (WebMCP `registration_source` is `null`) and is reported as `invalid_protocol_value`. This epoch-millisecond clock is separate from the monotonic-second timestamps in `observe-web-session`. When console text capture is selected, already-delivered primitive values are retained verbatim with their original `argument_index`; objects, getters, and remote properties are never expanded. Text capture is `unavailable` when the argument inventory is unknown.
 - WebSocket observations retain direction, opcode, and payload byte length. A missing or malformed CDP opcode is returned as `null`, with opcode-dependent payload byte length and shape also `null`, and marks `websocket_frames` incomplete. Missing or wrong-typed payload data is not treated as empty; its byte length is `null`, and a text payload shape is unknown. Numeric opcode `0` and an observed empty payload remain distinct from missing data. Opcode is nullable in the public inspection schema, so clients must handle unknown protocol values. When shape capture is selected, text messages are classified as text or complete value-free JSON shape; binary bytes, hashes, prefixes, and raw payloads 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 fingerprints hash all decoded response bytes, including bodies above 64 KiB, subject to the [64 MiB per-message CDP limit](#request-boundary). Missing or undecodable cache bodies 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.

docs/cli.md

@@ -569,6 +569,17 @@ prints a failure, so it omits the same-session reuse and
 `parameter`) for a new invocation and rerun. For example, a Ghidra decompile
 deadline names `REA_GHIDRA_DECOMPILE_TIMEOUT_SECONDS`.
 
+`rea debug-with-gdb` and `rea debug-with-rizin` hold their debugger only for the
+command's own process. If cleanup cannot confirm the debugger stopped, they print
+`cleanup_incomplete` and exit nonzero without stopping it. `remediation.action`
+names the process, as does `retained_debugger_pids` in `details.diagnostics`, or
+in `details.diagnostics.cleanup_error.details.diagnostics` when the command also
+failed or was interrupted. Exiting closes the debugger's command and output pipes.
+REA does not establish what the debugger or its target does next, so check that
+process. An MCP server that shuts down likewise exits with status 1 instead of
+waiting on a debugger it could not stop. It names the process on stderr and as
+`retained_debugger_pids` in its error log.
+
 If one-shot analysis fails and closing its session also fails, the original
 error details remain inline and `details.cleanup_error` describes the later
 close failure. Each diagnostic retains its own reported resources; the later
@@ -601,6 +612,9 @@ Saved snapshots include only cache entries whose subject identifies the selected
 target. Unbound observations and observations of child artifacts remain available
 in the open session and its Evidence bundle; they are excluded from the target's
 saved replay entries without changing their subjects.
+The snapshot also retains available Evidence linked by selected records, including
+transitive sources with an unbound or different subject. Those sources are not
+replay entries for the selected target. Unrelated records remain excluded.
 
 Snapshots retain eligible target-scoped question histories in full, including
 mutation Evidence recorded while another target was active. Observations and
@@ -1294,6 +1308,18 @@ new file. A publication that has committed can win a later cancellation; the
 after commit, the `cleanup_incomplete` error reports the same file as
 `details.published_path` and `details.bytes`.
 
+When a failure retains a completed observation and no bundle committed,
+`details.source_closure` states what that observation cites. `publication` is
+`not_published`. `records` holds authentic source Evidence this invocation
+captured, within 64 KiB. `omitted` lists up to 64 other identities and
+`omitted_count` the total: `not_captured` means the invocation holds no such
+record, so nothing is shown for it or its ancestors; `not_inlined` means it was
+captured but is not in this response. `complete` is true only when every cited
+record is in `records` or embedded in the observation. If examining the closure
+itself fails, `collection_failure` names the error type and the closure state is
+unknown; the original failure is still returned. Nothing is retained after exit,
+and running the command again is a new observation.
+
 When the reader of piped output closes early, as `rea ... | head` does, the
 CLI stops writing, prints no error and exits `141` (`128 + 13`, the status of a
 SIGPIPE-terminated writer). The operation itself may have completed; only its

docs/client-configuration.md

@@ -45,6 +45,10 @@ the client's configured command or reproduce the client's environment. If doctor
 passes but the client disconnects, inspect the client's startup error and registered
 launcher; see [verification scope](client-verification.md#launch-environment-verification).
 
+`rea doctor` has the same limit for provider checks, and reports a provider setting
+the registration's file does not carry as a
+[non-blocking unknown](installation.md#check-readiness-for-your-task).
+
 ## Runtime settings and client deadlines
 
 `update_analysis_settings` changes curated defaults inside the running REA

docs/electron-observation.md

@@ -201,6 +201,9 @@ neither result means the requested host operation occurred. The existing
 `openExternal`, `openPath`, and `trashItem` blocks reject their promises.
 The `trashItem` block prevents that call from reaching the OS trash operation;
 it does not intercept other ways the application can remove files.
+Process rejection and uncaught-exception events are observed at dispatch without
+adding exception handlers. The runtime and application retain their warning,
+handler, and fatal-exit policies, including for rejected shell-block promises.
 CLI SIGINT/SIGTERM requests cancellation and awaits owned-process cleanup.
 Window IDs, URLs, and cached titles come from one main-process Electron snapshot,
 so a crashed renderer does not require a successful renderer-side title query.

docs/elf-notes.md

@@ -38,6 +38,12 @@ were examined; an unavailable section name does not erase those note facts.
 For GNU notes, `gnu.build_id_hex`, `gnu.abi_tag`, or `gnu.gold_version` are
 filled when the descriptor matches that type. Other owners keep `gnu` null.
 
+For Linux `CORE` notes, `core.prpsinfo`, `core.file`, `core.siginfo`, or
+`core.auxv` are decoded for process arguments and status (`NT_PRPSINFO`), mapped
+file memory regions and paths (`NT_FILE`), fault signal and code (`NT_SIGINFO`), and
+auxiliary vectors (`NT_AUXV`) across 32-bit and 64-bit images. Other owners or unparsed
+notes keep `core` null.
+
 `PT_NOTE` and `SHT_NOTE` are examined independently. The same file range may
 appear twice when both tables declare it.
 
@@ -48,7 +54,7 @@ the image never carried notes in another container or overlay.
 
 The selected file is limited to 64 MiB by default (512 MiB maximum). At most
 32 note entries are examined by default across every table. Retained note
-names, section names, and GNU descriptor text share a scan budget based on
+names, section names, GNU descriptor text, and core dump string representations share a scan budget based on
 the source bytes plus any decoded section-name table, capped at
 `max_file_bytes`. The raw source and decoded table must also fit together
 within that ceiling. If admission fails, the error reports the required

docs/evm-bytecode.md

@@ -28,6 +28,15 @@ The host must already provide util-linux `prlimit`, normally `/usr/bin/prlimit`;
 no system tools and changes no user configuration. [Upstream provenance and
 license](https://github.com/morluto/rea/blob/main/third_party/evmole/README.md) are packaged.
 
+Limiter discovery validates the selected path without launching it. A relative
+path or directory is reported as `not_configured`, a missing file as
+`executable_missing`, and denied access or another filesystem-probe failure as
+`runtime_unavailable`. Recovery names the selected path and
+`REA_EVM_PRLIMIT_COMMAND`; denied access asks for execute permission and
+filesystem access. The original admission or system code stays in diagnostics.
+An available filesystem probe does not verify the selected tool's version or
+util-linux compatibility, and REA does not substitute another command.
+
 ## Evidence semantics
 
 - `encoding` is required: raw bytes or UTF-8 hex, with optional `0x` prefix and

docs/firmware-analysis.md

@@ -26,6 +26,9 @@ REA does not install these tools, their dependencies or system packages. It
 checks the tool's `--version` banner when executing a request. Binwalk must
 print `binwalk <version>` and Unblob must print the version alone. Inspection and
 extraction have independent prerequisites; one does not require the other.
+Capability discovery recommends `REA_BINWALK_COMMAND` for missing inspection
+and `REA_UNBLOB_COMMAND` for missing extraction. Configuring the other engine
+does not enable the unavailable operation.
 Both require util-linux `prlimit`, normally `/usr/bin/prlimit`; override its
 absolute path with `REA_FIRMWARE_PRLIMIT_COMMAND` if needed. `rea doctor` does
 not currently inspect firmware readiness. Engine selection/version checks run

docs/flutter-analysis.md

@@ -109,9 +109,13 @@ server; older servers keep the complete request after schema discovery.
   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
+  and Dart/toolchain strings the build carries. `dart_version` accepts the
+  labeled `Dart VM version:` form and the unlabeled Dart
+  `Version::String()` form (`<version> (<channel>) (<commit time>)` with an
+  optional ` on "<os>_<arch>"` suffix) that release engines embed. An
+  unlabeled match is a candidate based on its shape; it is not bound to a
+  Dart symbol and another string can match. `null` means no supported pattern
+  was found, not proof that no version exists. 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;
@@ -122,10 +126,13 @@ server; older servers keep the complete request after schema discovery.
   Compressed `SHT_NOTE` sections are skipped with an unsupported-compression
   limitation and partial coverage; IDs from other note ranges remain candidates.
   Encoded note storage is never interpreted as note headers or inflated.
-- 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.
+- Version and toolchain 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. Unlabeled runs that miss the
+  `Version::String()` grammar are ignored and do not set
+  `dart_version_capped`.
 - **`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

docs/index.md

@@ -51,7 +51,6 @@ For runtime instrumentation, see the [Frida guide](./frida-instrumentation.md).
 | --------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
 | Identify a desktop installation                     | [Runtime candidates, evidence and follow-up files](desktop-applications.md)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
 | Inspect native functions, references, or types      | [Native binaries](native-investigation.md), [Go metadata](go-binaries.md), [Rust dependencies](rust-binaries.md), [CPython pyc files](python-pyc.md), [PE certificates](pe-certificates.md), [PE imports](pe-imports.md), [PE TLS/load config](pe-load-config.md), [ELF notes](elf-notes.md), [Linux kernel modules](kernel-modules.md), [Apple IPS](apple-ips.md), [Erlang BEAM](erlang-beam.md), [offline ELF layout](binary-diagnostics.md), [recorded crashes](recorded-crashes.md)                                                                               |
-| Inspect native functions, references, or types      | [Native binaries](native-investigation.md), [Go metadata](go-binaries.md), [Rust dependencies](rust-binaries.md), [CPython pyc files](python-pyc.md), [PE certificates](pe-certificates.md), [PE imports](pe-imports.md), [PE TLS/load config](pe-load-config.md), [ELF notes](elf-notes.md), [Apple IPS](apple-ips.md), [Erlang BEAM](erlang-beam.md), [offline ELF layout](binary-diagnostics.md), [recorded crashes](recorded-crashes.md)                                                                                                                          |
 | Select a native provider or import format           | [Ghidra semantics](ghidra-provider.md), [Windows](windows-ghidra-p0.md), [DOS](ghidra-dos.md), [raw bytes](raw-binary.md), [GameCube DOL](ghidra-dol.md), [MIPS](ghidra-mips.md), [PSP](ghidra-psp.md), [NativeAOT](ghidra-nativeaot.md), [IDA](ida-provider.md), [Kuna](kuna-provider.md), [JEB](jeb-analysis.md)                                                                                                                                                                                                                                                    |
 | Analyze JavaScript or Electron artifacts            | [Static analysis](javascript-artifact-reconstruction.md), [tracing and comparisons](javascript-application-workflows.md), [application graph](javascript-application-graph.md), [source recovery](javascript-recovery.md)                                                                                                                                                                                                                                                                                                                                             |
 | Compare static JavaScript with runtime observations | [Runtime attribution](javascript-runtime-observation.md), [reconciliation](javascript-runtime-reconciliation.md), [Electron observation](electron-observation.md), [Electron fuses](electron-fuses.md)                                                                                                                                                                                                                                                                                                                                                                |

docs/installation.md

@@ -390,6 +390,20 @@ A scoped report has `scope.mode: "explicit"`. Its `scope_checks` determine
 `environment_healthy` summarizes the full audit. `--target` adds a target check,
 but the report remains audit-wide unless a scope option is also supplied.
 
+Provider checks read the terminal running doctor. A client-started server may
+receive a different environment, and doctor cannot observe it. When an available
+provider relies on a setting that a registration's file does not record, or
+records with other text, doctor adds a non-blocking
+`registration_environment:<client>` check with `classification: "unknown"`. The
+setting names appear in that registration's `provider_settings`. Selecting the
+client or an affected provider places the check in `scope_checks`; it never
+changes `healthy`, because the client may still pass the setting on. Preview
+`rea setup --client <client-id> --skill=false --dry-run` from that terminal and
+approve it to record the settings, then restart the client. To see what the
+running server can use, call `binary_session` in the client and read
+`result.tool_availability`. For Hopper, only an explicit `HOPPER_LAUNCHER_PATH`
+is compared.
+
 An invalid `REA_LOG_LEVEL` fails every scope, including one client or the
 installed skill, because the executable rejects it before any command. A
 selected client also fails for every other setting that MCP startup rejects,
@@ -666,7 +680,8 @@ Hopper is separate commercial software with its own license. Its free demo has
 vendor-defined limits, and a paid license is optional. REA reuses any detected
 installation and preserves Hopper during uninstall.
 
-The automated Hopper installer supports macOS 12+ and x86_64 hosts running
+The automated Hopper installer supports macOS 12+, subject to the minimum
+macOS version of the current Hopper release, and x86_64 hosts running
 Ubuntu 24.04+, Fedora 41+, Nobara 44+, Arch Linux, CachyOS, or Omarchy. The
 official Linux packages are x86_64-only, so setup does not offer Hopper on other
 Linux architectures; configure [Ghidra](#ghidra) there instead. REA itself runs on any
@@ -709,6 +724,15 @@ and digest, validates the application bundle, and atomically installs it to
 operator can choose its demo mode or activate an existing license. Homebrew and
 administrator access are not used.
 
+Each Hopper release declares its own minimum macOS version, so the setup plan
+cannot confirm support before approval. Setup compares this Mac's version with
+the release metadata's `deployment_target` before downloading the DMG, and with
+the signed application's `LSMinimumSystemVersion` before copying it. If this Mac
+is older, setup fails with `incompatible_host_version`, names both versions, and
+installs nothing. Update macOS, or install a Hopper release that supports this
+macOS version yourself and set `HOPPER_LAUNCHER_PATH`. An existing Hopper
+installation is not checked against the current release's minimum.
+
 The Hopper launcher action used by REA creates a document from an executable;
 its supported command-line interface does not attach to an already-open
 document. While another live REA session owns the same target and loader

docs/javascript-artifact-reconstruction.md

@@ -876,8 +876,22 @@ and Evidence. If that expansion does not fit, the file node retains
 `static_analysis` and any admitted `semantic_module_analysis` observations,
 including their exact values and source locations. Use the public modules item
 view to read them. The graph marks those unexpanded relationships as unknown.
-If cumulative result capacity is exhausted, analysis stops; inventoried but
-unexamined files remain present with explicit unknown scopes.
+
+The parent charges everything it retains against the heap that was free when
+analysis started, less a transfer reserve. Semantic facts and expanded
+relationships stop at half of that capacity, and a source refused there keeps
+its static facts with the refusal reported. Static facts can continue to three
+quarters of it while the live heap also has room for another copy of what it
+holds, so results charged for earlier sources do not cost later sources their
+static facts when the heap can still hold them. When that share runs short,
+relationships that were reserved but not yet expanded are given up before any
+static facts, and each affected file reports the omission.
+
+If static facts themselves exhaust that share, or the parent's live heap cannot
+hold the next checkpoint, analysis stops; inventoried but unexamined files remain
+present with explicit unknown scopes. The failure reports the limit that applied
+in bytes, what was still available, the live heap in use and what the checkpoint
+needed.
 
 ### Large local JSON inputs
 
@@ -1002,7 +1016,12 @@ reason in the renderer observations.
 HTML script source ranges follow the HTML parser across LF, CRLF, and bare CR
 line endings, preserving UTF-16 columns.
 Inline JavaScript in local HTML, meaning classic or module `<script>` bodies
-without a `src` attribute, is not analyzed. Each HTML file that has one gets an
+without a `src` attribute, is not analyzed. Embedded SVG scripts use `href`
+(or `xlink:href` when `href` is absent) instead: an existing reference, even
+an empty one, suppresses the inline body. SVG's `src` and HTML-only `language`
+attributes do not suppress its JavaScript body. SVG references and inline
+bodies share the same inert parsing and source coordinates as HTML scripts;
+template content stays inert. Each HTML file that has an inline body gets an
 unknown `parse-html-inline-scripts` scope with the exact body count and the
 script type and source range of the first 100 bodies; beyond that, coverage is
 truncated under the `html_inline_script_locations_per_file` limit, and the

docs/javascript-runtime-reconciliation.md

@@ -93,6 +93,10 @@ coverage. It does not print captured source text.
 Runtime targets, frames, scripts, and workers become capture-scoped graph
 nodes. REA evaluates compatible static candidates conservatively:
 
+V8 Inspector targets whose protocol type is `node` use JavaScript asset/module
+candidates for their verified entry location. A target-only correspondence does
+not mark that script loaded; load-state evidence requires a script observation.
+
 1. A captured-source SHA-256 plus one matching mapped location yields
    a content-and-location match.
 2. When no static candidate exists at a mapped location, one unique exact

docs/kuna-provider.md

@@ -1,9 +1,27 @@
 # Kuna provider
 
-[Kuna](https://github.com/Noelo-Lab/kuna) is an Apache-2.0 decompiler distributed as one `kuna` executable. For a new installation, use Kuna 1.988, REA's recommended verified release. REA uses a `kuna` you supply; it does not download, install, update or bundle Kuna. REA accepts other releases, which it reports as unverified in `rea doctor` and in each analysis profile. The verified-release flag compares the reported version; it does not authenticate the executable. The 1.988 real-provider checks ran on macOS arm64; Linux and other hosts have not been verified with that release.
+[Kuna](https://github.com/Noelo-Lab/kuna) is an Apache-2.0 decompiler distributed as one `kuna` executable. For a new installation, use Kuna 1.990, REA's recommended verified release. REA uses a `kuna` you supply; it does not download, install, update or bundle Kuna. REA accepts other releases; `rea doctor` and analysis Evidence report them as unverified. Verification compares the reported version; it does not authenticate the executable. Analysis profiles retain the observed version and executable digest, independently of REA's current verified-release policy. The 1.990 real-provider checks ran on macOS arm64; Linux and other hosts have not been verified with that release.
 
 Kuna is static: it reads the selected image and never executes it.
 
+This profile change makes earlier Kuna analysis snapshots and resumable directory
+exports incompatible, even with the same Kuna executable. Keep the old files:
+their recorded Evidence remains historical. To continue analysis, open the
+binary without the old snapshot and save a new one, or choose a new export
+directory. Existing snapshot Evidence can be extracted from its
+`evidence_bundle` JSON member and imported as a separate Evidence bundle; that
+does not restore cached workflow entries or resume the old directory export.
+For example, using `jq` and a new output path:
+
+```bash
+jq -e '.evidence_bundle | select(type == "object")' old-snapshot.json > history.json
+rea evidence-import history.json --json
+```
+
+The CLI command validates the bundle in a one-shot session. To use the records
+in an MCP session, call `import_evidence_bundle` with the extracted file path.
+This validates the embedded bundle, not the rest of the snapshot container.
+
 ## Configure
 
 ```bash
@@ -30,6 +48,10 @@ completes. Cleanup retries keep both the observations and the exact file owner.
 | `REA_KUNA_TIMEOUT_MS`         | 120 s                                                                                                                                   | Deadline for one single-function decompile, and for the inventory run REA makes when the export ends without one.                                                                        |
 | `REA_KUNA_JOBS`               | 1                                                                                                                                       | `decompile-project --jobs`. 1 is Kuna's reference answer; more jobs finish sooner but lose callee-first prototypes, and change about 40–60% of function bodies on the measured binaries. |
 
+Pseudocode read from a parallel project export includes a prototype-context
+limitation in its Evidence, including in function dossiers and composed
+workflows. Early single-function results have their own source limitation.
+
 A single-function decompile, inventory, string, graph or xref run that exceeds its deadline fails as `provider_timeout`, naming the setting that owns the deadline and that raising it takes a restart. A project export that exceeds `REA_KUNA_PROJECT_TIMEOUT_MS` is an export failure instead: requests that need the export report `capability_unavailable` with the export's own retry guidance.
 
 Windows is not supported yet. Supported targets are ELF and thin Mach-O executables for x86-64, AArch64 and ARM; PE and universal Mach-O are rejected with a typed reason.

docs/mcp-contracts.md

@@ -218,6 +218,12 @@ Malformed provider replies return `unreadable_output` for
 `resolve_containing_procedure` before successful-result projection;
 cancellation and incomplete cleanup remain errors.
 
+For a procedure match, `procedure_references` with `direction: "incoming"` can
+provide the selected provider's reference kinds when supported. Its scope covers
+the procedure body, so compare target addresses when selecting references to the
+exact entry. It is a separate observation and does not establish runtime use or
+complete reference discovery.
+
 REA prepares complete MCP results within the pinned stdio client's 10 MiB
 receive-buffer budget, including both text and structured representations and
 room for the JSON-RPC envelope. Tool errors omit `structuredContent` from the
@@ -292,6 +298,12 @@ limitation and reason IDs index that page's `items.messages` dictionary. Expand
 those references to recover exact text, repeated messages, and absent versus
 empty optional fields. Failed nodes retain their original error.
 
+Retained `xrefs` Evidence also supports `native-observation` address pages and
+`native-context` source caveats. `xrefs` still returns complete Evidence by
+default and does not accept `detail`; when its MCP response exceeds the budget,
+follow the returned continuation to page the retained result. These addresses
+remain untyped; paging does not add reference kinds or establish completeness.
+
 These views reference complete source context through `source_context.digest`
 and caveat counts. Read it using the same parent with
 `view: {"kind":"native-context","facet":"overview"}`. The other facets are
@@ -320,6 +332,8 @@ Oversized native inventories likewise include a ready-to-call first
 `native-inventory` page. Reduce its limit or select fields when a row is large.
 An oversized `trace_application_feature` result includes a first `feature-trace`
 page of its retained terminal `paths`; page `nodes` or `edges` from there.
+An oversized `batch_decompile` result includes a first body-free
+`batch-decompile` page of its retained `items`; read bodies from there.
 
 A broad follow-up or `get_evidence_bundle` can also exceed the response budget; exporting
 preserves the complete session without sending it through a single MCP frame.
@@ -369,6 +383,12 @@ is null. `total` counts all matching heads; each page preserves complete records
 in stable unknown ID order. Filters apply before paging. Pages reflect current
 session state, so keep the unknown ledger unchanged while traversing it.
 
+`list_unknowns` only lists registered heads. Import restores the Unknown records
+present in a bundle; it does not turn residual question text embedded in Evidence
+into registry entries. An empty registry therefore does not establish that
+retained Evidence has no open questions. Inspect the relevant Evidence or its
+analysis view for its residual questions.
+
 Automatically retained workflow questions include their immutable result
 Evidence ID. Distinct results retain independent questions and source links;
 repeating the same result does not create another unknown. Workflow Evidence
@@ -388,9 +408,26 @@ that revision. If a list response overflows, retry on the same connection with
 the same filters and offset and a smaller limit; export if one head still cannot
 fit.
 
+To find missing sources before a downstream analysis refuses them, read
+`missing_direct_links` from the summary, or call `get_evidence_bundle` with
+`detail: "links"`. The result (`kind: "evidence-link-inspection"`) transfers no
+record payload. `filters` select the starting records, while membership is
+checked against every record in the session or file, so a filtered-out source is
+`retained`, not missing. Rows (`evidence_id`, `linked_evidence_id`, `status`) are
+in canonical order, default to `link_status: "missing"` and page with `offset` and
+`limit` (default 64); `links` counts every inspected link regardless of that
+selection. `traversal: "direct"` (the default) reads the selected records' own
+links and sets `retained_closure` to null, because it makes no closure claim.
+`traversal: "retained-closure"` also follows retained links, as composed workflows
+such as `build_call_path` do, and reports `complete` or `incomplete`; anything
+beyond a missing link remains unknown. Missing means absent from this session or
+file, not nonexistent, and neither status describes analysis coverage. Importing
+a missing source changes this membership without changing any Evidence ID.
+
 For a portable exported bundle, the CLI shares this projection:
 `rea inspect-evidence-bundle '{"path":"/tmp/evidence.json","detail":"summary","filters":{"operation":"analyze_function"}}' --format json`.
-`"detail":"records"` with an `evidence_id` filter returns that complete record.
+`"detail":"records"` with an `evidence_id` filter returns that complete record,
+and `"detail":"links"` reads link membership.
 The CLI validates and materializes that local bundle under its existing memory
 budget; summary delivery does not make arbitrary large files cheap to import.
 
@@ -620,6 +657,43 @@ never upgrades `trace_coverage`, and graph connectivity does not prove runtime
 reachability. The CLI keeps complete trace JSON; save it and use
 `inspect-analysis-view` with an `evidence-file` source.
 
+### Selected batch decompile results
+
+`batch_decompile` returns every pseudocode body by default. `detail: "summary"`
+retains the complete batch in the session first and returns its producer counts,
+`source_provider` (the composing workflow), `source_analysis_profile_digest`,
+`decompiling_provider`, procedure identity counts, total and largest
+`pseudocode` size, and `limitation_count`. A server that retains no Evidence
+refuses the summary before analysis. Further views use the returned
+`parent_evidence_id` through `inspect_analysis_view`, or a saved Evidence file or
+portable bundle, and never repeat decompilation.
+
+A `batch-decompile` view selects one `collection` and repeats the batch counts
+and provider identities:
+
+- `items`: rows in retained order with their original `index`, so duplicate
+  selectors stay distinct. A row keeps the selector in `address`, its `procedure`
+  identity and `status`, plus `pseudocode_code_units` or its exact `error` and
+  `limitations`. `where.indices`, `where.addresses`, `where.procedure_addresses`
+  and `where.statuses` filter exactly and combine; an item with unknown identity
+  never matches `procedure_addresses`. Omitted filters select all, and empty
+  arrays select none. `offset` and `limit` default to 0 and 64.
+  `include_pseudocode: true` adds each selected body.
+- `pseudocode`: one `index`'s text in UTF-16 code-unit pages. `offset` and `limit`
+  default to 0 and 16,384, and `coverage` measures that body. An offset inside a
+  surrogate pair is refused, and a limit ending inside one stops before it. An
+  error item has no text and is refused with a pointer to `items`.
+
+Page `coverage` counts the selection only: exhausting it establishes nothing
+about functions outside the batch. Evidence `limitations` aggregate every item's
+provider observations and are not attributed to one successful item; only error
+items keep their own. A page over the response budget offers a smaller selection
+from the same `source`: fewer rows, or one body's `pseudocode` pages when a single
+row does not fit. An inline source offers none, because the record cannot be
+repeated in that reply; resend it with a smaller selection.
+`export_evidence_bundle` remains the complete alternative. The CLI has no batch command; save the Evidence and use
+`inspect-analysis-view` with an `evidence-file` source or `--evidence-bundle`.
+
 Completed `inspect_artifact` and atomic `inventory_artifact` Evidence also
 support summary and inventory views. A summary reports the manifest's expected
 counts separately from available record counts, observed entry kinds, inventory
@@ -940,6 +1014,14 @@ is `unavailable` with a reason when ownership could not be revalidated, or
 at that bounded check. These are historical snapshots, not live process
 inventories, and do not claim that no short-lived descendant existed.
 
+After a failed close retires the target, `binary_session.cleanup_retry` names
+`close_binary` only while that session still retains the cleanup bound to the
+close error. Target-dependent availability then gives the same conditional
+cleanup guidance before suggesting `open_binary`. The hint disappears after
+cleanup succeeds. A pending flag alone does not establish this retry: active
+targets, an in-flight initial close, and external-mutation blockers keep their
+own recovery and do not receive this hint.
+
 `binary_session.cleanup_pending` reports whether the session retains a provider
 client, session file resource, or provider-candidate resource whose cleanup
 has not succeeded. It can be true with

docs/native-investigation.md

@@ -201,6 +201,11 @@ remains valid. This tool
 has no dedicated CLI command; the existing `decompile` command remains the
 single-procedure CLI operation.
 
+`detail: "summary"` retains the complete batch and returns only its counts,
+identities and sizes. `inspect_analysis_view` then reads rows or one body from
+the retained record, an Evidence file or a portable bundle without repeating
+decompilation: see [selected batch results](mcp-contracts.md#selected-batch-decompile-results).
+
 Each batch schedules one provider read at a time, but has no shared input-count
 or aggregate retained-output allowance. It retains every requested row and full
 pseudocode body through Evidence construction. The MCP response budget checks
@@ -295,6 +300,20 @@ relationships but reports their completeness as unknown. In particular, a resolv
 tail call observed in one direction does not prove that the reverse inventory is
 complete. An empty Hopper inventory cannot establish that no such call exists.
 No whole-binary reverse scan is performed.
+
+For large native traces, `trace_feature` accepts `detail: "summary"`. It retains
+the complete trace and returns counts and its parent Evidence ID. Use
+`inspect_analysis_view` with that ID and `view: {kind: "native-feature-trace",
+collection: "references", offset: 0, limit: 64}` to page the original rows;
+`matches` and `unknowns` are separate collections. Matches accept an exact
+`address`; references accept exact `source_address` and `target_address`
+filters. Row indices refer to the retained collection, and page coverage does not establish provider discovery
+completeness. `list_unknowns` pages the independently managed question heads.
+In the one-shot CLI, `rea trace ... --detail summary --evidence_output PATH`
+stores the trace and its linked source Evidence for later
+`inspect-analysis-view` calls. This bundle does not include the CLI session's
+Unknown registry; its original question texts remain available in the trace's
+`unknowns` pages.
 Regex searches use ECMAScript Unicode syntax in a cancellable worker with a
 matching deadline and a worker old-generation heap. Deadline or cancellation
 stops matching while leaving the Hopper API available. The deadline is
@@ -383,6 +402,11 @@ sites remain distinct. A static route does not establish runtime reachability.
 Swift/closure dispatch requires resolved static references; unavailable ABI
 forms remain unknown.
 
+Encoded Objective-C method records retain their own locations in dispatch
+Evidence. A handler node uses the implementation address with an unknown file
+offset unless that function already has an independently located graph node;
+the method-record offset does not locate the handler's code.
+
 Incomplete reference kinds do not discard known typed calls or partial
 provider call references. A `provider_call` with `observed` resolution retains
 the provider-reported call site and endpoint without claiming generic direct,
@@ -428,7 +452,12 @@ Within one value trace, each exact procedure selector is attempted at most once,
 including failed analyses. A failed seed returns the provider's original typed
 failure and corrective guidance; it cannot establish a trace or target identity.
 Failures while expanding a validated seed remain unknown observations and do not
-establish canonical function identity. A successful seed without a native def-use
+establish canonical function identity. A cancelled expansion read returns the
+cancellation. An expansion read whose provider cannot confirm cleanup of
+resources it owns stops the trace: the result is that provider's own
+`cleanup_incomplete` failure with its resources and any bound cleanup retry, and the work
+admitted before the stop is retained as partial Evidence with a located unknown
+at the stopping read. A successful seed without a native def-use
 model still returns its procedure and an explicit coverage unknown. A new trace
 can retry failed selectors.
 
@@ -437,8 +466,9 @@ page omits retained nodes, and the edges they source, before `offset` or after
 `next_offset`. Every page repeats `traversal`: `complete`, or `truncated` with
 `reasons` naming each observed cause: the `max_depth`, `max_functions`,
 `max_call_sites`, `max_nodes` or `max_edges` budget, the fixed
-`node_payload_bytes` allowance, `provider_pcode_truncated`, or
-`ambiguous_call_candidates`. Each cause also has a located unknown. A budget is
+`node_payload_bytes` allowance, `provider_pcode_truncated`,
+`ambiguous_call_candidates`, or, only in a stopped trace's partial Evidence,
+`cleanup_incomplete`. Each cause also has a located unknown. A budget is
 reported only after it rejects work, so an exactly filled budget is not
 truncation; a depth or candidate limit is reported only when no other path
 analyzes or queues that destination. `total_nodes` and `total_edges` count
@@ -993,7 +1023,9 @@ with a path-specific gap and incomplete capture coverage. Null alone does not
 establish that an accessibility value was successfully read or absent.
 
 An initial PID or window mismatch is `invalid_request` on `pid` or
-`window_id` and leaves provider availability intact. An unavailable executable
+`window_id` and leaves provider availability intact. A selected window that
+exists off screen is not a mismatch: selection succeeds and the permission and
+capture checks report their own outcome. An unavailable executable
 URL or unresolved executable path leaves identity unknown and reports
 `capability_unavailable`. Identity or window loss
 after the initial capture remains an observation failure with collected facts.

docs/readme/README_ar.md

@@ -20,6 +20,8 @@
 
 **[الموقع](https://rea.tools/ar/) · [الأدلة](https://rea.tools/ar/guides/) · [دراسات الحالة](https://rea.tools/ar/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="انضم إلى مجتمع REA على Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [البدء السريع](#البدء-السريع) · [كيف يعمل REA](#كيف-يعمل-rea) · [ما الذي يمكنك تحليله](#ما-الذي-يمكنك-تحليله) · [دراسات الحالة](#دراسات-الحالة) · [إنشاء بيئات التعلم المعزز](#build-rl-environments) · [الأسئلة الشائعة](#الأسئلة-الشائعة) · [أدوات GNU objdump وRizin](../reverse-engineering-tools.md) · [التوثيق](#التوثيق)
 
 <code>npx rea-agents setup</code>

docs/readme/README_de.md

@@ -20,6 +20,8 @@
 
 **[Website](https://rea.tools/de/) · [Anleitungen](https://rea.tools/de/guides/) · [Fallstudien](https://rea.tools/de/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Der REA-Community auf Discord beitreten" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Schnellstart](#schnellstart) · [So funktioniert REA](#so-funktioniert-rea) · [Was du analysieren kannst](#was-du-analysieren-kannst) · [Fallstudien](#fallstudien) · [RL-Umgebungen erstellen](#build-rl-environments) · [Häufige Fragen](#häufige-fragen) · [Dokumentation](#dokumentation)
 
 <code>npx rea-agents setup</code>

docs/readme/README_el.md

@@ -19,6 +19,8 @@
 
 **[Ιστότοπος](https://rea.tools/) · [Οδηγοί](https://rea.tools/guides/) · [Μελέτες περίπτωσης](https://rea.tools/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Έλα στην κοινότητα REA στο Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Γρήγορη εκκίνηση](#γρήγορη-εκκίνηση) · [Πώς λειτουργεί το REA](#πώς-λειτουργεί-το-rea) · [Τι μπορείς να αναλύσεις](#τι-μπορείς-να-αναλύσεις) · [Μελέτες περίπτωσης](#μελέτες-περίπτωσης) · [Δημιουργία περιβαλλόντων RL](#build-rl-environments) · [Συχνές ερωτήσεις](#συχνές-ερωτήσεις) · [Τεκμηρίωση](#τεκμηρίωση)
 
 <code>npx rea-agents setup</code>

docs/readme/README_es.md

@@ -20,6 +20,8 @@
 
 **[Sitio web](https://rea.tools/es/) · [Guías](https://rea.tools/es/guides/) · [Casos prácticos](https://rea.tools/es/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Únete a la comunidad de REA en Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Inicio rápido](#inicio-rápido) · [Cómo funciona REA](#cómo-funciona-rea) · [Qué puedes analizar](#qué-puedes-analizar) · [Casos prácticos](#casos-prácticos) · [Crear entornos de RL](#build-rl-environments) · [Preguntas frecuentes](#preguntas-frecuentes) · [Documentación](#documentación)
 
 <code>npx rea-agents setup</code>

docs/readme/README_fa.md

@@ -20,6 +20,8 @@
 
 **[وب‌سایت](https://rea.tools/fa/) · [راهنماها](https://rea.tools/fa/guides/) · [نمونه‌های عملی](https://rea.tools/fa/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="به انجمن REA در دیسکورد بپیوندید" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [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) · [ساخت محیط‌های RL](#build-rl-environments) · [پرسش‌های متداول](#پرسش‌های-متداول) · [مستندات](#مستندات)
 
 <code>npx rea-agents setup</code>

docs/readme/README_fr.md

@@ -20,6 +20,8 @@
 
 **[Site web](https://rea.tools/fr/) · [Guides](https://rea.tools/fr/guides/) · [Études de cas](https://rea.tools/fr/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Rejoignez la communauté REA sur Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Démarrage rapide](#démarrage-rapide) · [Comment fonctionne REA](#comment-fonctionne-rea) · [Ce que vous pouvez analyser](#ce-que-vous-pouvez-analyser) · [Études de cas](#études-de-cas) · [Créer des environnements RL](#build-rl-environments) · [Questions fréquentes](#questions-fréquentes) · [Documentation](#documentation)
 
 <code>npx rea-agents setup</code>

docs/readme/README_id.md

@@ -20,6 +20,8 @@
 
 **[Situs Web](https://rea.tools/id/) · [Panduan](https://rea.tools/id/guides/) · [Contoh Penggunaan](https://rea.tools/id/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Bergabunglah dengan komunitas REA di Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Mulai Cepat](#mulai-cepat) · [Cara Kerja REA](#cara-kerja-rea) · [Hal yang Dapat Anda Analisis](#hal-yang-dapat-anda-analisis) · [Contoh Penggunaan](#contoh-penggunaan) · [Membangun lingkungan RL](#build-rl-environments) · [FAQ](#faq) · [Dokumentasi](#dokumentasi)
 
 <code>npx rea-agents setup</code>

docs/readme/README_it.md

@@ -20,6 +20,8 @@
 
 **[Sito web](https://rea.tools/it/) · [Guide](https://rea.tools/it/guides/) · [Showcase](https://rea.tools/it/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Unisciti alla community REA su Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Avvio rapido](#avvio-rapido) · [Come funziona REA](#come-funziona-rea) · [Cosa puoi analizzare](#cosa-puoi-analizzare) · [Showcase](#showcase) · [Creare ambienti RL](#build-rl-environments) · [FAQ](#faq) · [Documentazione](#documentazione)
 
 <code>npx rea-agents setup</code>

docs/readme/README_ja.md

@@ -20,6 +20,8 @@
 
 **[ウェブサイト](https://rea.tools/ja/) · [ガイド](https://rea.tools/ja/guides/) · [事例](https://rea.tools/ja/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="DiscordのREAコミュニティに参加" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [クイックスタート](#クイックスタート) · [REA の仕組み](#rea-の仕組み) · [分析できる対象](#分析できる対象) · [事例](#事例) · [RL 環境を構築する](#build-rl-environments) · [よくある質問](#よくある質問) · [ドキュメント](#ドキュメント)
 
 <code>npx rea-agents setup</code>

docs/readme/README_ko.md

@@ -20,6 +20,8 @@
 
 **[웹사이트](https://rea.tools/ko/) · [가이드](https://rea.tools/ko/guides/) · [사례](https://rea.tools/ko/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Discord에서 REA 커뮤니티에 참여하세요" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [빠른 시작](#빠른-시작) · [REA의 작동 방식](#rea의-작동-방식) · [분석할 수 있는 대상](#분석할-수-있는-대상) · [사례](#사례) · [RL 환경 만들기](#build-rl-environments) · [자주 묻는 질문](#자주-묻는-질문) · [문서](#문서)
 
 <code>npx rea-agents setup</code>

docs/readme/README_ml.md

@@ -20,6 +20,8 @@
 
 **[വെബ്സൈറ്റ്](https://rea.tools/) · [ഗൈഡുകൾ](https://rea.tools/guides/) · [ഉദാഹരണങ്ങൾ](https://rea.tools/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Discord-ൽ REA കമ്മ്യൂണിറ്റിയിൽ ചേരുക" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [ദ്രുതാരംഭം](#ദ്രുതാരംഭം) · [REA എങ്ങനെ പ്രവർത്തിക്കുന്നു](#rea-എങ്ങനെ-പ്രവർത്തിക്കുന്നു) · [എന്തൊക്കെ വിശകലനം ചെയ്യാം](#എന്തൊക്കെ-വിശകലനം-ചെയ്യാം) · [ഉദാഹരണങ്ങൾ](#ഉദാഹരണങ്ങൾ) · [RL പരിതസ്ഥിതികൾ നിർമ്മിക്കുക](#build-rl-environments) · [പതിവുചോദ്യങ്ങൾ](#പതിവുചോദ്യങ്ങൾ) · [ഡോക്യുമെന്റേഷൻ](#ഡോക്യുമെന്റേഷൻ)
 
 <code>npx rea-agents setup</code>

docs/readme/README_pl.md

@@ -20,6 +20,8 @@
 
 **[Strona internetowa](https://rea.tools/pl/) · [Przewodniki](https://rea.tools/pl/guides/) · [Przykłady zastosowań](https://rea.tools/pl/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Dołącz do społeczności REA na Discordzie" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Szybki start](#szybki-start) · [Jak działa REA](#jak-działa-rea) · [Co możesz analizować](#co-możesz-analizować) · [Przykłady zastosowań](#przykłady-zastosowań) · [Tworzenie środowisk RL](#build-rl-environments) · [FAQ](#faq) · [Dokumentacja](#dokumentacja)
 
 <code>npx rea-agents setup</code>

docs/readme/README_pt-BR.md

@@ -20,6 +20,8 @@
 
 **[Site](https://rea.tools/pt-BR/) · [Guias](https://rea.tools/pt-BR/guides/) · [Casos de uso](https://rea.tools/pt-BR/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Participe da comunidade REA no Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Início rápido](#início-rápido) · [Como o REA funciona](#como-o-rea-funciona) · [O que você pode analisar](#o-que-você-pode-analisar) · [Casos de uso](#casos-de-uso) · [Criar ambientes de RL](#build-rl-environments) · [Perguntas frequentes](#perguntas-frequentes) · [Documentação](#documentação)
 
 <code>npx rea-agents setup</code>

docs/readme/README_ro.md

@@ -20,6 +20,8 @@
 
 **[Site web](https://rea.tools/) · [Ghiduri](https://rea.tools/guides/) · [Studii de caz](https://rea.tools/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Alătură-te comunității REA pe Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Pornire rapidă](#pornire-rapidă) · [Cum funcționează REA](#cum-funcționează-rea) · [Ce poți analiza](#ce-poți-analiza) · [Studii de caz](#studii-de-caz) · [Construiește medii RL](#build-rl-environments) · [Întrebări frecvente](#întrebări-frecvente) · [Documentație](#documentație)
 
 <code>npx rea-agents setup</code>

docs/readme/README_ru.md

@@ -20,6 +20,8 @@
 
 **[Сайт](https://rea.tools/ru/) · [Руководства](https://rea.tools/ru/guides/) · [Примеры](https://rea.tools/ru/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Присоединяйтесь к сообществу REA в Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Быстрый старт](#быстрый-старт) · [Как работает REA](#как-работает-rea) · [Что можно анализировать](#что-можно-анализировать) · [Примеры](#примеры) · [Создание RL-окружений](#build-rl-environments) · [Частые вопросы](#частые-вопросы) · [Документация](#документация)
 
 <code>npx rea-agents setup</code>

docs/readme/README_th.md

@@ -20,6 +20,8 @@
 
 **[เว็บไซต์](https://rea.tools/th/) · [คู่มือ](https://rea.tools/th/guides/) · [ตัวอย่างการใช้งาน](https://rea.tools/th/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="เข้าร่วมชุมชน REA บน Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [เริ่มต้นอย่างรวดเร็ว](#เริ่มต้นอย่างรวดเร็ว) · [REA ทำงานอย่างไร](#rea-ทำงานอย่างไร) · [สิ่งที่วิเคราะห์ได้](#สิ่งที่วิเคราะห์ได้) · [ตัวอย่างการใช้งาน](#ตัวอย่างการใช้งาน) · [สร้างสภาพแวดล้อม RL](#build-rl-environments) · [คำถามที่พบบ่อย](#คำถามที่พบบ่อย) · [เอกสาร](#เอกสาร)
 
 <code>npx rea-agents setup</code>

docs/readme/README_tr.md

@@ -20,6 +20,8 @@
 
 **[Web sitesi](https://rea.tools/tr/) · [Rehberler](https://rea.tools/tr/guides/) · [Örnek çalışmalar](https://rea.tools/tr/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Discord'da REA topluluğuna katılın" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Hızlı başlangıç](#hızlı-başlangıç) · [REA nasıl çalışır](#rea-nasıl-çalışır) · [Neleri analiz edebilirsiniz](#neleri-analiz-edebilirsiniz) · [Örnek çalışmalar](#örnek-çalışmalar) · [RL ortamları oluşturun](#build-rl-environments) · [Sık sorulan sorular](#sık-sorulan-sorular) · [GNU objdump ve Rizin araçları](../reverse-engineering-tools.md) · [Belgeler](#belgeler)
 
 <code>npx rea-agents setup</code>

docs/readme/README_uk.md

@@ -20,6 +20,8 @@
 
 **[Сайт](https://rea.tools/uk/) · [Посібники](https://rea.tools/uk/guides/) · [Приклади](https://rea.tools/uk/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Приєднуйтеся до спільноти REA в Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Швидкий старт](#швидкий-старт) · [Як працює REA](#як-працює-rea) · [Що можна аналізувати](#що-можна-аналізувати) · [Приклади](#приклади) · [Створення RL-середовищ](#build-rl-environments) · [Поширені запитання](#поширені-запитання) · [Документація](#документація)
 
 <code>npx rea-agents setup</code>

docs/readme/README_ur-Latn.md

@@ -20,6 +20,8 @@
 
 **[Website](https://rea.tools/) · [Guides](https://rea.tools/guides/) · [Showcases](https://rea.tools/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Discord par REA community mein shamil hon" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Jaldi shuru karein](#jaldi-shuru-karein) · [REA kaise kaam karta hai](#rea-kaise-kaam-karta-hai) · [Kya kuch analyze kar sakte hain](#kya-kuch-analyze-kar-sakte-hain) · [Amli misaalein](#amli-misaalein) · [RL environments banayein](#build-rl-environments) · [FAQ](#faq) · [Documentation](#documentation)
 
 <code>npx rea-agents setup</code>

docs/readme/README_vi.md

@@ -20,6 +20,8 @@
 
 **[Trang web](https://rea.tools/vi/) · [Hướng dẫn](https://rea.tools/vi/guides/) · [Các nghiên cứu điển hình](https://rea.tools/vi/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="Tham gia cộng đồng REA trên Discord" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [Bắt đầu nhanh](#bắt-đầu-nhanh) · [Cách REA hoạt động](#cách-rea-hoạt-động) · [Những gì bạn có thể phân tích](#những-gì-bạn-có-thể-phân-tích) · [Các nghiên cứu điển hình](#các-nghiên-cứu-điển-hình) · [Xây dựng môi trường RL](#build-rl-environments) · [Câu hỏi thường gặp](#câu-hỏi-thường-gặp) · [Tài liệu](#tài-liệu)
 
 <code>npx rea-agents setup</code>

docs/readme/README_zh-TW.md

@@ -20,6 +20,8 @@
 
 **[網站](https://rea.tools/zh-TW/) · [指南](https://rea.tools/zh-TW/guides/) · [案例](https://rea.tools/zh-TW/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="加入 REA 的 Discord 社群" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [快速開始](#快速開始) · [REA 的工作原理](#rea-的工作原理) · [可以分析什麼](#可以分析什麼) · [案例](#案例) · [建立 RL 環境](#build-rl-environments) · [常見問題](#常見問題) · [文件](#文件)
 
 <code>npx rea-agents setup</code>

docs/readme/README_zh.md

@@ -20,6 +20,8 @@
 
 **[网站](https://rea.tools/zh-CN/) · [指南](https://rea.tools/zh-CN/guides/) · [案例](https://rea.tools/zh-CN/showcase/)**
 
+<a href="https://discord.com/invite/GkcryMnJDM"><img alt="加入 REA 的 Discord 社区" src="https://img.shields.io/badge/Join%20us%20on%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" height="40"></a>
+
 [快速开始](#快速开始) · [REA 的工作原理](#rea-的工作原理) · [可以分析什么](#可以分析什么) · [案例](#案例) · [构建 RL 环境](#build-rl-environments) · [常见问题](#常见问题) · [GNU objdump 和 Rizin 工具](../reverse-engineering-tools.md) · [文档](#文档)
 
 <code>npx rea-agents setup</code>

docs/reverse-engineering-tools.md

@@ -14,7 +14,13 @@ can run.
 
 ## GNU objdump
 
-`inspect_with_objdump` and `rea inspect-with-objdump` capture stdout and stderr as inline Evidence. Artifact inspection Evidence records SHA-256 digests before and after the provider call and whether those observed digests differ; this cannot prove the exact bytes consumed if the artifact changes during execution. Choose one artifact operation: `file_headers`, `section_headers`, `symbols`, `relocations`, `dwarf`, `disassemble`, or `disassemble_all`. `disassemble` maps to `-d`; `disassemble_all` maps to `-D` and has different symbol-boundary semantics. The pathless `inspect_with_objdump {"operation":"architectures"}` MCP call or `rea list-objdump-architectures` CLI command maps to `-i` and produces subjectless provider-capability Evidence describing BFD targets compiled into the installed binary, not a guarantee that a target is supported.
+REA sends GNU-style arguments to the selected command but does not verify its
+implementation or version. Evidence identifies the command profile as
+`objdump.command` and retains the selected executable; it does not claim that
+the executable is GNU objdump. Other implementations may accept some commands
+while differing in supported options and output.
+
+`inspect_with_objdump` and `rea inspect-with-objdump` capture stdout and stderr as inline Evidence. Artifact inspection Evidence records SHA-256 digests before and after the provider call and whether those observed digests differ; this cannot prove the exact bytes consumed if the artifact changes during execution. Choose one artifact operation: `file_headers`, `section_headers`, `symbols`, `relocations`, `dwarf`, `disassemble`, or `disassemble_all`. `disassemble` maps to `-d`; `disassemble_all` maps to `-D` and has different symbol-boundary semantics. The pathless `inspect_with_objdump {"operation":"architectures"}` MCP call or `rea list-objdump-architectures` CLI command maps to `-i` and produces subjectless Evidence containing the selected command’s output, not a guarantee that any target is supported. GNU objdump uses this option to list its compiled-in BFD targets.
 
 Failure to launch the selected executable reports provider unavailability, the
 configured command and the host error under `boundary: "provider-launch"`.
@@ -25,7 +31,7 @@ their rejected fields in input issues before artifact reads or provider launch.
 
 For DWARF, REA does not follow local debug links by default. Set `follow_debug_links` to `true` to pass `-WK`; external debug-file identities are not bound to Evidence when following links is enabled. The CLI flag is `--follow-debug-links`; `--no-follow-debug-links` or `--follow-debug-links=false` keeps the default. A separate `true` or `false` word after the flag is not its value. REA clears `DEBUGINFOD_URLS` in the objdump child environment to disable debuginfod network access on GNU binutils versions that do not support `-WE`. The complete raw output is authoritative; REA does not normalize it into a version-independent disassembly schema.
 
-For a known native symbol, use `operation: "disassemble", symbol: "my_function"` in MCP or `--symbol my_function` in CLI. REA passes `--disassemble=my_function` to objdump before collecting output, rather than producing the whole disassembly and filtering it afterward. Omit `symbol` to keep full code-section disassembly. Other operations reject this selector. Use a nonempty, NUL-free name with well-formed Unicode; unsupported values are rejected before reading the artifact or starting objdump. Symbol names are passed unchanged; matching, versioned symbols, aliases and the end of the disassembly follow the installed GNU objdump's semantics. Evidence records the selector and actual arguments. An empty match does not establish artifact-wide absence, and the selected observation does not contain the omitted disassembly; request the complete operation when needed.
+For a known native symbol, use `operation: "disassemble", symbol: "my_function"` in MCP or `--symbol my_function` in CLI. REA passes `--disassemble=my_function` to objdump before collecting output, rather than producing the whole disassembly and filtering it afterward. Omit `symbol` to keep full code-section disassembly. Other operations reject this selector. Use a nonempty, NUL-free name with well-formed Unicode; unsupported values are rejected before reading the artifact or starting objdump. Symbol names are passed unchanged; matching, versioned symbols, aliases and the end of the disassembly follow the selected executable's semantics. Evidence records the selector and actual arguments. An empty match does not establish artifact-wide absence, and the selected observation does not contain the omitted disassembly; request the complete operation when needed.
 
 ```sh
 rea inspect-with-objdump ./program disassemble

docs/specs/agent-environments/fidelity-compare/spec.md

@@ -86,7 +86,10 @@ credentials and resets it between sequences.
   environment refused the call), `not-served`, `no-tool` and `unmapped-id`, each with a JSON pointer and short values.
 - The report gives agreement overall, per probe and per tool, agreement ignoring item order beside it, the counts per
   category, the ids mapped and the values normalised, the records left uncompared, and each step's differences (at most
-  20 per step).
+  20 per step). Each step gives `differences_total` before that cap, `differences_omitted`, and
+  `differences_truncated`, true when entries were omitted. Equality and categories use all differences.
+  Omitted entries are not retained and `--full-output` cannot recover them; further investigation uses the original
+  capture and probe Evidence.
 
 ## Non-goals
 

docs/web-network-captures.md

@@ -97,9 +97,13 @@ it does not invent an empty string or body bytes.
 The observed Playwright 1.58.2 profile can report a bracketed IPv6
 `serverIPAddress`, forwarded from Chromium's `remoteIPAddress` through
 [Playwright's HAR tracer](https://github.com/microsoft/playwright/blob/v1.58.2/packages/playwright-core/src/server/har/harTracer.ts).
+Chromium DevTools exports (declared as `WebInspector` `537.36`, the frozen
+AppleWebKit version in Chromium's user agent) report the same bracketed form:
+[its HAR writer](https://github.com/ChromeDevTools/devtools-frontend/blob/63739e8ec45ca01b9e4be9b8b05946acdadd8f1f/front_end/models/har/Log.ts)
+strips the port but not the brackets.
 REA preserves the original spelling in reported Evidence and removes brackets
-only from its validation copy, only for this declared producer and a valid IPv6
-literal. Invalid addresses, bracketed IPv4, ports and zone identifiers remain
+only from its validation copy, only for these declared producers and a valid
+IPv6 literal. Invalid addresses, bracketed IPv4, ports and zone identifiers remain
 validation failures. Producer declarations do not authenticate origin.
 
 Native mitmproxy decoding requires a caller-supplied **mitmdump 12.2.3 on

docs/web-runtime.md

@@ -22,7 +22,7 @@ Results retain:
 - Main-document script inventory, exact session script IDs, resource start coordinates, sourceURL/source-map declarations, producer hash and independently computed UTF-8 text digest.
 - Armed-window requests from the selected main frame and selected origins, complete reported initiator objects (including stack hierarchy/descriptions), derived callsites, exact script-ID associations and unresolved async parent IDs. Source association does not establish UI causality. Callsite and listener `line_number`/`column_number` values are CDP's zero-based lines and UTF-16 columns in the resource named by `url`. With a `script_id`, that is the script's enclosing resource: the script file for an external script, or the HTML document for an inline script. `resource_start` is where the script text starts in that resource. A callsite with an empty `script_id` comes from the request's reported initiator rather than a script call frame, such as markup that the parser read in a document or a module import in a script file. It still locates a position in `url`, but has no session-script association, so it cannot be joined to captured source or `resource_start`. `trace_web_source_location` takes one-based, script-relative lines and zero-based UTF-16 columns.
 
-For an external script, where `url` is the script file, use `line = line_number + 1` and keep `column = column_number`; this also applies to an empty-`script_id` initiator position in a script file. For an inline script, which requires a `script_id` and its `resource_start`, use `line = line_number - resource_start.line_number + 1`. Subtract `resource_start.column_number` from `column_number` only when `line_number === resource_start.line_number`; keep the reported column on later lines. Columns remain zero-based. This conversion requires known source-start and column coordinates; preserve unavailable coordinates as unknown.
+For an external script, where `url` is the script file, use `line = line_number + 1` and keep `column = column_number`; this also applies to an empty-`script_id` initiator position in a script file. For an inline script, which requires a `script_id` and its `resource_start`, use `line = line_number - resource_start.line_number + 1`. Subtract `resource_start.column_number` from `column_number` only when `line_number === resource_start.line_number`; keep the reported column on later lines. Columns remain zero-based. This conversion requires known source-start and column coordinates; preserve unavailable coordinates as unknown. Chrome reports a negative start column for some inline scripts, such as one in a helper frame, and positions on that script's first line include it. REA keeps those columns as reported rather than failing the observation; a negative column does not locate a column in `url`, but the subtraction above still gives the script-relative column.
 
 The coverage counter interval starts at backend acceptance before the armed receipt and ends at the resetting sample after the local timer. Counts may include execution during these command intervals; request events are restricted to the locally armed window. The result retains both backend timestamps and the distinct local clock.
 

patches/release-please+17.3.0.patch

@@ -81,6 +81,51 @@
                  // In this case, there is no squashed merge commit. This could be a simple
                  // merge commit, a rebase merge commit, or a direct commit to the branch.
                  // Fallback to fetching the list of commits from the REST API. In the future
+@@ -1154,15 +1154,40 @@
+      * @returns {string} HTML URL of the new file
+      */
+     async createFileOnNewBranch(filename, contents, newBranchName, baseBranchName) {
+-        // create or update new branch to match base branch
+-        await this.forkBranch(newBranchName, baseBranchName);
+-        // use the single file upload API
++        // Preserve existing notes until the replacement file is committed.
++        // A failed write must not leave an open proposal pointing at a missing file.
++        if (!(await this.getBranchSha(newBranchName))) {
++            const baseSha = await this.getBranchSha(baseBranchName);
++            if (!baseSha) {
++                throw new Error(`Unable to find base branch: ${baseBranchName}`);
++            }
++            await this.createNewBranch(newBranchName, baseSha);
++        }
++        let existingSha;
++        try {
++            const { data } = await this.octokit.repos.getContent({
++                owner: this.repository.owner,
++                repo: this.repository.repo,
++                path: filename,
++                ref: newBranchName,
++            });
++            if (Array.isArray(data) || data.type !== 'file' || !data.sha) {
++                throw new Error(`Expected a file at ${filename} on branch: ${newBranchName}`);
++            }
++            existingSha = data.sha;
++        }
++        catch (error) {
++            if (!(error && error.status === 404)) {
++                throw error;
++            }
++        }
+         const { data: { content }, } = await this.octokit.repos.createOrUpdateFileContents({
+             owner: this.repository.owner,
+             repo: this.repository.repo,
+             path: filename,
+             // contents need to be base64 encoded
+-            content: Buffer.from(contents, 'binary').toString('base64'),
++            content: Buffer.from(contents, 'utf8').toString('base64'),
++            ...(existingSha ? { sha: existingSha } : {}),
+             message: 'Saving release notes',
+             branch: newBranchName,
+         });
 --- a/build/src/github.d.ts
 +++ b/build/src/github.d.ts
 @@ -47,6 +47,8 @@

scripts/ci/pr-readiness.d.mts

@@ -3,6 +3,7 @@ export interface ReadinessPullRequest {
   readonly state: "open" | "closed";
   readonly draft: boolean;
   readonly body: string | null;
+  readonly user: { readonly id: number } | null;
   readonly head: { readonly sha: string };
   readonly base: { readonly ref: string };
 }
@@ -53,7 +54,7 @@ export interface ReadinessGitHub {
   >;
 }
 export function evaluateReadiness(
-  pull: Pick<ReadinessPullRequest, "body" | "draft">,
+  pull: Pick<ReadinessPullRequest, "body" | "draft" | "user">,
   template: string,
 ): ReadinessResult;
 export function runReadinessCheck(input: {

scripts/ci/pr-readiness.mjs

@@ -2,6 +2,14 @@ import { readFile } from "node:fs/promises";
 
 const result = (state, description) => ({ state, description });
 
+// GitHub account IDs remain stable when a login changes.
+const EXEMPT_AUTHOR_IDS = new Set([
+  76467478, // morluto
+  208281322, // kaoru0822-kitauji
+  161784452, // N0zoM1z0
+  342580529, // gensokyo-dev
+]);
+
 const OPEN_HEADS_QUERY = `query ReadinessOpenHeads($owner: String!, $name: String!, $after: String) {
   repository(owner: $owner, name: $name) {
     pullRequests(first: 100, after: $after, states: OPEN, orderBy: {field: CREATED_AT, direction: ASC}) {
@@ -60,11 +68,11 @@ function readinessLines(markdown) {
   let comment = false;
   for (const source of lines) {
     if (fence) {
+      const closing = /^ {0,3}(`+|~+)\s*$/u.exec(source);
       if (
-        new RegExp(
-          `^ {0,3}${fence.character}{${fence.length},}\\s*$`,
-          "u",
-        ).test(source)
+        closing !== null &&
+        closing[1][0] === fence.character &&
+        closing[1].length >= fence.length
       )
         fence = undefined;
       continue;
@@ -105,6 +113,11 @@ export function evaluateReadiness(pull, template) {
       "pending",
       "Draft PR: complete Author readiness before requesting review.",
     );
+  if (EXEMPT_AUTHOR_IDS.has(pull.user?.id))
+    return result(
+      "success",
+      "PR author is exempt from the readiness declaration.",
+    );
   const expected = readinessLines(template).flatMap((line) => {
     const match = /^- \[ \] (.+)$/u.exec(line);
     return match ? [match[1]] : [];
@@ -210,6 +223,7 @@ export async function runReadinessCheck({ github, context, core }) {
     latest.head.sha !== pull.head.sha ||
     latest.body !== pull.body ||
     latest.draft !== pull.draft ||
+    latest.user?.id !== pull.user?.id ||
     latest.base.ref !== pull.base.ref
   )
     verdict = result(

scripts/ci/scopes.mjs

@@ -76,6 +76,7 @@ const PATH_SCOPES = {
     "src/frida/frida-bindings.d.ts",
     "src/objdump/**",
     "src/cli/reverseEngineeringCommands.ts",
+    "src/process/ownedDebuggerCleanupError.ts",
     "src/process/ownedDebuggerSession.ts",
     "src/rizin/**",
     "src/contracts/*ToolContracts.ts",
@@ -103,6 +104,7 @@ const PATH_SCOPES = {
     "tests/boundary/providers/RizinInferiorTty.test.ts",
     "tests/boundary/providers/RizinTestInferiorTty.ts",
     "tests/boundary/process/ownedDebuggerSession.test.ts",
+    "tests/boundary/cli/debuggerOneShotExit.test.ts",
     "tests/boundary/cli/gdbStartupCleanupCli.test.ts",
     "tests/boundary/cli/reverseEngineeringCommands.test.ts",
     "tests/boundary/mcp/cutter*.test.ts",
@@ -163,6 +165,9 @@ const PATH_SCOPES = {
     "tests/boundary/mcp/nativeDemangleMcp.test.ts",
     "tests/boundary/mcp/nativeMachoSummaryMcp.test.ts",
     "tests/boundary/providers/native/nativeSignatureTargetBinding.test.ts",
+    "tests/boundary/providers/native/nativeUiOffscreenWindow.test.ts",
+    "tests/boundary/process/nativeUiVerifierLane.test.ts",
+    "tests/conformance/native/ui-offscreen.swift",
     "tests/boundary/cli/nativeSignatureCli.test.ts",
     "tests/boundary/cli/nativeMachoArm64_32Cli.test.ts",
     "tests/boundary/cli/nativeMachoEncryptionCli.test.ts",
@@ -699,7 +704,9 @@ export const CI_LANES = {
   agentenv: sameJobs(PATH_SCOPES.agentenv, ["agentenv"]),
   nativeaot: sameJobs(PATH_SCOPES.nativeaot, ["nativeaot"]),
   ghidra_bridge: sameJobs(PATH_SCOPES.ghidra_bridge, ["ghidra_bridge"]),
-  website: sameJobs(PATH_SCOPES.website, ["website"]),
+  website: sameJobs(PATH_SCOPES.website, ["website"], {
+    earlyInputs: [".agents/skills/reverse-engineer-anything/**"],
+  }),
 };
 // An input owner omitted from the registry must never silently lose a lane.
 for (const [name, paths] of Object.entries(PATH_SCOPES))

scripts/electron-active-hook.cjs

@@ -78,7 +78,6 @@ let correlationSequence = 0;
 
 const shape = (value, depth = 0) => {
   if (depth > 2 || value === null) return value === null ? "null" : "nested";
-  if (arrayIsArray(value)) return "array";
   switch (typeof value) {
     case "string":
       return "string";
@@ -91,7 +90,12 @@ const shape = (value, depth = 0) => {
     case "function":
       return "function";
     case "object":
-      return "object";
+      try {
+        return arrayIsArray(value) ? "array" : "object";
+      } catch {
+        // Revoked proxies retain their typeof identity but reject IsArray.
+        return "object";
+      }
     default:
       return "unknown";
   }
@@ -1321,20 +1325,25 @@ try {
   boundaries.patchApplicationEffects(electron.app);
   boundaries.patchNativeAddonLoading();
   boundaries.patchChildProcess();
-  process.once("uncaughtException", (cause, origin) => {
-    recordRuntime("process-lifecycle", "uncaught-exception", "observed", {
-      argument_shape_values: [cause, origin],
-      error: true,
-    });
-    throw cause;
-  });
-  process.once("unhandledRejection", (reason, promise) => {
-    recordRuntime("process-lifecycle", "unhandled-rejection", "observed", {
-      argument_shape_values: [reason, promise],
-      error: true,
-    });
-    throw reason;
-  });
+  // Observe dispatch without adding listeners: their presence changes Node's
+  // rejection fallback and Electron's uncaught-exception listener-count policy.
+  const processEmit = process.emit;
+  process.emit = function observedProcessEmit(event, ...args) {
+    if (this === process) {
+      const kind =
+        event === "unhandledRejection"
+          ? "unhandled-rejection"
+          : event === "uncaughtExceptionMonitor"
+            ? "uncaught-exception"
+            : null;
+      if (kind !== null)
+        recordRuntime("process-lifecycle", kind, "observed", {
+          argument_shape_values: args,
+          error: true,
+        });
+    }
+    return applyFunction(processEmit, this, arguments);
+  };
 } catch (cause) {
   globalThis.__reaElectronActiveHookError = true;
   recordRuntime(

scripts/fixtures/flutter-gallery-expectations.json

@@ -3,19 +3,19 @@
   "snapshot_hash": "65817c30a78bb44c3dc3771876b6010a",
   "engines": {
     "x86_64": {
-      "dart_version": null,
+      "dart_version": "2.14.0-151.0.dev (dev) (Tue May 25 10:58:39 2021 -0700)",
       "toolchain_lines": [
         "Android (5900059 based on r365631c) clang version 9.0.8 (https://android.googlesource.com/toolchain/llvm-project 207d7abc1a2abf3ef8d4301736d6a7ebc224a290) (based on LLVM 9.0.8svn)"
       ]
     },
     "armeabi-v7a": {
-      "dart_version": null,
+      "dart_version": "2.14.0-151.0.dev (dev) (Tue May 25 10:58:39 2021 -0700)",
       "toolchain_lines": [
         "Android (5900059 based on r365631c) clang version 9.0.8 (https://android.googlesource.com/toolchain/llvm-project 207d7abc1a2abf3ef8d4301736d6a7ebc224a290) (based on LLVM 9.0.8svn)"
       ]
     },
     "arm64-v8a": {
-      "dart_version": null,
+      "dart_version": "2.14.0-151.0.dev (dev) (Tue May 25 10:58:39 2021 -0700)",
       "toolchain_lines": [
         "Android (5900059 based on r365631c) clang version 9.0.8 (https://android.googlesource.com/toolchain/llvm-project 207d7abc1a2abf3ef8d4301736d6a7ebc224a290) (based on LLVM 9.0.8svn)"
       ]

scripts/lib/browser-fixture-http.mjs

@@ -0,0 +1,26 @@
+import { createServer } from "node:http";
+
+/** Interpret request targets once, before a fixture routes or forwards them. */
+export const createBrowserFixtureHttpServer = (handleRequest) =>
+  createServer((request, response) => {
+    let target;
+    try {
+      // A leading // is an origin-form path, not a replacement authority.
+      const requested = new URL(
+        request.url?.startsWith("/")
+          ? `http://fixture.invalid${request.url}`
+          : request.url,
+      );
+      if (requested.protocol !== "http:" && requested.protocol !== "https:")
+        throw new TypeError("Unsupported fixture request protocol");
+      target = { pathname: requested.pathname, search: requested.search };
+    } catch {
+      response.writeHead(400, {
+        "Content-Type": "text/plain; charset=utf-8",
+        "X-Content-Type-Options": "nosniff",
+      });
+      response.end("Invalid HTTP request target");
+      return;
+    }
+    handleRequest(request, response, target);
+  });

scripts/lib/browser-scenario-service-worker-setup-e2e.mjs

@@ -1,11 +1,12 @@
 import assert from "node:assert/strict";
 import { once } from "node:events";
-import { createServer, request } from "node:http";
+import { request } from "node:http";
 import { connect } from "node:net";
 
 import { createBrowserScenarioProvider } from "../../dist/composition/browserScenario.js";
 import { browserScenarioSchema } from "../../dist/domain/browserScenario.js";
 import { PlaywrightScenarioServiceWorkers } from "../../dist/browser/PlaywrightScenarioServiceWorkers.js";
+import { createBrowserFixtureHttpServer } from "./browser-fixture-http.mjs";
 
 const beforeDeadline = async (promise, label, timeoutMs = 2_000) => {
   let timer;
@@ -26,7 +27,7 @@ const beforeDeadline = async (promise, label, timeoutMs = 2_000) => {
 
 // Real loopback proxy: Playwright's first transport is forwarded to Chrome,
 // then only the policy owner's discovery/handshake is deliberately withheld.
-const startProxy = async (endpoint, mode, ownerOnly = false) => {
+export const startProxy = async (endpoint, mode, ownerOnly = false) => {
   const upstream = new URL(endpoint);
   const sockets = new Set();
   let versions = 0;
@@ -47,38 +48,44 @@ const startProxy = async (endpoint, mode, ownerOnly = false) => {
     socket.once("close", () => closed());
     entered();
   };
-  const server = createServer((incoming, response) => {
-    const versionRequest = incoming.url.replace(/\/$/, "") === "/json/version";
-    if (versionRequest) versions += 1;
-    if (
-      enabled &&
-      mode === "http-discovery" &&
-      versions === 2 &&
-      versionRequest
-    ) {
-      stall(incoming.socket);
-      return;
-    }
-    const outgoing = request(new URL(incoming.url, upstream), (received) => {
-      const chunks = [];
-      received.on("data", (chunk) => chunks.push(chunk));
-      received.on("end", () => {
-        const content = Buffer.concat(chunks).toString("utf8");
-        response.writeHead(received.statusCode, {
-          "content-type": "application/json",
+  const server = createBrowserFixtureHttpServer(
+    (incoming, response, target) => {
+      const versionRequest =
+        target.pathname.replace(/\/$/, "") === "/json/version";
+      if (versionRequest) versions += 1;
+      if (
+        enabled &&
+        mode === "http-discovery" &&
+        versions === 2 &&
+        versionRequest
+      ) {
+        stall(incoming.socket);
+        return;
+      }
+      const destination = new URL(upstream);
+      destination.pathname = target.pathname;
+      destination.search = target.search;
+      const outgoing = request(destination, (received) => {
+        const chunks = [];
+        received.on("data", (chunk) => chunks.push(chunk));
+        received.on("end", () => {
+          const content = Buffer.concat(chunks).toString("utf8");
+          response.writeHead(received.statusCode, {
+            "content-type": "application/json",
+          });
+          response.end(
+            content.replaceAll(
+              `127.0.0.1:${upstream.port}`,
+              `127.0.0.1:${server.address().port}`,
+            ),
+          );
         });
-        response.end(
-          content.replaceAll(
-            `127.0.0.1:${upstream.port}`,
-            `127.0.0.1:${server.address().port}`,
-          ),
-        );
       });
-    });
-    outgoing.on("error", () => response.destroy());
-    response.on("close", () => outgoing.destroy());
-    outgoing.end();
-  });
+      outgoing.on("error", () => response.destroy());
+      response.on("close", () => outgoing.destroy());
+      outgoing.end();
+    },
+  );
   server.on("connection", (socket) => {
     sockets.add(socket);
     socket.on("close", () => sockets.delete(socket));

scripts/lib/browser-scenario-service-workers-e2e.mjs

@@ -1,6 +1,5 @@
 import assert from "node:assert/strict";
 import { once } from "node:events";
-import { createServer } from "node:http";
 import { setTimeout as delay } from "node:timers/promises";
 
 import { chromium } from "playwright-core";
@@ -11,6 +10,7 @@ import { createBrowserScenarioProvider } from "../../dist/composition/browserSce
 import { browserScenarioSchema } from "../../dist/domain/browserScenario.js";
 import { runScenarioCli } from "./browser-scenario-verifier.mjs";
 import { verifyScenarioPolicySetup } from "./browser-scenario-service-worker-setup-e2e.mjs";
+import { createBrowserFixtureHttpServer } from "./browser-fixture-http.mjs";
 
 const diagnosticError = (error) =>
   error instanceof Error
@@ -26,38 +26,38 @@ const diagnosticError = (error) =>
       }
     : { kind: typeof error };
 
-const startSite = async () => {
+export const startSite = async () => {
   let otherOrigin;
-  const server = createServer((request, response) => {
-    const path = new URL(request.url, "http://localhost").pathname;
-    if (path === "/sw.js") {
-      response.setHeader("content-type", "application/javascript");
-      response.end(`self.addEventListener("install", () => self.skipWaiting());
+  const server = createBrowserFixtureHttpServer(
+    (_request, response, { pathname: path }) => {
+      if (path === "/sw.js") {
+        response.setHeader("content-type", "application/javascript");
+        response.end(`self.addEventListener("install", () => self.skipWaiting());
         self.addEventListener("activate", event => event.waitUntil(clients.claim()));
         self.addEventListener("fetch", event => {
           if (new URL(event.request.url).pathname === "/probe")
             event.respondWith(new Response("worker"));
         });`);
-    } else if (path === "/probe") {
-      response.end("network");
-    } else if (path === "/failed-navigation") {
-      response.destroy();
-    } else if (path === "/stalled-image.png") {
-      response.writeHead(200, { "content-type": "image/png" });
-      response.flushHeaders();
-    } else if (path === "/pending-image" || path === "/pending-child") {
-      response.setHeader("content-type", "text/html");
-      response.end(
-        '<!doctype html><img id="pending-image" src="/stalled-image.png">',
-      );
-    } else if (path === "/pending-frame") {
-      response.setHeader("content-type", "text/html");
-      response.end(
-        '<!doctype html><iframe id="pending-frame" src="/pending-child"></iframe>',
-      );
-    } else {
-      response.setHeader("content-type", "text/html");
-      response.end(`<!doctype html><p id="state">ready</p>
+      } else if (path === "/probe") {
+        response.end("network");
+      } else if (path === "/failed-navigation") {
+        response.destroy();
+      } else if (path === "/stalled-image.png") {
+        response.writeHead(200, { "content-type": "image/png" });
+        response.flushHeaders();
+      } else if (path === "/pending-image" || path === "/pending-child") {
+        response.setHeader("content-type", "text/html");
+        response.end(
+          '<!doctype html><img id="pending-image" src="/stalled-image.png">',
+        );
+      } else if (path === "/pending-frame") {
+        response.setHeader("content-type", "text/html");
+        response.end(
+          '<!doctype html><iframe id="pending-frame" src="/pending-child"></iframe>',
+        );
+      } else {
+        response.setHeader("content-type", "text/html");
+        response.end(`<!doctype html><p id="state">ready</p>
         <a id="navigate" href="${otherOrigin}/">Navigate</a>
         ${path === "/frame" ? `<iframe src="${otherOrigin}/"></iframe>` : ""}
         ${
@@ -72,8 +72,9 @@ const startSite = async () => {
         </script>`
             : ""
         }`);
-    }
-  });
+      }
+    },
+  );
   server.listen(0, "127.0.0.1");
   await once(server, "listening");
   return {

scripts/lib/browser-scenario-storage-e2e.mjs

@@ -1,5 +1,4 @@
 import assert from "node:assert/strict";
-import { createServer } from "node:http";
 import { once } from "node:events";
 import { setTimeout as delay } from "node:timers/promises";
 
@@ -11,6 +10,7 @@ import {
   scenarioProfiles,
 } from "./browser-scenario-verifier.mjs";
 import { withVerifierCleanup } from "./verifier-cleanup.mjs";
+import { createBrowserFixtureHttpServer } from "./browser-fixture-http.mjs";
 
 const scenarioFor = (browser, origin) =>
   browserScenarioSchema.parse({
@@ -149,28 +149,29 @@ const storagePage = (kind, port) => {
   </script></body></html>`;
 };
 
-const startStorageSite = async () => {
+export const startStorageSite = async () => {
   let port = 0;
-  const server = createServer((request, response) => {
-    const path = new URL(request.url ?? "/", "http://127.0.0.1").pathname;
-    const kind =
-      path === "/storage-main"
-        ? "main"
-        : path === "/storage-frame"
-          ? "frame"
-          : path === "/storage-debugger"
-            ? "debugger"
-            : path === "/storage-top"
-              ? "top"
-              : undefined;
-    if (kind === undefined) {
-      response.writeHead(404).end();
-      return;
-    }
-    response.setHeader("content-type", "text/html; charset=utf-8");
-    response.end(storagePage(kind, port));
-  });
-  server.listen(0, "::");
+  const server = createBrowserFixtureHttpServer(
+    (_request, response, { pathname: path }) => {
+      const kind =
+        path === "/storage-main"
+          ? "main"
+          : path === "/storage-frame"
+            ? "frame"
+            : path === "/storage-debugger"
+              ? "debugger"
+              : path === "/storage-top"
+                ? "top"
+                : undefined;
+      if (kind === undefined) {
+        response.writeHead(404).end();
+        return;
+      }
+      response.setHeader("content-type", "text/html; charset=utf-8");
+      response.end(storagePage(kind, port));
+    },
+  );
+  server.listen(0, "127.0.0.1");
   await once(server, "listening");
   port = server.address().port;
   return {

scripts/lib/kuna-reference-verifier.mjs

@@ -114,6 +114,46 @@ export const verifyKunaReferences = async ({
         );
       }
       assert.equal(trace.value.truncated, false);
+      const summary = await call("trace_feature", {
+        query: name,
+        detail: "summary",
+      });
+      assert.equal(
+        summary.value.parent_evidence_id,
+        trace.structured.evidence_id,
+      );
+      assert.equal(
+        summary.value.summary.references,
+        trace.value.references.length,
+      );
+      assert.equal(
+        summary.value.summary.unknowns,
+        trace.value.residual_unknowns.length,
+      );
+      for (const [collection, values] of [
+        ["matches", trace.value.matches],
+        ["references", trace.value.references],
+        ["unknowns", trace.value.residual_unknowns],
+      ]) {
+        const view = await call("inspect_analysis_view", {
+          source: {
+            kind: "retained-evidence",
+            evidence_id: trace.structured.evidence_id,
+          },
+          view: {
+            kind: "native-feature-trace",
+            collection,
+            offset: 0,
+            limit: 1,
+          },
+        });
+        assert.equal(view.value.coverage.total, values.length);
+        assert.deepEqual(
+          view.value.items.rows.map(({ value }) => value),
+          values.slice(0, 1),
+        );
+        assert.equal(view.value.items.collection, collection);
+      }
     }
     composed = {
       graph_ms: graph.ms,

scripts/lib/native-value-e2e.mjs

@@ -253,20 +253,9 @@ async function verifyNativeValueAdapters(
       functionEvidence.analysis_profile,
     );
     assert.deepEqual(functionProfile, upstreamProfile);
-    const comparison = await client.callTool(
-      {
-        name: "compare_functions",
-        arguments: {
-          left: analyzed.structuredContent,
-          right: analyzed.structuredContent,
-        },
-      },
-      { timeout: 180000 },
-    );
-    assert.notEqual(comparison.isError, true, JSON.stringify(comparison));
-    assert.notEqual(comparison.structuredContent, undefined);
-    const comparisonEvidence = parseEvidence(
-      requireMcpOutput(comparison, "compare_functions"),
+    const comparisonEvidence = await compareRetainedFunctionEvidence(
+      client,
+      functionEvidence,
     );
     assert.equal(
       comparisonEvidence.normalized_result.status,
@@ -395,21 +384,27 @@ async function verifyUnresolvedCallEvidence(client) {
   assert.equal(trace.goal_status, "unknown");
   assert.equal(trace.truncated, true);
   assert.ok(trace.residual_unknowns.length > 0);
+  const compared = await compareRetainedFunctionEvidence(client, evidence);
+  assert.equal(compared.normalized_result.status, "unknown");
+}
+
+/** Compare an already admitted function record without retransmitting its payload. */
+export async function compareRetainedFunctionEvidence(client, evidence) {
+  const reference = {
+    kind: "retained-evidence",
+    evidence_id: evidence.evidence_id,
+  };
   const compared = await client.callTool(
     {
       name: "compare_functions",
       arguments: {
-        left: analyzed.structuredContent,
-        right: analyzed.structuredContent,
+        left: reference,
+        right: reference,
       },
     },
     { timeout: 180000 },
   );
   assert.notEqual(compared.isError, true, JSON.stringify(compared));
   assert.notEqual(compared.structuredContent, undefined);
-  assert.equal(
-    parseEvidence(requireMcpOutput(compared, "compare_functions"))
-      .normalized_result.status,
-    "unknown",
-  );
+  return parseEvidence(requireMcpOutput(compared, "compare_functions"));
 }

scripts/lib/release-proposal.d.mts

@@ -1,5 +1,7 @@
 export function applyReleasePleasePatch(directory?: string): Promise<void>;
 
+export function describeReleaseProposalError(error: unknown): string;
+
 export function releaseProposalSearchDepth(
   run: (command: string, args: string[]) => Promise<string>,
   sourceSha: string,

scripts/lib/release-proposal.mjs

@@ -13,6 +13,15 @@ const patch = fileURLToPath(
   new URL("../../patches/release-please+17.3.0.patch", import.meta.url),
 );
 
+export function describeReleaseProposalError(error) {
+  if (!(error instanceof Error)) return String(error);
+  const description = error.message || error.name || "Unknown error";
+  const status = error.status;
+  return Number.isInteger(status) && status >= 400 && status <= 599
+    ? `${description} (HTTP ${status})`
+    : description;
+}
+
 export async function applyReleasePleasePatch(
   directory = dirname(require.resolve("release-please/package.json")),
 ) {

scripts/lib/test-discovery.d.mts

@@ -10,5 +10,10 @@ export function compareTestDiscovery(
     readonly file: string;
     readonly projects: readonly string[];
   }[];
+  readonly wrongProjects: readonly {
+    readonly file: string;
+    readonly expected: string;
+    readonly actual: string;
+  }[];
   readonly discovered: number;
 };

scripts/lib/test-discovery.mjs

@@ -1,4 +1,5 @@
 import { relative, resolve } from "node:path";
+import { projectForTest } from "./test-projects.mjs";
 
 /** Compare actual Vitest discovery with repository-owned test paths. */
 export const compareTestDiscovery = (expectedPaths, discovery, root) => {
@@ -36,6 +37,14 @@ export const compareTestDiscovery = (expectedPaths, discovery, root) => {
       const projects = projectsByFile.get(file);
       return projects.length > 1 ? [{ file, projects: sorted(projects) }] : [];
     }),
+    wrongProjects: sorted(expected).flatMap((file) => {
+      const projects = projectsByFile.get(file);
+      if (!projects || projects.length !== 1) return [];
+      const expectedProject = projectForTest(file).name;
+      return projects[0] === expectedProject
+        ? []
+        : [{ file, expected: expectedProject, actual: projects[0] }];
+    }),
     discovered: projectsByFile.size,
   };
 };

scripts/lint/index.mjs

@@ -1,7 +1,6 @@
 import noBufferDeepEquality from "./no-buffer-deep-equality.mjs";
 import cancellationCause from "./cancellation-cause.mjs";
 
-import noBulkAppendSpread from "./rules/no-bulk-append-spread.mjs";
 import noEvidenceJsonPreadmission from "./rules/no-evidence-json-preadmission.mjs";
 import noFixtureRootCleanup from "./rules/no-fixture-root-cleanup.mjs";
 import noPositionalToolContract from "./rules/no-positional-tool-contract.mjs";
@@ -12,7 +11,6 @@ export default {
   rules: {
     "no-buffer-deep-equality": noBufferDeepEquality,
     "cancellation-cause": cancellationCause,
-    "no-bulk-append-spread": noBulkAppendSpread,
     "no-evidence-json-preadmission": noEvidenceJsonPreadmission,
     "no-fixture-root-cleanup": noFixtureRootCleanup,
     "no-positional-tool-contract": noPositionalToolContract,

scripts/lint/rules/no-bulk-append-spread.d.mts

@@ -1,3 +0,0 @@
-import type { RuleTester } from "oxlint/plugins-dev";
-declare const rule: Parameters<RuleTester["run"]>[1];
-export default rule;

scripts/lint/rules/no-bulk-append-spread.mjs

@@ -1,57 +0,0 @@
-const bulkMethods = new Set(["push", "unshift", "splice"]);
-
-/** Syntax that leaves the evaluated callee or receiver unchanged. */
-const transparent = new Set([
-  "ChainExpression",
-  "ParenthesizedExpression",
-  "TSAsExpression",
-  "TSNonNullExpression",
-  "TSSatisfiesExpression",
-  "TSTypeAssertion",
-  "TSInstantiationExpression",
-]);
-
-const unwrap = (node) => {
-  while (node && transparent.has(node.type)) node = node.expression;
-  return node;
-};
-
-const propertyName = (member) => {
-  if (!member.computed && member.property.type === "Identifier")
-    return member.property.name;
-  if (member.computed && member.property.type === "Literal")
-    return member.property.value;
-  return undefined;
-};
-
-export default {
-  meta: {
-    type: "problem",
-    docs: {
-      description:
-        "Append collections element-wise instead of spreading them into call arguments.",
-    },
-    schema: [],
-    messages: {
-      spread:
-        "Spreading into {{method}}() makes each element a call argument, and engines bound argument count by stack size, so a data-sized collection throws RangeError. Use appendAll from src/domain/arrayAppend.ts, or document the structural bound in a disable comment.",
-    },
-  },
-  create(context) {
-    return {
-      CallExpression(node) {
-        const callee = unwrap(node.callee);
-        if (callee?.type !== "MemberExpression") return;
-        const method = propertyName(callee);
-        if (!bulkMethods.has(method)) return;
-        for (const argument of node.arguments)
-          if (argument.type === "SpreadElement")
-            context.report({
-              node: argument,
-              messageId: "spread",
-              data: { method },
-            });
-      },
-    };
-  },
-};

scripts/prepare-website.py

@@ -7,6 +7,7 @@ import xml.etree.ElementTree as ET
 from zipfile import ZIP_DEFLATED, ZipFile, ZipInfo
 
 from website import PageMetadata, page_url
+from website_agents import prepare_discovery, prepare_markdown
 
 
 EXAMPLE_FILES = (
@@ -81,6 +82,8 @@ def prepare_assets(site):
     prepare_sharing_image(site)
     prepare_sitemap(site)
     prepare_notes_example(site)
+    prepare_discovery(Path(__file__).resolve().parent.parent, site)
+    prepare_markdown(site)
 
 
 if __name__ == "__main__":

scripts/release-proposal.mjs

@@ -4,6 +4,7 @@ import { appendFile } from "node:fs/promises";
 import { parseArgs, promisify } from "node:util";
 import {
   applyReleasePleasePatch,
+  describeReleaseProposalError,
   releaseProposalSearchDepth,
 } from "./lib/release-proposal.mjs";
 
@@ -63,7 +64,7 @@ async function main() {
 
 await main().catch((error) => {
   process.stderr.write(
-    `Release proposal failed: ${error instanceof Error ? error.message : String(error)}\n`,
+    `Release proposal failed: ${describeReleaseProposalError(error)}\n`,
   );
   process.exitCode = 1;
 });

scripts/test-website-agents.py

@@ -0,0 +1,109 @@
+"""Check the public build artifacts and Markdown content contract."""
+
+import hashlib
+import json
+from pathlib import Path
+import tempfile
+import unittest
+from zipfile import ZipFile
+
+from bs4 import BeautifulSoup
+import yaml
+
+from website_agents import prepare_discovery, prepare_markdown
+
+
+ROOT = Path(__file__).resolve().parent.parent
+
+
+class AgentPublishingChecks(unittest.TestCase):
+    def test_published_skill_digest_covers_every_reference_and_catalogs_resolve(self):
+        site = ROOT / "website/_site"
+        index = json.loads((site / ".well-known/agent-skills/index.json").read_text())
+        self.assertEqual(index["$schema"], "https://schemas.agentskills.io/discovery/0.2.0/schema.json")
+        skill, = index["skills"]
+        self.assertEqual(skill["name"], "reverse-engineer-anything")
+        self.assertEqual(skill["type"], "archive")
+        artifact = site / skill["url"].removeprefix("https://rea.tools/")
+        self.assertEqual(skill["digest"], "sha256:" + hashlib.sha256(artifact.read_bytes()).hexdigest())
+        source = ROOT / ".agents/skills/reverse-engineer-anything"
+        with ZipFile(artifact) as archive:
+            expected = {f"{source.name}/{p.relative_to(source).as_posix()}": p.read_bytes()
+                        for p in source.rglob("*.md")}
+            self.assertEqual(set(archive.namelist()), set(expected))
+            for name, data in expected.items():
+                self.assertEqual(archive.read(name), data)
+            manifest = archive.read(f"{source.name}/SKILL.md").decode()
+            metadata = yaml.safe_load(manifest[4:].split("\n---\n", 1)[0])
+            self.assertEqual(skill["description"], metadata["description"])
+        catalog = json.loads((site / ".well-known/ai-catalog.json").read_text())
+        self.assertEqual(catalog["specVersion"], "1.0")
+        entry, = catalog["entries"]
+        self.assertEqual(entry["identifier"], "urn:air:rea.tools:skill:reverse-engineer-anything")
+        self.assertEqual(entry["url"], skill["url"])
+        self.assertNotIn("data", entry)
+        ard = json.loads((site / ".well-known/ard.json").read_text())
+        resource, = ard["entries"]
+        self.assertEqual(resource["url"], skill["url"])
+        self.assertTrue(2 <= len(resource["representativeQueries"]) <= 5)
+
+    def test_reference_changes_invalidate_digest_and_yaml_scalars_are_parsed(self):
+        with tempfile.TemporaryDirectory() as directory:
+            root = Path(directory)
+            source = root / ".agents/skills/reverse-engineer-anything"
+            source.mkdir(parents=True)
+            (source / "SKILL.md").write_text(
+                '---\nname: reverse-engineer-anything\ndescription: >-\n  Inspect programs:\n  preserve evidence.\n---\n# REA\n'
+            )
+            (source / "references").mkdir()
+            reference = source / "references/evidence.md"
+            reference.write_text("First revision")
+            site = root / "site"
+            def build():
+                prepare_discovery(root, site)
+                return json.loads((site / ".well-known/agent-skills/index.json").read_text())["skills"][0]
+            before = build()
+            self.assertEqual(before["description"], "Inspect programs: preserve evidence.")
+            self.assertEqual(build(), before)
+            reference.write_text("Second revision")
+            self.assertNotEqual(build()["digest"], before["digest"])
+
+    def test_markdown_preserves_content_and_resources_without_publishing_excluded_pages(self):
+        with tempfile.TemporaryDirectory() as directory:
+            site = Path(directory)
+            page = site / "zh-CN/guides/test/index.html"
+            page.parent.mkdir(parents=True)
+            page.write_text('''<html><head><title>REA: test</title></head><body>
+              <nav>Outside navigation</nav><main><aside>Shortened display paths.</aside>
+              <h1>Inspect a program</h1><nav>Contents navigation</nav>
+              <ol><li><span class="flow-label">Read</span><span class="flow-node"><code>input.exe</code></span><span class="flow-note">Keep evidence.</span></li></ol>
+              <p><a href="../other/?q=a%2Fb#code">Another guide</a></p>
+              <pre><code>const value = a &lt; b ? "x_y" : "&amp;";\n  keepIndent();</code></pre>
+              <details><summary>Evidence</summary><p>Hidden evidence remains available.</p></details>
+              <img src="../../../assets/diagram.svg" alt="Investigation diagram">
+              <button>Copy code</button><script>doNotPublish()</script></main></body></html>''')
+            fixture = site / "examples/notes-web/index.html"
+            fixture.parent.mkdir(parents=True)
+            fixture.write_text("<main>Target application</main>")
+            fallback = site / "zh-CN/fallback/index.html"
+            fallback.parent.mkdir(parents=True)
+            fallback.write_text('<head><meta name="robots" content="noindex, follow"></head><main>English fallback notice.</main>')
+            prepare_markdown(site)
+            markdown = page.with_suffix(".md").read_text()
+            self.assertIn('# Inspect a program', markdown)
+            self.assertIn('const value = a < b ? "x_y" : "&";\n  keepIndent();', markdown)
+            self.assertIn("Hidden evidence remains available.", markdown)
+            self.assertIn("Shortened display paths.", markdown)
+            self.assertNotIn("Read`input.exe`", markdown)
+            self.assertIn("https://rea.tools/zh-CN/guides/other/?q=a%2Fb#code", markdown)
+            self.assertIn("https://rea.tools/assets/diagram.svg", markdown)
+            for text in ("Outside navigation", "Contents navigation", "Copy code", "doNotPublish"):
+                self.assertNotIn(text, markdown)
+            self.assertFalse(fixture.with_suffix(".md").exists())
+            self.assertFalse(fallback.with_suffix(".md").exists())
+            html = BeautifulSoup(page.read_text(), "html.parser")
+            self.assertEqual(html.find("link", rel="alternate")["href"], "https://rea.tools/zh-CN/guides/test/index.md")
+
+
+if __name__ == "__main__":
+    unittest.main()

scripts/test-website.py

@@ -539,11 +539,13 @@ class WebsiteMaintenanceChecks(unittest.TestCase):
                             self.assertIn(b"color: black", response.read())
                         with self.assertRaises(HTTPError) as error:
                             urlopen(base + prefix + "missing")
-                        self.assertEqual(error.exception.code, 404)
+                        with contextlib.closing(error.exception) as response:
+                            self.assertEqual(response.code, 404)
                         if prefix != "/":
                             with self.assertRaises(HTTPError) as error:
                                 urlopen(base + "/guide/")
-                            self.assertEqual(error.exception.code, 404)
+                            with contextlib.closing(error.exception) as response:
+                                self.assertEqual(response.code, 404)
                     finally:
                         server.shutdown()
                         worker.join()

scripts/verify-adb.mjs

@@ -15,8 +15,10 @@ if (options.help) {
   process.stdout.write(
     "Usage: npm run verify:adb [-- --serial SERIAL --pull PACKAGE]\n" +
       "Requires a caller-selected adb binary (REA_ADB_PATH or PATH) and a\n" +
-      "connected device or emulator. Verifies all five operations against\n" +
-      "the live device; --pull acquires a real package and re-digests the\n" +
+      "connected device or emulator. Verifies observation and transfer operations\n" +
+      "against the live device. --install-apk is unavailable until APK identity\n" +
+      "and device cleanup ownership can be established safely.\n" +
+      "--pull acquires a real package and re-digests the\n" +
       "files on disk against the returned SHA-256 values.\n",
   );
   process.exit(0);
@@ -356,79 +358,6 @@ try {
             `versionCode=${String(details.version_code)}`,
           );
       }
-      if (options.installApk !== null && options.installApk !== undefined) {
-        const probePackage = "com.rea.adb.probe";
-        const installedProbe = await run("install_adb_package", {
-          serial,
-          apk_path: options.installApk,
-          replace: false,
-        });
-        if (installedProbe !== null) {
-          check(
-            "install reports the APK digest",
-            typeof installedProbe.sha256 === "string" &&
-              installedProbe.sha256.length === 64,
-            `${installedProbe.bytes} bytes`,
-          );
-          const resolved = await run("resolve_adb_packages", {
-            serial,
-            query: "rea.adb.probe",
-          });
-          if (resolved !== null)
-            check(
-              "resolve finds the installed probe uniquely",
-              resolved.matches.length === 1 &&
-                resolved.matches[0]?.package_name === probePackage,
-              `${resolved.matches.length} match(es)`,
-            );
-          const started = await run("start_adb_app", {
-            serial,
-            package: probePackage,
-          });
-          if (started !== null)
-            check(
-              "app start resolves and launches the launcher activity",
-              typeof started.component === "string" &&
-                started.component.startsWith(probePackage),
-              String(started.component),
-            );
-          const observedProcesses = await run("list_adb_processes", { serial });
-          if (observedProcesses !== null)
-            check(
-              "started app appears in the process listing",
-              observedProcesses.processes.some(
-                (process) => process.name === probePackage,
-              ),
-            );
-          const stopped = await run("stop_adb_app", {
-            serial,
-            package: probePackage,
-          });
-          if (stopped !== null)
-            check(
-              "force-stop completes the dynamic-analysis run",
-              stopped.package_name === probePackage,
-            );
-          const removed = await run("uninstall_adb_package", {
-            serial,
-            package: probePackage,
-          });
-          if (removed !== null)
-            check(
-              "uninstall completes the lifecycle",
-              removed.package_name === probePackage,
-            );
-          const afterRemoval = await run("resolve_adb_packages", {
-            serial,
-            query: "rea.adb.probe",
-          });
-          if (afterRemoval !== null)
-            check(
-              "resolve reports zero matches after removal",
-              afterRemoval.matches.length === 0,
-            );
-        }
-      }
     } finally {
       await rm(transferRoot, { recursive: true, force: true });
     }
@@ -459,7 +388,7 @@ if (failures > 0) {
 }
 
 function parseArguments(argv) {
-  const parsed = { help: false, serial: null, pull: null, installApk: null };
+  const parsed = { help: false, serial: null, pull: null };
   for (let index = 0; index < argv.length; index += 1) {
     const argument = argv[index];
     if (argument === "--help" || argument === "-h") parsed.help = true;
@@ -470,8 +399,10 @@ function parseArguments(argv) {
       parsed.pull = argv[index + 1] ?? null;
       index += 1;
     } else if (argument === "--install-apk") {
-      parsed.installApk = argv[index + 1] ?? null;
-      index += 1;
+      process.stderr.write(
+        "--install-apk is unavailable: APK identity and device cleanup ownership cannot be established safely\n",
+      );
+      process.exit(1);
     } else {
       process.stderr.write(`unknown argument: ${argument}\n`);
       process.exit(2);

scripts/verify-binary-layout.mjs

@@ -38,15 +38,16 @@ if (
   !isAbsolute(strace ?? "")
 )
   throw new Error(
-    "verify:binary:layout requires Linux x64, absolute REA_PWNTOOLS_PYTHON (pwntools 4.15.0/pyelftools 0.33/Unicorn 2.1.2) and REA_VERIFY_STRACE_COMMAND. gcc, ld, strip and the i686/ARM/MIPS/MIPSel linux-gnu cross gcc and readelf must be available. No target is executed.",
+    "verify:binary:layout requires Linux x64, absolute REA_PWNTOOLS_PYTHON (pwntools 4.15.0/pyelftools 0.33/Unicorn 2.1.2) and REA_VERIFY_STRACE_COMMAND. gcc, ld, strip and the AArch64/i686/ARM/MIPS/MIPSel linux-gnu cross gcc and readelf must be available. No target is executed.",
   );
 const run = createVerifierRun();
 const execute = promisify(execFile);
-const ELF32_TOOLCHAINS = [
-  ["i686-linux-gnu", "EM_386", "little"],
-  ["arm-linux-gnueabi", "EM_ARM", "little"],
-  ["mips-linux-gnu", "EM_MIPS", "big"],
-  ["mipsel-linux-gnu", "EM_MIPS", "little"],
+const CROSS_TOOLCHAINS = [
+  ["aarch64-linux-gnu", "EM_AARCH64", 64, "little"],
+  ["i686-linux-gnu", "EM_386", 32, "little"],
+  ["arm-linux-gnueabi", "EM_ARM", 32, "little"],
+  ["mips-linux-gnu", "EM_MIPS", 32, "big"],
+  ["mipsel-linux-gnu", "EM_MIPS", 32, "little"],
 ];
 for (const command of [
   "gcc",
@@ -54,7 +55,7 @@ for (const command of [
   "strip",
   "/usr/bin/prlimit",
   strace,
-  ...ELF32_TOOLCHAINS.flatMap(([prefix]) => [
+  ...CROSS_TOOLCHAINS.flatMap(([prefix]) => [
     `${prefix}-gcc`,
     `${prefix}-readelf`,
   ]),
@@ -347,14 +348,15 @@ try {
       cases++;
     }
   }
-  // ELF32 profiles: cross-compiled, never executed, checked against GNU readelf.
+  // Cross profiles: cross-compiled, never executed, checked against GNU readelf.
   for (const [
     prefix,
     machine,
+    bits,
     byteOrder,
     kind,
     flags,
-  ] of ELF32_TOOLCHAINS.flatMap((toolchain) => [
+  ] of CROSS_TOOLCHAINS.flatMap((toolchain) => [
     [...toolchain, "executable", ["-fPIE", "-pie"]],
     [...toolchain, "relocatable", ["-c", "-fPIC"]],
   ])) {
@@ -423,16 +425,18 @@ try {
         `readelf relocation count: ${name}`,
       );
       for (const [entryIndex, [, offset, info]] of rows.entries()) {
-        // ELF32 r_info uses the low byte for type and the upper 24 bits for
-        // the symbol index. readelf has already decoded the file's byte order.
+        // ELF32 r_info packs an 8-bit type under a 24-bit symbol index;
+        // ELF64 packs a 32-bit type under a 32-bit index. readelf has already
+        // decoded the file's byte order.
         const decodedInfo = BigInt(`0x${info}`);
+        const typeBits = bits === 32 ? 8n : 32n;
         expectedRelocations.push({
           section_index: section.index,
           entry_index: entryIndex,
           reported_offset: BigInt(`0x${offset}`),
-          type: Number(decodedInfo & 0xffn),
+          type: Number(decodedInfo & ((1n << typeBits) - 1n)),
           symbol_table_index: section.link,
-          symbol_index: Number(decodedInfo >> 8n),
+          symbol_index: Number(decodedInfo >> typeBits),
         });
       }
     }
@@ -445,7 +449,7 @@ try {
       const value = await inspect(mode, path);
       assert.deepEqual(value.architecture, {
         machine,
-        bits: 32,
+        bits,
         byte_order: byteOrder,
       });
       assert.equal(
@@ -1312,7 +1316,8 @@ try {
       }
     }
   }
-  unsupported.writeUInt16LE(183, 18);
+  // EM_RISCV: AArch64 (183) is now an admitted profile.
+  unsupported.writeUInt16LE(243, 18);
   const core = Buffer.from(object);
   core.writeUInt16LE(4, 16);
   for (const [name, bytes, category] of [
@@ -1337,7 +1342,7 @@ try {
       "invalid_input",
     ],
     ["undersized-program-header", undersizedProgramHeader, "invalid_input"],
-    ["arm64", unsupported, "unsupported_target"],
+    ["riscv64", unsupported, "unsupported_target"],
     ["core", core, "unsupported_target"],
   ]) {
     const path = join(root.path, name);

scripts/verify-bun-standalone.mjs

@@ -8,7 +8,7 @@ import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 
 const root = await PrivateRuntimeRoot.create({ prefix: "rea-bun-verifier-" });
-const commands = await createVerifierCommands();
+let commands;
 const run = (command, args, options = {}) =>
   commands.exec(command, args, {
     timeout: 120_000,
@@ -19,6 +19,7 @@ const run = (command, args, options = {}) =>
 // only after every owned command family has settled.
 let commandsClosed = false;
 await withVerifierCleanup(async () => {
+  commands = await createVerifierCommands();
   const source = join(root.path, "hello.ts");
   await writeFile(source, 'console.log("REA_BUN_VERIFICATION");\n');
   const docker = process.argv.includes("--docker");
@@ -104,7 +105,7 @@ await withVerifierCleanup(async () => {
   );
 }, [
   async () => {
-    await commands.close();
+    if (commands !== undefined) await commands.close();
     commandsClosed = true;
   },
   async () => {

scripts/verify-claude-client.mjs

@@ -2,7 +2,7 @@ import assert from "node:assert/strict";
 import { createHash } from "node:crypto";
 import { mkdir, mkdtemp, readFile, writeFile } from "node:fs/promises";
 import { tmpdir } from "node:os";
-import { join } from "node:path";
+import { isAbsolute, join } from "node:path";
 import { fileURLToPath } from "node:url";
 import Ajv2020 from "ajv/dist/2020.js";
 import { createCatalogIdentity } from "../dist/catalogIdentity.js";
@@ -15,6 +15,7 @@ import {
   TOOL_CONTRACTS,
   toolContract,
 } from "../dist/contracts/toolContracts.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
@@ -306,7 +307,9 @@ await withVerifierCleanup(async () => {
           assert(
             parsed.plannedActions.every(
               (a) =>
-                a.target === configPath || a.target.startsWith(skillDirectory),
+                a.target === configPath ||
+                (isAbsolute(a.target) &&
+                  isPathWithinRoot(skillDirectory, a.target)),
             ),
           );
         }

scripts/verify-commandcode-client.mjs

@@ -6,6 +6,7 @@ import { isAbsolute, join, relative } from "node:path";
 import { fileURLToPath } from "node:url";
 import Ajv2020 from "ajv/dist/2020.js";
 import { TOOL_CONTRACTS } from "../dist/contracts/toolContracts.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
@@ -54,6 +55,7 @@ const expectedNames = TOOL_CONTRACTS.map(
 ).sort();
 let clientVersion = null;
 let expectedSkill;
+let skillCallId;
 let inventoryChecked = false;
 let catalogSize = null;
 let skillLoaded = false;
@@ -180,10 +182,24 @@ await withVerifierCleanup(async () => {
           finalSeen = true;
           return { role: "assistant", content: "REA_CLIENT_COMPATIBILITY_OK" };
         }
-        if (stage === 0)
-          return call("activate_skill", { name: "reverse-engineer-anything" });
+        if (stage === 0) {
+          const request = call("activate_skill", {
+            name: "reverse-engineer-anything",
+          });
+          skillCallId = request.tool_calls[0].id;
+          return request;
+        }
         assert(
-          strings.some((text) => text.includes(expectedSkill)),
+          skillCallId !== undefined &&
+            expectedSkill !== undefined &&
+            body.messages.some(
+              (message) =>
+                message.role === "tool" &&
+                message.tool_call_id === skillCallId &&
+                textValues(message.content).some((text) =>
+                  text.includes(expectedSkill),
+                ),
+            ),
           "The model must receive the complete activated skill body",
         );
         skillLoaded = true;
@@ -322,8 +338,8 @@ await withVerifierCleanup(async () => {
           result.plannedActions.every(
             (action) =>
               action.target === configPath ||
-              action.target === skillDirectory ||
-              action.target.startsWith(skillDirectory + "/"),
+              (isAbsolute(action.target) &&
+                isPathWithinRoot(skillDirectory, action.target)),
           ),
           "Refuse setup writes outside the disposable account's configuration and skill",
         );
@@ -342,9 +358,16 @@ await withVerifierCleanup(async () => {
     assert.deepEqual(repeated.appliedActions, []);
     assert.equal(await readFile(configPath, "utf8"), configured);
     assert.equal(await readFile(`${configPath}.rea.backup`, "utf8"), original);
-    expectedSkill = (await readFile(join(skillDirectory, "SKILL.md"), "utf8"))
-      .replace(/^---\r?\n[\s\S]*?\r?\n---\r?\n/, "")
-      .trim();
+    const installedSkill = await readFile(
+      join(skillDirectory, "SKILL.md"),
+      "utf8",
+    );
+    const frontmatter = /^---\r?\n[\s\S]*?\r?\n---(?:\r?\n|$)/u.exec(
+      installedSkill,
+    );
+    assert(frontmatter, "Installed skill must have authored frontmatter");
+    expectedSkill = installedSkill.slice(frontmatter[0].length).trim();
+    assert(expectedSkill.length > 0, "Installed skill body must not be empty");
     const response = await saveVerifierCommandOutput(
       () =>
         commands.exec(

scripts/verify-copilot-client.mjs

@@ -12,6 +12,7 @@ import {
 import { advertisedInputJsonSchema } from "../dist/contracts/toolSchemaMetadata.js";
 import { presentAdvertisedInputSchema } from "../dist/contracts/compactInputSchema.js";
 import { compactPresentationForMcpInputSchemaProfile } from "../dist/config/mcpInputSchemaProfile.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
@@ -300,8 +301,8 @@ await withVerifierCleanup(async () => {
             parsed.plannedActions.every(
               (a) =>
                 a.target === configPath ||
-                a.target === skillDirectory ||
-                a.target.startsWith(skillDirectory + "/"),
+                (isAbsolute(a.target) &&
+                  isPathWithinRoot(skillDirectory, a.target)),
             ),
           );
         }

scripts/verify-deepseek-client.mjs

@@ -56,6 +56,7 @@ let catalog;
 let patternCount = 0;
 let skillRequested = false;
 let skillLoaded = false;
+let expectedSkillBody;
 let evidenceSeen = false;
 let finalSeen = false;
 let artifactDigest;
@@ -169,11 +170,34 @@ await withVerifierCleanup(async () => {
         const toolMessages = (body.messages ?? []).filter(
           (message) => message.role === "tool",
         );
-        const serialized = JSON.stringify(toolMessages);
-        skillLoaded ||=
-          serialized.includes(
-            '<skill_content name=\\"reverse-engineer-anything\\">',
-          ) && serialized.includes(dirname(skillPath));
+        skillLoaded ||= toolMessages.some((message) => {
+          if (
+            typeof message.content !== "string" ||
+            expectedSkillBody === undefined
+          )
+            return false;
+          const open =
+            '<skill_content name="reverse-engineer-anything">\n<skill_resources>\n';
+          const start = message.content.indexOf(open);
+          if (start < 0) return false;
+          const resourcesStart = start + open.length;
+          const instructions = `\n</skill_resources>\n\n<skill_instructions>\n${expectedSkillBody}\n</skill_instructions>\n</skill_content>`;
+          const resourcesEnd = message.content.indexOf(
+            instructions,
+            resourcesStart,
+          );
+          if (resourcesEnd < 0) return false;
+          const expectedPath = dirname(skillPath)
+            .replaceAll("&", "&amp;")
+            .replaceAll("<", "&lt;")
+            .replaceAll(">", "&gt;");
+          const resources = message.content.slice(resourcesStart, resourcesEnd);
+          return (
+            !resources.includes("</skill_content>") &&
+            !resources.includes("</skill_resources>") &&
+            resources.includes(expectedPath)
+          );
+        });
         for (const message of toolMessages) {
           const found = findClientResult(
             message.content,
@@ -281,11 +305,15 @@ await withVerifierCleanup(async () => {
     const configured = await readFile(configPath, "utf8");
     assert.deepEqual(withoutRea(configured), withoutRea(original));
     assert(configured.includes("!!js process.cwd()"));
-    assert(
-      (await readFile(skillPath, "utf8")).includes(
-        "name: reverse-engineer-anything",
-      ),
+    const installedSkill = await readFile(skillPath, "utf8");
+    assert(installedSkill.includes("name: reverse-engineer-anything"));
+    const frontmatter = /^---\r?\n[\s\S]*?^---(?:\r?\n|$)/mu.exec(
+      installedSkill,
     );
+    assert(frontmatter, "Installed skill must have its authored frontmatter");
+    expectedSkillBody = installedSkill.slice(frontmatter[0].length).trim();
+    assert(expectedSkillBody.length > 0);
+
     lifecycle.push("setup");
     const repeated = await runRea(
       ["setup", "--client", "dsh", "--yes"],

scripts/verify-flutter.mjs

@@ -104,6 +104,10 @@ await withVerifierCleanup(
     }
     // The same checks apply to the default selection and each remaining ABI.
     const preferred = await run("inspect_dart_aot", { path: apk });
+    assert.ok(
+      abis.some((abi) => abi.abi === preferred.abi),
+      "Default AOT selection must be one of the identified ABIs",
+    );
     for (const abi of abis) {
       const aot =
         abi.abi === preferred.abi

scripts/verify-frida.mjs

@@ -2,7 +2,7 @@ import { basename } from "node:path";
 
 import { FridaInstrumentationService } from "../dist/application/frida/FridaInstrumentationService.js";
 import { FridaInstrumentationManager } from "../dist/frida/FridaInstrumentationManager.js";
-import { isValidFridaRemoteAddress } from "../dist/contracts/fridaRemoteAddress.js";
+import { isValidFridaRemoteAddress } from "../dist/domain/fridaRemoteAddress.js";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 
 const options = parseArguments(process.argv.slice(2));
@@ -80,6 +80,10 @@ try {
       !Array.isArray(observed.messages)
     )
       throw new Error("Frida returned no message observation");
+    if (!Number.isSafeInteger(observed.pid) || observed.pid <= 0)
+      throw new Error("Frida returned no valid target PID");
+    if (remote !== undefined && observed.pid !== options.pid)
+      throw new Error("Frida observed a different target PID");
     const messageObserved = observed.messages.some(
       (message) =>
         typeof message === "object" &&

scripts/verify-gemini-client.mjs

@@ -2,13 +2,14 @@ import assert from "node:assert/strict";
 import { createHash } from "node:crypto";
 import { mkdir, mkdtemp, readFile, writeFile } from "node:fs/promises";
 import { tmpdir } from "node:os";
-import { join, resolve, sep } from "node:path";
+import { isAbsolute, join, relative, resolve } from "node:path";
 import { fileURLToPath } from "node:url";
 import Ajv2020 from "ajv/dist/2020.js";
 import {
   TOOL_CONTRACTS,
   toolContract,
 } from "../dist/contracts/toolContracts.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
@@ -65,6 +66,7 @@ let fixture;
 let passed = false;
 let evidenceSeen = false;
 let skillLoaded = false;
+let expectedSkillBody;
 let skillRequested = false;
 let finalSeen = false;
 let catalogSize = null;
@@ -150,13 +152,21 @@ await withVerifierCleanup(async () => {
         if (args.includes("--dry-run")) {
           assert(parsed.plannedActions.some((a) => a.target === configPath));
           assert(
-            parsed.plannedActions.some((a) =>
-              a.target.startsWith(join(profile, ".agents", "skills") + sep),
-            ),
+            parsed.plannedActions.some((a) => {
+              const skills = join(profile, ".agents", "skills");
+              return (
+                isAbsolute(a.target) &&
+                relative(skills, a.target) !== "" &&
+                isPathWithinRoot(skills, a.target)
+              );
+            }),
           );
           assert(
-            parsed.plannedActions.every((a) =>
-              a.target.startsWith(profile + sep),
+            parsed.plannedActions.every(
+              (a) =>
+                isAbsolute(a.target) &&
+                relative(profile, a.target) !== "" &&
+                isPathWithinRoot(profile, a.target),
             ),
           );
         }
@@ -189,11 +199,15 @@ await withVerifierCleanup(async () => {
     assert.deepEqual(JSON.parse(repeated.stdout).appliedActions, []);
     assert.equal(await readFile(configPath, "utf8"), configured);
     assert.equal(await readFile(`${configPath}.rea.backup`, "utf8"), original);
-    assert(
-      (await readFile(skillPath, "utf8")).includes(
-        "name: reverse-engineer-anything",
-      ),
+    const installedSkill = await readFile(skillPath, "utf8");
+    assert(installedSkill.includes("name: reverse-engineer-anything"));
+    const frontmatter = /^---\r?\n([\s\S]*?)\r?\n---(?:\r?\n([\s\S]*))?/u.exec(
+      installedSkill,
     );
+    assert(frontmatter, "Installed skill must have authored frontmatter");
+    expectedSkillBody = (frontmatter[2] ?? "").trim();
+    assert(expectedSkillBody.length > 0);
+
     const target = join(lab, "fixture α测试");
     await mkdir(target);
     await writeFile(join(target, "app.js"), CLIENT_JAVASCRIPT_FIXTURE_SOURCE);
@@ -268,16 +282,32 @@ await withVerifierCleanup(async () => {
           .flatMap((c) => c.parts ?? [])
           .filter((p) => p.functionResponse)
           .map((p) => p.functionResponse);
-        skillLoaded ||= responses.some(
-          (r) =>
-            r.name === "activate_skill" &&
-            r.response?.output?.includes(
-              `<activated_skill name="reverse-engineer-anything">`,
-            ) &&
-            r.response.output.includes(
+        skillLoaded ||= responses.some((response) => {
+          const output = response.response?.output;
+          if (
+            response.name !== "activate_skill" ||
+            typeof output !== "string" ||
+            expectedSkillBody === undefined
+          )
+            return false;
+          const start = `<activated_skill name="reverse-engineer-anything">\n  <instructions>\n    ${expectedSkillBody}\n  </instructions>\n\n  <available_resources>\n    `;
+          const at = output.indexOf(start);
+          if (at < 0) return false;
+          const resourcesStart = at + start.length;
+          const resourcesEnd = output.indexOf(
+            "\n  </available_resources>\n</activated_skill>",
+            resourcesStart,
+          );
+          if (resourcesEnd < 0) return false;
+          const resources = output.slice(resourcesStart, resourcesEnd);
+          return (
+            !resources.includes("</activated_skill>") &&
+            !resources.includes("</available_resources>") &&
+            resources.includes(
               join(profile, ".agents", "skills", "reverse-engineer-anything"),
-            ),
-        );
+            )
+          );
+        });
         const observed = findClientResult(
           responses,
           isClientJavaScriptEvidence,

scripts/verify-grok-client.mjs

@@ -6,6 +6,7 @@ import { isAbsolute, join, relative } from "node:path";
 import { fileURLToPath } from "node:url";
 import Ajv2020 from "ajv/dist/2020.js";
 import { TOOL_CONTRACTS } from "../dist/contracts/toolContracts.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
@@ -402,8 +403,8 @@ enabled = false
             parsed.plannedActions.every(
               (action) =>
                 action.target === configPath ||
-                action.target === skillDirectory ||
-                action.target.startsWith(skillDirectory + "/"),
+                (isAbsolute(action.target) &&
+                  isPathWithinRoot(skillDirectory, action.target)),
             ),
             "Refuse writes outside the owned profile/skill",
           );

scripts/verify-http3-network-captures.mjs

@@ -373,6 +373,7 @@ try {
 }
 const report = {
   run_id: run.run_id,
+  input_mode: inspectExisting ? "existing-captures" : "producer-attempt",
   status: primaryError || cleanupFailures.length ? "failed" : "passed",
   finished_at: new Date().toISOString(),
   ...(primaryError
@@ -386,8 +387,13 @@ const report = {
   verifier,
   producer_audit: producerAudit,
   scope:
-    "Real Linux reverse HTTP/3, mitmproxy 12.2.3, aioquic 1.2.0 and guest cURL",
+    "Linux reverse HTTP/3 capture verification for mitmproxy 12.2.3, aioquic 1.2.0 and guest cURL",
   limitations: [
+    ...(inspectExisting
+      ? [
+          "Existing capture producer provenance was not authenticated by this attempt.",
+        ]
+      : []),
     "No local/WireGuard, browser, QUIC v2, migration, 0-RTT, mTLS or replay claim.",
     "No HAR, interruption, certificate-failure or body-completeness coverage in this profile.",
   ],

scripts/verify-il2cpp-metadata.mjs

@@ -17,7 +17,7 @@ import { hashVerifierFile } from "./lib/verifier-file-resources.mjs";
 const args = process.argv.slice(2);
 if (args.includes("--help") || args.includes("-h")) {
   console.log(
-    "Usage: npm run verify:il2cpp:metadata -- [--metadata PATH]\nPATH is a caller-owned, unencrypted global-metadata.dat (version 24, 27, 29, 31 or 39). Without --metadata, only the advertised contract is verified.",
+    "Usage: npm run verify:il2cpp:metadata -- [--metadata PATH]\nPATH is a caller-owned, unencrypted global-metadata.dat (version 24, 27, 29, 31 or 39). Without --metadata, only the MCP tool name is checked.",
   );
   process.exit(0);
 }
@@ -178,4 +178,8 @@ await withVerifierCleanup(async () => {
     );
   }
 }, [() => client.close(), () => transport.close(), () => resources.close()]);
-console.log("All asserted IL2CPP metadata CLI/MCP checks passed");
+console.log(
+  path === null
+    ? "IL2CPP metadata MCP tool presence check passed"
+    : "All asserted IL2CPP metadata CLI/MCP checks passed",
+);

scripts/verify-jeb.mjs

@@ -68,7 +68,9 @@ try {
     });
     if (!decompiled.ok) throw new Error(decompiled.error.message);
     const text = decompiled.value.normalized_result.text;
-    if (!text.includes(target.marker))
+    if (text.trim().length === 0)
+      throw new Error("Decompiled pseudo-code was empty");
+    if (target.marker !== undefined && !text.includes(target.marker))
       throw new Error(
         "Decompiled pseudo-code did not contain the fixture marker",
       );
@@ -83,7 +85,7 @@ try {
       open_project_observed: openRefusal === null,
       ...(openRefusal === null ? {} : { open_project_refusal: openRefusal }),
       unit_count: units.value.normalized_result.units.length,
-      decompiled_marker_observed: true,
+      decompiled_marker_observed: target.marker !== undefined,
     };
   }, [
     async () => {
@@ -110,7 +112,7 @@ async function resolveTarget(selected, directory) {
     return {
       path: "/engine-side/target",
       itemAddress: selected.itemAddress,
-      marker: selected.marker ?? "",
+      marker: selected.marker,
     };
   await writeFile(
     join(directory, "Hello.java"),
@@ -153,6 +155,8 @@ function parseArguments(args) {
     values.set(flag, value);
     index += 1;
   }
+  if (values.get("--marker") === "")
+    throw new Error("--marker must not be empty");
   return {
     help: false,
     ...(values.has("--item-address")

scripts/verify-native-ui.mjs

@@ -245,7 +245,13 @@ try {
     scope,
     { environment: process.env },
   );
-  if (mismatch.ok || !mismatch.error.message.includes("target-mismatch"))
+  // A selection failure carries its helper code in the issue, not in its message.
+  if (
+    mismatch.ok ||
+    !mismatch.error.issues?.some(({ message }) =>
+      message.startsWith("target-mismatch:"),
+    )
+  )
     throw new Error("Real helper did not reject changed target bytes");
   report = {
     ok: true,

scripts/verify-omp-client.mjs

@@ -2,9 +2,10 @@ import assert from "node:assert/strict";
 import { createHash } from "node:crypto";
 import { mkdir, mkdtemp, readFile, writeFile } from "node:fs/promises";
 import { homedir, tmpdir } from "node:os";
-import { join, relative } from "node:path";
+import { isAbsolute, join, relative } from "node:path";
 import { fileURLToPath } from "node:url";
 import { TOOL_CONTRACTS } from "../dist/contracts/toolContracts.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
@@ -314,8 +315,8 @@ await withVerifierCleanup(async () => {
             parsed.plannedActions.every(
               (action) =>
                 action.target === configPath ||
-                action.target === skillDirectory ||
-                action.target.startsWith(skillDirectory + "/"),
+                (isAbsolute(action.target) &&
+                  isPathWithinRoot(skillDirectory, action.target)),
             ),
             "Refuse setup writes outside the owned config/skill",
           );

scripts/verify-opencode-client.mjs

@@ -2,10 +2,11 @@ import assert from "node:assert/strict";
 import { createHash } from "node:crypto";
 import { mkdir, mkdtemp, readFile, writeFile } from "node:fs/promises";
 import { tmpdir } from "node:os";
-import { isAbsolute, join, relative, sep } from "node:path";
+import { isAbsolute, join } from "node:path";
 import { fileURLToPath } from "node:url";
 import Ajv2020 from "ajv/dist/2020.js";
 import { TOOL_CONTRACTS } from "../dist/contracts/toolContracts.js";
+import { isPathWithinRoot } from "../dist/domain/localPath.js";
 import { createVerifierCommands } from "./lib/verify-commands.mjs";
 import { saveVerifierCommandOutput } from "./lib/verifier-command-output.mjs";
 import { withVerifierCleanup } from "./lib/verifier-cleanup.mjs";
@@ -49,19 +50,15 @@ const skillDirectory = join(
   "skills",
   "reverse-engineer-anything",
 );
-const isManagedSkillTarget = (target) => {
-  const value = relative(skillDirectory, target);
-  return (
-    value === "" ||
-    (value !== ".." && !value.startsWith(`..${sep}`) && !isAbsolute(value))
-  );
-};
 assert.equal(
-  isManagedSkillTarget(`${skillDirectory}-sibling/AGENTS.md`),
+  isPathWithinRoot(skillDirectory, `${skillDirectory}-sibling/AGENTS.md`),
   false,
 );
 assert.equal(
-  isManagedSkillTarget(join(skillDirectory, "..", "outside", "AGENTS.md")),
+  isPathWithinRoot(
+    skillDirectory,
+    join(skillDirectory, "..", "outside", "AGENTS.md"),
+  ),
   false,
 );
 let clientVersion = null;
@@ -271,7 +268,10 @@ await withVerifierCleanup(async () => {
           );
           assert(
             parsed.plannedActions.every(
-              (a) => a.target === configPath || isManagedSkillTarget(a.target),
+              (a) =>
+                a.target === configPath ||
+                (isAbsolute(a.target) &&
+                  isPathWithinRoot(skillDirectory, a.target)),
             ),
           );
         }