* feat(native): inspect recorded crashes through pwntools and pwndbg
* fix(native): bind core note semantics to raw owner bytes
* fix(native): validate recorded core ABI and evidence scalars
* fix(native): require complete recorded note interpretations
* fix(native): bind physical core notes to original headers
Bring in #927's merge of main (#1037 at skill version 30):
- README_ja/ar/ko/zh: keep the merged tables and count
inspect_dyld_shared_cache in application workflows (13 -> 14).
- Skill: #927 now uses version 31, so this branch moves to 32.
- Regenerate the product catalog, conformance manifest, skill digest and
package metadata (138 tools).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Bring in #927's merge of main and resolve conflicts:
- README.md: main's simplified English README has no tool-family table.
- README_ja/ar/ko/zh: keep main's reworded tables and count
inspect_dyld_shared_cache in application workflows (13 -> 14).
- Skill: #927 now uses version 29, so this branch moves to 30.
- Regenerate the product catalog, conformance manifest, skill digest and
package metadata (136 tools). The MCP tool catalog is now built rather
than committed (#997), so it is removed here.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
fix(release): support built-in credentials and refresh release boundary
GitHub now creates approval-required PR workflows for GITHUB_TOKEN-authored
opened/synchronize/reopened events. Restore optional custom credentials and
cover prepare/publish without a separate secret. Record the verified 5.0.0
checkpoint, public 133-tool catalog and package update in the installation guide.
Assisted-by: Sakuya (gensokyo-skills:sakuya-checkpointed-execution)
* fix(release): publish from explicit source checkpoints
Dispatch preparation and publication against a frozen release branch.
Bind package builds and provenance to the same tagged commit, enable
release-branch PR CI, and document the maintainer workflow.
Assisted-by: Sakuya (gensokyo-skills:sakuya-checkpointed-execution)
* fix(release): stop publication when the checkpoint has moved
Reject unknown phases and a release-branch tip that differs from the
dispatch SHA before Release Please creates a tag. Keep registry jobs and
catalog pushes on their own phase, and run website checks for release PRs.
* fix(release): require a token that can trigger release CI
Stop falling back to GITHUB_TOKEN, which cannot start workflow runs for the release pull request it creates.
* test(release): require the CI-capable release token
Lock the workflow to RELEASE_PLEASE_TOKEN and fail preparation when that secret is missing.
* style(release): format the release-token test
---------
Co-authored-by: N0zoM1z0 <1705004398@qq.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
* fix(release): publish from explicit source checkpoints
Dispatch preparation and publication against a frozen release branch.
Bind package builds and provenance to the same tagged commit, enable
release-branch PR CI, and document the maintainer workflow.
Assisted-by: Sakuya (gensokyo-skills:sakuya-checkpointed-execution)
* fix(release): drive checkpoint releases from frozen branches
Use branch pushes for Release Please preparation and reviewed-merge publication.
Reject moved source refs before tagging and bind registry provenance to the
triggering release commit. Remove phase inputs and document the three steps.
Assisted-by: Sakuya (gensokyo-skills:sakuya-checkpointed-execution)
Assisted-by: Cirno (gensokyo-skills:cirno-radical-simplification)
* fix(release): extend npm propagation window and improve diagnostics
The release verification step polls npm for up to 20 minutes (80 attempts
× 15s), but npm registry propagation can take longer. When the version
doesn't appear within the window, the workflow fails even though the
package was already published successfully.
Changes:
- Extend polling window from 20 to 30 minutes (120 attempts × 15s)
- Add --force flag to npm view to bypass cache on each poll
- Report full retry window in diagnostics (attempts × sleep time)
- Distinguish propagation delay from CLI verification failure in error message
Fixes#621
* fix(release): drop no-op --force from npm propagation polling
npm view --force is accepted but only prints a protections warning on a read-only command; --prefer-online already forces fresh registry metadata on each poll.
---------
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
src/generatedMcpToolCatalog.ts had zero runtime importers and churned on every contract edit, guaranteeing merge conflicts on long-lived branches. It is now gitignored and regenerated by build:unlocked; typecheck and knip depend on build via turbo. Tests consume the build output unchanged. Adds a regen-generated CI workflow that pushes tracked generated metadata (product catalog, skill frontmatter, verification manifest/ledger, package metadata, error schema) to same-repo PR branches.
Carry #952's Apple ownership into the cache work: the cache reader and
fixture move to src/artifacts/apple/, the cache facts, platforms and
resolution schema to src/domain/apple/, the cache service to
src/application/apple/, and the lane to scripts/verify/apple/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Electron guide currently uses a teaching fixture. Add a real Notion clipboard case with REA findings, short source excerpts and responsive figures, alongside the existing DX-Ball reconstruction.
- Trace the saved Notion 7.6.1 preload API through its wrappers to the clipboard receiver. Attribute the separate web-cache module probes to their original harness. Public content uses relative source locations and generic example data; packages and complete captures stay local.
- Add a Showcases index and links from the homepage and JavaScript guide. Simplify homepage copy while preserving the approved headline and closing question.
- Show copyable agent installation prompts and the direct setup command. Keep terminal cursors blinking consistently and respect reduced-motion settings.
- Provide one Notes example ZIP. Generate it from the six public source files before website checks and manual Pages deployment; verify its contents against the source. Pages remains manual-only.
Validation:
- `npm run check:fast`
- `python3 scripts/prepare-website.py` and `python3 scripts/verify-website.py`
- Browser checks at `/` and `/rea/`: downloads, prompt/command copying, responsive layouts, keyboard navigation, reduced motion and DX-Ball comparison controls.
- Published REA 4.1.0 preload analysis and feature traces; source/claim/privacy review.
- Real Hopper/Ghidra verification: not performed (website-only change).
macOS 11 and later ship system libraries only inside the dyld shared
cache. trace_dylib_resolution therefore left every /usr/lib and /System
load undetermined, and REA had no way to see what a cache contains.
inspect_dyld_shared_cache (CLI inspect-dyld-shared-cache) reads a cache
file set selected by path, without an open target. Its pure TypeScript
reader parses modern and legacy dyld_cache_header layouts and reports:
- UUID, architecture, platform, OS version, Mac Catalyst alternate,
cache type, shared region, max slide;
- every VM mapping;
- each subcache by file suffix, checked by UUID. A missing or
mismatched subcache makes coverage partial;
- the complete image list.
For each requested install path it reads that image's load commands
through the cache mappings with the existing Mach-O reader.
trace_dylib_resolution accepts an optional shared_cache. Absolute
candidates found in a cache that serves the process's CPU family resolve
with the new outcome and status `shared-cache`. That status also takes
precedence over later in-bundle fallbacks such as a back-deployed
libswiftCore. The result names the cache UUID and OS version it used.
verify:dyld-shared-cache checks every image in the host cache (3,649 on
macOS 26.6), with dependencies and their re-export, weak, upward and
delay-init attributes, against `dyld_info -all_dyld_cache
-linked_dylibs`. It also checks CLI/MCP parity, and that each
shared-cache edge in the fixture trace is a library dyld actually loads.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(pages): reserve publishing for the REA website
* docs(website): add worked guides and agent prompt examples
* ci(website): check relevant PRs and publish only on demand
* docs(website): clarify IDA coverage and analysis session lifetime
project_apple_application_graph accepted only IPA inventories. It now
also projects macOS .app directories and ZIP or DMG inventories that
contain applications. It reports application roots, platforms, and
every nested bundle with its path-convention role, layout, parent,
Info.plist path, executable candidates, and signing paths:
- roles: app extensions, XPC services (including inside frameworks),
frameworks, login items, system and driver extensions, plug-ins,
resource bundles, and helper or nested applications;
- privileged helpers in Contents/Library/LaunchServices and launchd
plists in Contents/Library/LaunchAgents or LaunchDaemons;
- direct files of Contents/Helpers;
- symlinks by path. Their targets are not inventoried, so versioned
frameworks are described through their real Versions/<version>/
directories.
AppleDouble sidecars (._* and __MACOSX/) are excluded from roles. A DMG
whose contents could not be inventoried is reported as partial. Other
inventory formats remain input errors. The projection still reads no
plist or CMS semantics; it points callers at inspect_plist.
The result schema is additive. projection_id now covers the new fields,
so IPA projection IDs differ from earlier versions.
verify:macos-bundle compiles a source-owned app with clang (CLT only).
It packs it as a directory, a ditto ZIP, and an APFS DMG, and checks
identical anatomy through the CLI with stdio MCP parity. It runs in the
macOS CI job.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(website): add local introduction and DX-Ball showcase
* feat(website): illustrate REA and DX-Ball reconstruction
Add two explanatory figures and an interactive comparison of recorded DX-Ball x86 instructions with maintained C. Preserve the full instruction listing, caller excerpt, constant reads, figure prompts and evidence provenance.
Validation: Chrome preview at local root and /rea/; five comparison steps; 22 clipboard checks; keyboard navigation; responsive widths 1440, 768, 390 and 320. Verify assembly against saved REA Evidence and C against the pinned public source. Existing execution and compiler results are cited, not rerun.
* docs(website): show the REA queries behind source recovery
Present compact function, caller and byte-read request excerpts alongside recorded evidence. Keep the overview and case narrative focused on REA, with provider provenance in contributor notes. Replace raster references with maintained SVG source diagrams while retaining the established layout and image-generation prompts.
Validation: local and /rea/ Chrome previews, five comparison steps, 22 clipboard checks, responsive widths and keyboard navigation passed. SVG XML and all 41 local links/assets passed checks.
* feat(browser): trace captured website module imports
Compose verified exported scripts, inert native import syntax and a narrow
Chromium resolution port. Preserve query/fragment identity, import-map context,
all reported capture candidates and explicit unknown execution. Share named
CLI/MCP contracts and retain native reports, diagnostics and cleanup failures.
Add focused boundary regressions and a real module-loading oracle to the
existing conditional Chrome lane. No captured application code is executed
by module tracing, and no selected website asset is refetched.
Refs #851, #774
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* fix(browser): preserve module contexts and source bindings
Keep capture completeness inline, validate actual source text against its
digest, and separate unsupported URL contexts from malformed input.
Use the importer origin for native module resolution and the selected map
base independently, including mixed HTTP/HTTPS and escaped HTML attributes.
Refs #851, #774
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* feat(javascript): recover readable modules with Wakaru
Integrate pinned upstream source behind a typed optional recovery port,
preserve source identity and producer provenance, and publish verified
artifacts for existing JavaScript analysis through CLI and MCP.
Add focused real-tool CI with genuine bundler fixtures and bounded
resource, cancellation, cleanup, schema and packaged-artifact checks.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* fix(javascript): align recovery catalog and browser verification
Share canonical Wakaru provider identity, describe CLI mode options, and
format the new workflow test. Disconnect the real page transport after
its observation readiness progress instead of a fixed delay.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* refactor(javascript): consume the investigation evidence writer
Verify that the SDK recovery handler records complete Evidence through the newly shared investigation owner.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
* chore(test): account for isolated recovery fixture compilers
* fix(javascript): validate recovery paths before acquiring the engine
Return field-specific invalid_format diagnostics for relative paths and verify compiled CLI/MCP parity even when the optional engine is unavailable.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* feat(web): export captured scripts for JavaScript application analysis
Add capture adapters, conservative source path projection, verified exclusive publication and a shared CLI/MCP workflow. Retain source identities, missing content and competing URL versions in a complete manifest for existing static analysis.
Refs: #774Closes: #799
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* test(web): include script export in independent effect audit
Register the new contract family in the independent inventory audit and verify the filesystem, session, network and idempotence metadata of the named export contract.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* fix(web): preserve capture limitation values in export manifests
Align the export boundary with passive capture’s string contract instead of rejecting an admitted empty limitation. Cover the publisher boundary and regenerate canonical schemas and identities.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* feat(browser): capture correlated network content and lifecycle evidence
Keep request identities and redirect relationships across ordered scenario events.
Capture selected browser-exposed bytes and duplicate headers through modular
readers, settle bounded reads before cleanup, and preserve unavailable states.
Keep content-selection mismatches and legacy network coverage unknown during
comparison. Verify real CLI/MCP, installed package, and browser lifecycle paths.
Refs #774
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* fix(browser): align capture defaults and redact encoded media metadata
Correct the complete documented request and legacy default expectation caught by CI. Cover and redact encoded declared values in captured body media types.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
* test(browser): verify polling capture cutoff completeness
Accept incomplete event coverage only when the continuously polling fixture has explicit unfinished requests; retain failures for other missing or truncated sections. Include actual completeness and unfinished transactions in diagnostics.
Assisted-by: Yukari (gensokyo-skills:yukari-boundary-analysis)
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
SafeOutputTree.create() called fchmod(0o700) on a directory handle and
commit() fsynced a directory handle. Both are POSIX-only: Windows rejects
them with EPERM ("operation not permitted, fchmod"/"fsync"), so every
extract_artifact call failed with a generic io error and produced no output.
Skip the redundant directory chmod and directory fsync on Windows. mkdir
already applies the requested mode, and file contents are synced in write().
Add a Windows regression test and run it, plus the archive-safety test, in
the curated Windows CI lane so the behavior is covered.
* feat(ghidra): enable Windows x64 read-only analysis
* test(ghidra): exercise native Windows runtime boundaries
* test(ghidra): verify current Windows inventory and isolate ACL observer
* fix(ghidra): align Windows admission and address P0 review
morluto
Ysopwnable
Kaoru0822
Yunare Maia
Rudy Mizrahi Celekli
Dery Ferdika