Fixes#1868 — when CLAUDE_CONFIG_DIR is set to a non-default location
(e.g. ~/.config/claude for XDG compliance, or a multi-tenant install
path), the plugin still wrote state files to the hardcoded ~/.claude/
path, leaving stale state and breaking CLAUDE_CONFIG_DIR's purpose.
Resolution precedence (highest first):
1. SECURITY_WARNINGS_STATE_DIR — plugin-specific override (existing)
2. CLAUDE_CONFIG_DIR/security — CC's config-dir env (new — #1868)
3. ~/.claude/security — default fallback (unchanged)
Empty-string env vars (e.g. CLAUDE_CONFIG_DIR= in a misconfigured
shell) are treated as not-set so the empty path doesn't collide with
os.path.join and silently write to /security at the filesystem root.
Implementation: a single state_dir() helper in _base.py is the source
of truth for resolution. All five modules that previously had inline
SECURITY_WARNINGS_STATE_DIR / ~/.claude/security resolutions
(_base.py, session_state.py, ensure_agent_sdk.py, llm.py, and one
site in security_reminder_hook.py) now call state_dir() instead.
Re-implementing the precedence inline risks drift — one module gets
a future fix, others don't.
The helper is called per-invocation rather than cached at import time
so test monkeypatches of the env vars take effect, and so a long-
running test or future shared-process scenario can change the env
between calls and have the next call observe the new value. The
per-call cost is negligible compared to the subprocess-spawn cost
the hooks pay every fire in production.
Three hardcoded ~/.claude/security strings remain but are NOT
functional resolutions:
- _base.py:39: the fallback BRANCH inside state_dir() itself
- ensure_agent_sdk.py:6, :11: docstring text describing default
location for users
Verified locally on macOS Python 3.13:
- py_compile clean on all 5 modified files.
- Existing 45 smoke + extensibility tests still pass.
- 14 new tests in test_claude_config_dir.py (added to internal test
suite at sg-staging/tests/, not in this PR):
* 7 resolution-semantics: default fallback, CLAUDE_CONFIG_DIR
override, SECURITY_WARNINGS_STATE_DIR beats both, tilde
expansion, empty-string handling (CLAUDE_CONFIG_DIR= must
fall back, NOT join to /security).
* 4 static-shape: each of session_state / ensure_agent_sdk /
llm / security_reminder_hook either imports state_dir from
_base OR has zero resolution patterns. Catches the
regression where someone adds a new state-file writer and
re-implements resolution inline, missing the
CLAUDE_CONFIG_DIR branch.
* 3 end-to-end: with CLAUDE_CONFIG_DIR set, get_state_file /
get_lock_file return paths under <CLAUDE_CONFIG_DIR>/security/;
save_state round-trip writes a file to the redirected path
and re-reads the same contents.
- 59/59 pass total (45 existing + 14 new) in 2.54s.
NOT verified end-to-end with a real CC instance setting
CLAUDE_CONFIG_DIR. The shape tests catch the regression class
(hardcoded ~/.claude/), and the end-to-end test pins the behavior
that user state files actually land at the redirected path.
Closes#1868.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixesanthropics/claude-plugins-official#2056 — on Windows, when the
worktree contains an untracked file whose name has a character undefined
in cp1252 (accented capitals like Á Í Ï Ð Ý, most CJK, emoji), the
UserPromptSubmit hook crashes:
Exception in thread Thread-5 (_readerthread):
UnicodeDecodeError: 'charmap' codec can't decode byte 0x81
Traceback (most recent call last):
File diffstate.py, line 338, in _list_untracked
for p in r.stdout.split('\\0'):
AttributeError: 'NoneType' object has no attribute 'split'
Non-blocking (UPS failures still let the prompt through) but the
baseline-untracked snapshot is silently lost, so the Stop-hook review
mis-handles pre-existing untracked files.
Root cause (reporter's diagnosis, verified):
1. core.quotePath=false makes git emit raw UTF-8 for non-ASCII filenames.
2. subprocess.run(..., text=True) decodes via
locale.getpreferredencoding(False) in strict mode — on Windows that
is cp1252, in which 0x81 / 0x8D / 0x8F / 0x90 / 0x9D are undefined.
Those bytes appear in the UTF-8 encodings of Á (C3 81), Í (C3 8D),
Ï (C3 8F), Ð (C3 90), Ý (C3 9D), and a large fraction of CJK / emoji
codepoints.
3. The decode runs in the subprocess reader thread. The thread raises
UnicodeDecodeError, threading prints 'Exception in thread Thread-N',
subprocess.run returns with stdout=None. The handler then does
None.split('\\0') -> AttributeError, which is NOT in the narrow
except (TimeoutExpired, FileNotFoundError, OSError) tuple, so it
escapes the helper, propagates out of UserPromptSubmit's
ThreadPoolExecutor.result(), and exits the hook non-zero.
This is internally inconsistent: gitutil._git_diff_range,
security_reminder_hook._reflog_amend_lookup (line ~540), and the commit
diff loop (line ~1115) already do bytes + decode utf-8/replace, with
comments explicitly noting that text=True would crash. The fix below
extends that established pattern to the helpers that were holdouts.
Affected helpers (6 total):
- diffstate._list_untracked <- reporter, hot path, CRITICAL
- diffstate.capture_git_baseline <- reporter, latent
- diffstate.get_baseline_file_content <- audit, file content read, HIGH
- gitutil._git_name_only <- reporter, latent
- gitutil._git_status_porcelain <- reporter, latent
- gitutil._git_reflog_recent_commits <- audit, embeds %gs commit msg, HIGH
For each one:
- Drop text=True from subprocess.run.
- Decode r.stdout / r.stderr as .decode('utf-8', errors='replace').
- Add ValueError to the except tuple as defense against any future
strict-decode regression (UnicodeDecodeError is a ValueError
subclass; including it explicitly degrades the helper to its
empty/None return instead of escaping out of the hook).
Verified locally on macOS Python 3.13:
- py_compile clean on both files.
- 45 existing smoke + extensibility tests still pass.
- 21 new internal tests (not in this PR — added to the team's local
test suite at staging/tests/test_unicode_decode.py):
* 18 static-shape parametrized: each of the 6 fixed helpers has
no text=True in its subprocess calls, contains errors='replace',
and lists ValueError in its except.
* Deterministic end-to-end: create real git repo + Ávila_report.txt
untracked, call _list_untracked, verify it returns
{'Ávila_report.txt': <mtime>} without crashing.
* Deterministic end-to-end: same for capture_git_baseline (verifies
the latent stderr-warning case stays valid).
* Deterministic end-to-end: get_baseline_file_content on a file
whose content has 山田太郎 + 🎉; verify the bytes round-trip
through the decode.
- 66/66 tests pass total (45 existing + 21 new).
NOT verified end-to-end on Windows — would need actual cp1252 strict
decode to fire. Reporter has the deterministic repro and will
re-verify on their Win11 / Python 3.14.x setup before merge.
Not in this PR (defense-in-depth, lower risk):
- 3 git rev-parse calls returning path output (gitutil._find_git_index,
_git_toplevel, _git_dir) could fail on Windows if cwd is in a
non-ASCII install directory. Same fix shape but unreported and
much lower probability — worth a separate follow-up if anyone
actually hits it.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes#410, #2037, #2045, #1640, #1280, #1329, #1341, #255,
anthropics/claude-code#46720 (partial closes on overlap with other rules).
The plugin's substring-only XSS / browser-DOM rules
(new_function_injection, react_dangerously_set_html, document_write_xss,
innerHTML_xss, outerHTML_xss, insertAdjacentHTML_xss) fired on any file
containing the trigger substring — including:
* Markdown documentation explaining XSS sinks
* Blog posts / READMEs that name browser APIs
* Python tutorials referencing dangerouslySetInnerHTML
* Plugin skill files with example HTML strings
* .yaml / .json configs that happen to contain the literal string
* .gitignore / Dockerfile / Makefile
These constructs have no meaning outside JS/TS source. Add a
path_filter: lambda p: p.endswith(_JS_EXTS) to each so they fire only
on .js, .jsx, .ts, .tsx, .mjs, .cjs, .mts, .cts, .vue, .svelte.
Cross-checked against the existing _JS_EXTS-gated rules
(regex_exec_substring, child_process_exec, exec_substring) — same
pattern, same constant, same intent. Uses the module-level _JS_EXTS
tuple so future extension changes propagate to all 6 rules atomically.
Verified locally on macOS Python 3.13:
- py_compile clean.
- 45-test existing smoke + extensibility suite still passes.
- 151 new parametrized tests in test_xss_gate.py (added to internal
test suite this PR doesn't ship): each gated rule x every
JS-family extension accepts, x every non-JS path (.md / .py /
.yaml / .json / .txt / .html / Dockerfile / Makefile / .gitignore
/ .sh / .go / .rs / .rb) rejects. 196 tests pass total.
Doesn't address everything in the false-positive cluster — issues that
require Python-rule gating (#1114 .env.schema exec), tighter substring
scoping (#660 pickle in usernames), or hook-protocol changes (#1358
exit-2 vs warning, #1375 plain-text-vs-JSON output) need separate PRs.
This PR covers the JS-substring subset cleanly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixes anthropics/claude-plugins-official#2071 — on macOS where the
default `python3` is Apple's Command Line Tools Python 3.9.6, the
plugin's agentic commit reviewer silently does not run, even when the
user has a newer Python installed.
Three compounding factors in the bug:
1. `sg-python.sh` only checks the major version (`3`), so it always
picks 3.9 even when 3.10+ is on PATH.
2. `claude_agent_sdk` requires Python >=3.10 — pip install on 3.9
returns "No matching distribution" -> bootstrap returns BUILD_FAILED.
3. Even with a hand-built 3.12 venv, `llm.py` imports the SDK
in-process into the hook's interpreter (still 3.9), which raises
SyntaxError. The existing venv-probe in `ensure_agent_sdk.py` uses
the venv's own Python (3.12) so it reports NOOP_VENV (healthy) while
the consumer fails — misleading telemetry on top of silent feature
degradation.
Per BQ telemetry, 14,073 external macOS users hit
sdk_bootstrap=BUILD_FAILED in the past 4 days (the default-macOS
cohort), out of ~86K total external installed users. Combined with
~20K other users in similar broken-bootstrap states (Windows pre-#2055,
Linux <3.10), about half the installed base has a silently-broken
agentic reviewer.
This PR implements the reporter's items #1, #3, and #4. Item #2
(running the SDK out-of-process) is deferred as a bigger refactor.
Item #1 — hooks/sg-python.sh — prefer >=3.10 binaries via 3-pass probe:
Pass 1: python3.13 / 3.12 / 3.11 / 3.10 (>=3.10 by name, highest wins)
Pass 2: bare python3 / python / py -3 (accept only if reported >=3.10)
Pass 3: bare python3 / python / py -3 (any Python 3, FALLBACK so
pattern checks still work on macOS-default 3.9 — no regression
vs today; SDK-dependent paths detect the version mismatch
inside Python and degrade cleanly via item #4)
Item #4 — ensure_agent_sdk.py — health-check honesty:
Added HOOK_PY_INCOMPATIBLE=6 outcome with short-circuit at top of main():
if sys.version_info < (3, 10):
return HOOK_PY_INCOMPATIBLE, "hook_py", f"py_{...}"
Telemetry consequences after rollout: sdk_bootstrap=6 is a new clean
bucket; some users currently miscounted in sdk_bootstrap=3 BUILD_FAILED
(wasted pip cycles) and sdk_bootstrap=1 NOOP_VENV (falsely-healthy)
move to sdk_bootstrap=6. The remaining NOOP_VENV count becomes
trustworthy.
Item #3 — ensure_agent_sdk.py — one-time user-visible notice:
When outcome == HOOK_PY_INCOMPATIBLE and a marker file at
`~/.claude/security/.agentic_unavailable_notice_v<pv>` doesn't exist,
the SessionStart response includes hookSpecificOutput.additionalContext
+ systemMessage explaining the situation. Marker file is plugin-
version-keyed so a future fix (e.g. shipping out-of-process SDK) can
bump pv and re-notify users.
BUILD_FAILED is intentionally excluded from the notice — it covers
transient causes where a permanent banner would mislead.
Verified locally on macOS Python 3.13:
- py_compile clean on both files.
- Existing 45-test smoke + extensibility suite: 45/45 PASS in 2.50s.
- Unit test of simulated 3.9 path: HOOK_PY_INCOMPATIBLE returned with
correct phase/kind; notice shown on first call, suppressed on
second, reshown on bumped pv; BUILD_FAILED correctly does NOT
trigger notice.
NOT verified: actual Python 3.9 behavior end-to-end (would need a 3.9
install). Worth a follow-up smoke test in a 3.9 venv before next
release. The unit test simulating 3.9 covers the logic but not the
runtime invocation through the shim.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Picked up from sethconvex's PR #1966 (auto-closed by membership gate),
split off from #1980 (Convex plugin entry refresh) so the editorial
addition to claude-automation-recommender gets its own review.
Changes:
- SKILL.md: add `convex` to the package.json dep-detection grep, update
the Database row in the indicator table to name Convex, and add a
Convex MCP row to the MCP recommendation table.
- references/mcp-servers.md: new "Convex MCP" section in the Databases
group (Supabase / Convex / PostgreSQL / Neon / Turso), and a row in
the Detection Patterns quick reference.
Convex publishes its MCP server via the `convex` npm package
(`npx convex mcp start`), exposing tables, function-spec, data,
run-once-query, logs, env list/set/get. Same row pattern as the
existing database/backend MCP entries.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The first round of this PR removed SKIP_WIN32, fixed venv_py to use
Scripts/python.exe, and added Lib/site-packages to the consumer glob —
all necessary. Windows verification (Win11 ARM64, Py 3.13, Git Bash)
showed two more blockers, both addressed here.
1. Pip dependency resolver picks unbuildable cryptography on ARM64.
Without --prefer-binary, pip picks a cryptography version with no
published ARM64 wheel and tries to build it from source. That needs
Rust/Cargo, almost never present on user machines → BUILD_FAILED
with err_kind=other:cryptography. A binary wheel exists for an
adjacent version (cryptography-46.0.3-cp311-abi3-win_arm64.whl);
--prefer-binary tells pip to pick it. Cross-platform safe (no-op
where the latest version already has a wheel).
2. pywin32 .pth files aren't processed by sys.path.insert().
With the venv built, ensure_agent_sdk.py's post-build probe passes
(it runs from venv_py, where Python's site.py at startup processes
pywin32.pth and registers win32/, win32/lib/ plus runs
pywin32_bootstrap.py to set the DLL search dir). But llm.py runs in
the hook's SYSTEM Python and adds the venv via sys.path.insert(),
which doesn't trigger site.py at all. Without the bootstrap, the
SDK's mcp.client.stdio → mcp.os.win32.utilities chain raises
ModuleNotFoundError: pywintypes and the agentic reviewer falls back
to single-shot silently — exactly the symptom this PR is trying to
fix. The probe says NOOP_VENV; the actual consumer fails. Probe and
consumer use different Pythons.
Replicate what site.py would do: after inserting site-packages,
also insert win32/ and win32/lib/, then exec pywin32_bootstrap.py.
Pulled into a shared helper _inject_agent_sdk_venv_into_syspath()
so both consumer sites (3P SDK fallback, agentic_review fallback)
call the same code — Windows handling stays in one place.
Verified on macOS (POSIX path unchanged):
- Helper end-to-end test: POSIX-layout venv detected + fake package
imports successfully via the injected path
- Windows-layout venv also detected; win32 branch correctly skipped
via sys.platform check
- Both files pass py_compile
Credit: @mhegazy verified the previous commit on Win11 ARM64 / Py 3.13
/ Git Bash, surfaced both issues end-to-end, and provided the exact
fix patterns. This commit applies them with the pywin32 part factored
into a shared helper (vs. inlining at both consumer sites).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The agentic reviewer is silently no-op on Windows today. SessionStart
bootstrap (ensure_agent_sdk.py) short-circuits with SKIP_WIN32 because
the consumer glob in llm.py only matches POSIX venv layout
(lib/pythonX.Y/site-packages). On Windows, venvs use Lib/site-packages
(capital L, no pythonX.Y subdir), so even if a venv got built the
glob wouldn't find its contents.
Result: Windows users on default installs (no system-wide
claude_agent_sdk) get layer 1 (pattern warnings) and layer 2 (single-
shot LLM diff review) but not layer 3 — the cross-file agentic review
that catches IDOR, auth-bypass, cross-file SSRF, and other things that
need to read related files. Plugin description claims layer 3 but it
silently doesn't run.
Three changes:
1. llm.py — extend the consumer glob (2 sites: 3P SDK fallback at
~L297, agentic_review fallback at ~L1090) to also match the Windows
Lib/site-packages layout, so a venv built on Windows is actually
discoverable.
2. ensure_agent_sdk.py — remove the sys.platform == 'win32' early-exit
so the SessionStart bootstrap builds the venv on Windows too.
Outcome code 4 (formerly SKIP_WIN32) is retired but not reused so
pre-fix telemetry rows still decode correctly.
3. ensure_agent_sdk.py — venv_py path now branches on sys.platform:
Windows venvs put the interpreter at Scripts\python.exe; POSIX
uses bin/python. Previously assumed POSIX, so even with the glob
fix, the post-build SDK-importability probe would fail on Windows.
Verified locally on macOS:
- glob test: both layouts now match (POSIX venv detected, simulated
Windows venv also detected via the new Lib/site-packages branch)
- both files pass py_compile
- POSIX path unchanged (sys.platform != 'win32' so old branch runs)
Not verified on Windows in this commit — needs an actual Windows
runner to confirm the venv build + SDK import + subprocess plumbing
all work end-to-end. The SDK spawns a child claude.exe; Windows
process plumbing has its own quirks (shell semantics, path escaping)
that may surface separately. Worth a controlled rollout (one-week
soak under env-var opt-in before flipping default).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixes#2043. On Git Bash for Windows, Claude Code hands script paths to
the shim in POSIX form (`/c/Users/...`). We exec a Windows `python.exe`
(the `python3` Microsoft Store stub fails the probe), and Windows Python
interprets the leading `/` as the root of the current drive — `/c/...`
becomes `C:\c\Users\...` or `D:\c\Users\...` depending on which
drive the shell happens to be on, fails with ENOENT, and every
Edit/Write/MultiEdit blocks until the session restarts.
Convert absolute path args via `cygpath -w` (a Git Bash builtin) before
exec. Guarded by `command -v cygpath` so macOS/Linux fall straight
through unchanged; `cygpath -w` is idempotent on already-Windows paths
so the rare mixed-form case is safe. Only `/*` paths are converted —
Windows-form paths reaching the shim are already openable by python.exe.
Verified locally:
- cygpath absent on macOS → guard skips → POSIX behavior unchanged
- end-to-end shim invocation with a POSIX path on macOS exits 0
- stubbed cygpath -w on /c/Users/test/hook.py produces C:\Users\test\hook.py
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Aligns the compose service name, local config filename, and all
log/restart commands with the image and binary name. Adds an explicit
-config arg since the image CMD still defaults to the legacy
/etc/mcp-gateway path.
🏠 Remote-Dev: homespace
- Replace doc references with platform.claude.com URLs (overview,
quickstart, security, deploy-compose, deploy-helm, console,
troubleshooting, reference, WIF)
- Swap the POC mcp-proxy image for the public registry digest used in
the published quickstart
🏠 Remote-Dev: homespace
Adds the /create-docker-mcp-tunnel command, which drives the MCP tunnels
Docker Compose quickstart end to end: preflight checks, certificate
generation, proxy config, cloudflared, an optional sample FastMCP server,
and verification from Managed Agents and the Messages API.
Migrated from anthropic-experimental/mcp-tunnel-skills.
🏠 Remote-Dev: homespace
Paths containing spaces (common on Windows, e.g. C:\Users\Some User\...)
cause shell word-splitting when CLAUDE_PLUGIN_ROOT is unquoted, resulting
in hooks erroring with "No such file or directory" on every tool call.
Wraps the path in double quotes for all five affected hook commands.
Fixes the pattern reported in issue #57946. Closes the fix surfaced in PR #1921.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- modernize-harden: never edits legacy/ anymore. Writes findings plus a
reviewed unified diff to analysis/<system>/security_remediation.patch.
A second security-auditor pass reviews each hunk (RESOLVES / PARTIAL /
INTRODUCES-RISK) before presenting. The user reviews and applies the
patch deliberately, then re-runs to verify. This makes every command
consistent with the recommended deny Edit(legacy/**) workspace setting,
so the README's exception note is gone.
- modernize-map: restructure the parse-target list around three stack-
agnostic principles (dispatcher targets are variables; code-storage
joins live in config; entry points live in deployment descriptors), with
COBOL/Java/web/CLI examples on equal footing rather than COBOL-dominant.
Same protections against false dead-code findings, less stack-specific.
- security-auditor agent: rephrase coverage items in stack-neutral terms
(record layouts/temp datasets, resource ACLs, deployment scripts/job
definitions, batch input records) so the checklist reads naturally for
COBOL, Java EE, .NET, and web targets alike.
- README: drop the harden exception note; describe the patch workflow.
Fixes found by running the discovery workflow against the AWS CardDemo
mainframe sample (~50 KLOC of COBOL/CICS/JCL/BMS/VSAM):
- modernize-assess: add scc -> cloc -> find/wc fallback chain with the
COCOMO-II formula so Step 1 works when scc isn't installed; same for
portfolio-mode cloc/lizard. Drop the reference to a specific
agent-spawning tool name (just "in parallel"). Sharpen the structural-
map subagent prompt: 5-12 domains, subgraph clustering, ~40-edge cap,
repo-relative paths, dangling-reference check.
- modernize-map: expand the parse-target list with the things a
literal-minded reader would miss on a real mainframe codebase — CICS
CSD DEFINE TRANSACTION/FILE for entry points and online file I/O,
EXEC CICS file ops, SELECT...ASSIGN TO joined with JCL DD,
EXEC SQL table refs (not JCL DD), SEND/RECEIVE MAP, dynamic
data-name XCTL resolution, COBOL fixed-format column slicing. Without
these the dead-code list is wrong (most CICS programs look unreachable).
Also write a machine-readable topology.json alongside the summary.
- modernize-extract-rules: add a Priority (P0/P1/P2) field with a
heuristic, and an optional Suspected-defect field. modernize-brief
reads P0 rules to build the behavior contract, but the Rule Card had
no priority slot — the chain was broken.
- modernize-brief: read the new P0 tags; flag low-confidence P0 rules as
SME blockers.
- modernize-reimagine: drop "for the demo" wording.
- security-auditor agent: add mainframe/COBOL coverage items (RACF,
JCL/PROC creds, BMS field validation, DB2 dynamic SQL, copybook PII)
and mark web-only items as such so it adapts to the target stack.
- README: add Optional Tooling section and a symlink example for the
expected layout.
- modernize-brief: read TOPOLOGY.html (what modernize-map actually
produces) instead of nonexistent TOPOLOGY.md, and tell the user which
command produces each missing input.
- README: rewrite the Commands section to match actual command behavior —
correct output filenames, ordering (brief is the synthesis/approval gate
after discovery, not the first step), agent attributions, and required
args. Add a workspace-layout note and an explicit callout that
modernize-harden edits legacy/, which conflicts with the recommended
deny rule. Reconcile the Overview and Typical Workflow sequences.
- modernize-assess: generalize the production-runtime overlay step so it
no longer assumes a specific MCP server/tool; mark it optional. Fix
app/jcl/ -> legacy/$1/jcl/ for layout consistency.
- modernize-map: make TOPOLOGY.html self-contained (load Mermaid from a
CDN) so it renders in any browser; drop assumptions about an external
artifact renderer. Generalize the telemetry annotation note.
- business-rules-extractor agent: fix command cross-reference to the
actual command name.
- plugin.json: include the brief step in the workflow description.
/maker-setup now falls back to GitHub's tarball endpoint when git isn't
on PATH, instead of detouring through a package-manager git install.
curl and tar ship with macOS, Linux, and Windows 10 1803+ out of the
box, so this is zero-install on every target platform — and a CwC
attendee just needs the files once to flash a device, not git history.
- maker-setup.md: git-clone fast path, curl|tar (Unix) / curl+tar+
Rename-Item (PowerShell) fallback, normalizes the -main suffix
- m5-onboard/SKILL.md: drop git from required deps + per-OS git
bootstrap block; keep Python bootstrap
- README: git now listed as optional
Linear: CC-1975
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
🏠 Remote-Dev: homespace
* Add cwc-makers plugin: /keep-thinking Cardputer onboarding
Packages the Code-with-Claude Makers (claude.com/cwc-makers) Cardputer
experience as a one-command flow for event attendees:
- commands/keep-thinking.md: user entry point — clones
moremas/build-with-claude and runs the m5-onboard provisioning flow
- skills/m5-onboard/SKILL.md: vendored from upstream onboard/SKILL.md;
Installation section replaced with clone-location note; explicit
'relay physical button steps to user' directive added
- skills/cardputer-buddy/SKILL.md: post-onboarding app iteration
All three are user-invocable; /keep-thinking is the intended entry
point. Skill content is Apache-2.0 from the upstream repo.
Linear: CC-1975
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
🏠 Remote-Dev: homespace
* Rename /keep-thinking → /start-making
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
🏠 Remote-Dev: homespace
* Rename /start-making → /maker-setup
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
🏠 Remote-Dev: homespace
Several agent files used <example> blocks containing user: "..." /
assistant: "..." turn markers, embedded as \n-escaped strings inside
the YAML frontmatter description: field. Replace those with flat prose
trigger descriptions in description: and a 'When to invoke' section
in the agent body containing prose-bullet scenarios.
Affected files:
- 5 agent definitions:
- plugins/hookify/agents/conversation-analyzer.md
- plugins/pr-review-toolkit/agents/code-reviewer.md
- plugins/pr-review-toolkit/agents/pr-test-analyzer.md
- plugins/pr-review-toolkit/agents/type-design-analyzer.md
- plugins/pr-review-toolkit/agents/comment-analyzer.md
- 5 agent-development skill files updated to teach the new format:
- plugins/plugin-dev/skills/agent-development/SKILL.md
- .../references/triggering-examples.md
- .../references/agent-creation-system-prompt.md
- .../examples/complete-agent-examples.md
- .../examples/agent-creation-prompt.md
Routing is unaffected: each description: still names the trigger
surface in plain English.
Ports anthropics/skills#547 (b0cbd3d) so this repo matches the upstream
skills repo.
improve_description.py and run_loop.py now shell out to `claude -p` instead
of using the Anthropic SDK directly, so the description optimizer uses the
session's existing Claude Code auth and no longer requires a separate
ANTHROPIC_API_KEY. SKILL.md drops the stale extended-thinking reference and
adds guidance for updating an existing skill.
Several enterprise customers sync exclusively from this repo (not
anthropics/skills, whose README disclaims production use), so they have been
stuck on the old SDK-based path.
- analyze-sessions.mjs: track per-session start/end/tokens and emit
by_day[] in JSON output (date, dow, tokens, peak concurrency,
per-session spans). Hoist shared token-sum in commit loop.
- template.html: new "session timeline by day" section — horizontal
day pills (% of total + session count) drive a lane-packed gantt of
concurrent sessions, colored by project, with hover details and
←/→ keyboard nav. Extract drillList() helper and use it for both
top-prompts and cache-breaks (5 visible + "show more" toggle).
Generates an explorable HTML report of Claude Code session usage from
local ~/.claude/projects transcripts: total tokens, cache efficiency,
per-project/subagent/skill breakdowns, most expensive prompts with
transcript context, and cache breaks. Terminal-styled, single-file
output with sortable tables and expandable drill-downs.
Fixes#993 (Permission denied on hook scripts) without relying on
client-side +x preservation.
The hook executor spawns commands via /bin/sh -c, which requires +x
to execute a script directly. Prefixing with bash reads the script
as data — mode bits are irrelevant. This works on all Claude Code
versions, whereas the client-side chmod fix (claude-cli #24666) only
shipped in v2.1.86.
All 3 scripts declare #!/bin/bash and use bashisms ([[ ]], =~), so
bash (not sh) is the correct interpreter.
The version field forces a fresh cache path (1.0.0/ instead of
unknown/), ensuring the new hooks.json reaches users with stale
caches.
Deep-mode allows bounded local computation but must NOT use WebFetch
or WebSearch. Finding the solution on AoPS is not solving the problem.
Adds explicit NO WEB prompt block and orchestrator self-restraint note.
Found by Ralph's test run (skill solved 5/6 then started fetching
dgrozev.wordpress.com and artofproblemsolving.com for P6).
The skill that addresses the Proof-or-Bluff gap: self-verified 85.7% IMO
becomes <5% under human grading. Uses fresh-context verifiers armed with
specific failure patterns (not generic 'check logic').
Validated: 17/18 IMO+Putnam 2025 solved, 0 false positives, 2 novel proofs.
See eval data in anthropic monorepo sandbox/sandbox/ralph/math_skills/.
P0 follow-up for EA-471. Updates plugin-dev teaching materials to stop
recommending the commands/ directory layout for new plugins:
- command-development/SKILL.md: add legacy banner at top pointing to
skills/ format
- create-plugin.md: update scaffolding to create skills/<name>/SKILL.md
instead of commands/; mark commands/ as acceptable legacy alternative;
update all examples, tables, and testing instructions
- example-plugin: migrate example-command to skills/example-command/SKILL.md;
keep commands/example-command.md with a legacy-format note; update README
to reflect new preferred structure
Both formats remain loaded identically — this is a documentation change only.
Refs: anthropics/apps#26827
Co-authored-by: Henry Shi <henrys@anthropic.com>
* fix readme typo
* fix(plugin-dev): add missing .claude-plugin/plugin.json
The plugin-dev plugin was missing its required plugin.json manifest file,
causing the plugin to fail loading. This adds the missing configuration
file following the same format as other official plugins.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* Add README and setup documentation for Greptile plugin
- Add README.md with setup instructions for getting API key
- Document the GREPTILE_API_KEY environment variable requirement
- Add homepage, author URL, and keywords to plugin.json
- Update description to reflect Greptile as AI code review agent
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* feat: add c7 agent
* Update Context7 plugin for v2 API
- Update skill/agent/command to use new query-docs tool (replaces get-library-docs)
- Add query parameter usage for intelligent reranking
- Add version pinning support (e.g., /vercel/next.js/v15.1.8)
- Add tools and model metadata to agent
- Simplify docs to focus on workflow, not parameter details
- Add README.md with usage examples
* Switch Context7 MCP to remote HTTP server
* feat: update tools with better skill/agent format prompt
* fmt
* fix: installation guide
* Change Notion name to lowercase in marketplace.json
According to the SKILLS spec (see https://agentskills.io/specification#:~:text=Max%2064%20characters.%20Lowercase%20letters%2C%20numbers%2C%20and%20hyphens%20only.%20Must%20not%20start%20or%20end%20with%20a%20hyphen.) names should not contain uppercase letters. This prevents loading the marketplace in spec-compliant agents.
Update the name to be in lowercase.
* Fix empty array crash on bash 3.2 in setup-ralph-loop.sh
* Update Vercel plugin to point to vercel-labs/vercel-plugin
Replace the marketplace pointer for the Vercel plugin from
vercel/vercel-deploy-claude-code-plugin to vercel-labs/vercel-plugin.
* vercel-labs to vercel
* docs(ralph-loop): add Windows compatibility section
Retargeted from PR #124 (originally against plugins/ralph-wiggum/,
since renamed). Documents the Git Bash workaround for Windows users
hitting WSL bash resolution issues in the stop hook.
Original author: @stefanzvonar
* add(plugin): terraform — HashiCorp infrastructure-as-code
Adapted from PR #14 by @gautambaghel (HashiCorp).
Original: https://github.com/anthropics/claude-plugins-official/pull/14
* add(plugin): autofix-bot — DeepSource automated code review
Adapted from PR #23 by @jai-deepsource (DeepSource).
Original: https://github.com/anthropics/claude-plugins-official/pull/23
* add(plugin): stagehand — Browserbase browser automation
Adapted from PR #43 by @Kylejeong2 (Browserbase). PR's marketplace.json
had a syntax error (missing '},' before adjacent entry); entry
reconstructed from the diff.
Original: https://github.com/anthropics/claude-plugins-official/pull/43
* add(plugin): atomic-agents — BrainBlend-AI framework
Adapted from PR #46 by @KennyVaneetvelde (BrainBlend-AI).
Original: https://github.com/anthropics/claude-plugins-official/pull/46
* add(plugin): microsoft-docs — official Microsoft documentation MCP
Adapted from PR #55 by @TianqiZhang (Microsoft).
Original: https://github.com/anthropics/claude-plugins-official/pull/55
* add(plugin): bonfire — session-context workflow tooling
Adapted from PR #108 by @vieko (Vercel).
Original: https://github.com/anthropics/claude-plugins-official/pull/108
* Add intercom to marketplace
* Add neon to marketplace
* Remove qodo SHA
* Merge staging into add-plugin/intercom to resolve conflict
* Merge latest staging to resolve conflict
* Remove external_plugins changes from staging
Moved to external-plugins-staging branch for separate review.
---------
Co-authored-by: Han T. <han.tan@shopify.com>
Co-authored-by: Julien Tavernier <jtavernier@Juliens-MacBook-Pro.local>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: Daksh Gupta <daksh510@gmail.com>
Co-authored-by: Fahreddin Özcan <ozcanfahrettinn@gmail.com>
Co-authored-by: Matt Kotsenas <Matt.Kotsenas@gmail.com>
Co-authored-by: LuciferDono <pranavj821@gmail.com>
- Split pipeline into two steps (extract lines, then parse) mirroring the
original structure.
- set +e around the jq call so failures reach the $? check instead of
aborting under set -e.
- The "no text content" branch remains removed (that was the original bug —
all-tool-use turns now correctly yield empty text and the loop continues).
The state file lives at .claude/ralph-loop.local.md — project-scoped,
not session-scoped. The plugin's Stop hook fires in every Claude Code
session open in that project directory. So if session A starts a loop,
session B's Stop events also find the state file and block, feeding A's
prompt into B and consuming A's iteration budget.
This was masked by the transcript-parsing bug fixed in the previous
commit: that bug deleted the state file on the first Stop in any
session, so neither session looped. Fixing it exposed the leak.
Fix: setup writes CLAUDE_CODE_SESSION_ID into the frontmatter; the hook
compares against .session_id from its stdin JSON and exits silently on
mismatch. State files without session_id (written by old setup scripts)
fall through to preserve existing behavior.
Claude Code writes each assistant content block (text/tool_use/thinking)
as its own JSONL line. The hook's `grep role:assistant | tail -1` would
grab whichever block happened to be last — often tool_use — then jq's
text filter returned empty string, triggering the 'no text content' path
which deletes the state file and exits without blocking.
Net effect: the loop silently never fires. In one observed session, 62%
of assistant lines were tool_use-only; the hook deleted state on the
very first Stop event every time.
Fix: slurp all assistant lines with jq -rs, flatten to text blocks only,
take the last. Empty result is now non-fatal — no text means no <promise>
tag, so the loop continues. Also absorbs jq parse errors (control chars
in text) via || fallback instead of aborting under set -e.
Skills and commands are now merged, so SKILL.md name fields surface
in the UI as slash commands. Upper Space names like 'Hook Development'
become '/Hook Development', which doesn't work since user-invoked
commands don't support spaces.
Rename all affected SKILL.md name fields to lower-kebab-case:
- hookify: Writing Hookify Rules -> writing-hookify-rules
- plugin-dev: Agent Development -> agent-development
- plugin-dev: Command Development -> command-development
- plugin-dev: Hook Development -> hook-development
- plugin-dev: MCP Integration -> mcp-integration
- plugin-dev: Plugin Settings -> plugin-settings
- plugin-dev: Plugin Structure -> plugin-structure
- plugin-dev: Skill Development -> skill-development
Also update references in plugin-dev/README.md.
Use YAML block scalars (|) for multi-line description fields that contain
<example> blocks with colons and special characters. Without block scalars,
the YAML parser fails because it interprets lines like 'user:' and
'assistant:' as new key-value pairs.
Affected files:
- plugins/plugin-dev/agents/agent-creator.md
- plugins/plugin-dev/agents/skill-reviewer.md
- plugins/plugin-dev/agents/plugin-validator.md
- plugins/pr-review-toolkit/agents/code-simplifier.md
