Treat subtype success with is_error:true as a failed run so CI does not
show a misleading green check when the review never actually ran.
Fixes#1495
Co-authored-by: syf2211 <syf2211@users.noreply.github.com>
installClaudeCode() pipes `curl -fsSL | bash -s --`. Bash exits with
the status of the last command, so when curl fails (429 rate limit,
403, or connection error) `bash -s` still exits 0 on empty stdin and
the action logs "Claude Code installed successfully". The 3-attempt
retry loop never triggers because the first attempt looks successful,
and the run later dies with "Executable not found in $PATH: claude".
Prefix the pipeline with `set -o pipefail;` so curl's non-zero exit
propagates through the pipe and the retry loop can actually kick in.
Extracted into buildInstallCommand() with regression tests covering
both the old buggy shape and the fixed one.
When snapshotting PR-authored sensitive paths into .claude-pr/, cpSync with
dereference:true throws ENOENT if a symlink target is missing on the PR head
(e.g. .claude/CLAUDE.md -> ../AGENTS.md). Fall back to copying the symlink
itself so restoreConfigFromBase can continue and restore trusted base versions.
Fixes#1398
When classify_inline_comments is enabled, create_inline_comment buffers calls
without confirmed=true. The model frequently re-issues the call with confirmed=true
after reading the buffered reply, which posts the comment live but leaves the
original buffered entry behind. The post-session replay step then posts it again,
so every inline comment lands twice.
Reconcile the buffer on a live post: after a confirmed comment is created, remove
any buffered entry matching the same path, line, startLine and body so it cannot be
replayed. Extracts the reconciliation into src/mcp/inline-comment-buffer.ts (the MCP
server module starts a server on import) and adds unit tests.
Co-authored-by: archievi <13202986+archievi@users.noreply.github.com>
* Add agent-approval-check composite action
Require N human approvals on PRs that contain agent-authored commits.
Posts an agent-approval-check commit status that repos mark as a
required check on protected branches.
This is a sanitized port of the check Anthropic runs internally on
every agent-authored PR — same detection rules, /approve <sha>
comment flow, sibling-PR-same-SHA guard, and fail-closed semantics,
with the Anthropic-specific path exemptions and kill-switch removed
and config moved to action inputs.
Co-Authored-By: Claude <noreply@anthropic.com>
* Drop stray internal acronym from comment
* agent-approval-check: require write-access approvers, pin deps, pagination + doc fixes
🏠 Remote-Dev: homespace
* agent-approval-check: prettier
🏠 Remote-Dev: homespace
* agent-approval-check: verify approver write permission via REST; commits(last:100); docstring
🏠 Remote-Dev: homespace
* agent-approval-check: use headRefOid; drop pull_request_review trigger and correct threat-model docs
🏠 Remote-Dev: homespace
* agent-approval-check: stale-notification wording, no-retry-on-4xx, docstring API-call count
🏠 Remote-Dev: homespace
* agent-approval-check: fail-closed sibling guard on commits-ordering edge; drop stale 'reviewed' from README
🏠 Remote-Dev: homespace
* agent-approval-check: drop hardcoded API-call counts from logs; clarify author write-access requirement in README
🏠 Remote-Dev: homespace
* agent-approval-check: count all agent-email commits (close-reopen bypass); validate REQUIRED_APPROVALS>=1; exempt_head_branches warning
🏠 Remote-Dev: homespace
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Octavian Guzu <oct@anthropic.com>
`validateBranchName` rejects branch names containing `@`, even though
`git check-ref-format` permits `@` and GitHub itself accepts such
branches. PRs whose head or base branch contains an `@` fail validation
in-process before any git operation, so the action errors out
immediately.
Branch names with `@` show up in real workflows: ticket conventions
like "TICKET-123@add-feature" (#998), leading-prefix conventions like
"@hotfix/...", and agent tooling that appends "@<sessionid>" (#1305).
There is no workaround other than renaming the branch, which is often
not under the user's control.
Branch names are never passed through a shell (git calls use
execFileSync argv arrays), so `@` carries no injection risk. This is
the same reasoning used to add `#` in #1167, `+` in #1248, and `,` in
#1310. The bare name "@" (HEAD shorthand in git revision syntax) and
the "@{" reflog sequence are still rejected.
- Add `@` to the validateBranchName whitelist regex, including the
leading position (the leading-character rule blocks option injection
via `-`, which `@` cannot cause)
- Reject the bare name "@" with a dedicated check
- Update the surrounding comment, JSDoc, and error message to match
- Add test cases for @-containing names and bare "@"
Fixes#998
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Adds unit tests for previously-uncovered branches in
src/entrypoints/format-turns.ts:
- detectContentType: malformed-JSON fall-through (objects and arrays)
and the default python classification for non-python/non-js code
- formatResultContent: non-string inputs (number, plain object)
- groupTurnsNaturally / formatGroupedContent: the system_other path
for non-init system turns
Tests only; no source changes. format-turns.ts line coverage rises
from ~86% and the file's non-CLI logic is now fully exercised.
Co-authored-by: hk <solanamobilech@gmail.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Issue/PR comments (#512) and the issue/PR body (#710) are filtered to the
trigger timestamp so content created or edited after an authorized trigger
cannot be injected into Claude's prompt (TOCTOU protection). Reviews and
inline review comments were not: fetchGitHubData returned reviewData filtered
by actor only, and formatReviewComments renders it into the prompt, so a
review submitted or edited after the trigger reached Claude verbatim.
filterReviewsToTriggerTime already existed (added alongside the comment filter
in #512) but was only wired to the image-download list, never to the returned
reviewData.
Filter reviewData.nodes through filterReviewsToTriggerTime and each review's
inline comments through filterCommentsToTriggerTime, alongside the existing
actor filter, then build the review image-processing lists from those
already-filtered nodes (removing a now-redundant second filter pass).
Strengthen the two integration tests to assert post-trigger and edited-after
reviews/comments are dropped.
parseAllowedTools (used to decide which GitHub MCP servers to install)
hand-rolled a regex parse of claude_args, while the tools actually
granted to Claude are parsed by parseClaudeArgsToExtraArgs in
base-action/src/parse-sdk-options.ts using shell-quote. The two parsers
diverged on two inputs (#1357):
- Multiple values after a single flag: for
`--allowedTools "Read" "Grep" "mcp__github__get_commit"` the regex
captured only "Read", so the github MCP server was not installed even
though mcp__github__get_commit was granted — tool calls then failed.
- Commented-out lines: the regex counted tools on `#`-prefixed lines
that the SDK parser strips, installing servers that were never used.
Reimplement parseAllowedTools on the same shell-quote tokenizer and the
same "accumulating flag consumes all consecutive non-flag values"
semantics, stripping comment lines first, so the install decision agrees
with the tools that are actually granted. Unquoted glob patterns (e.g.
`mcp__github__*`), which shell-quote yields as glob objects, are
recovered to their literal text to preserve existing behavior.
Closes#1357
Co-authored-by: bymle <229636660+bymle@users.noreply.github.com>
* fix(parse-sdk-options): prevent shell-quote from collapsing unquoted Bash(X:*) rules to bare Bash
shell-quote's parse() tokenizes unquoted `(`, `)` as control operators
and barewords containing `*` as glob ops, all returned as non-string
objects. parseClaudeArgsToExtraArgs filtered those out, so an unquoted
`--allowedTools View,Bash(gh:*),Bash(cat:*)` collapsed to bare `Bash` —
silently widening scoped permission rules to unrestricted Bash(*).
Escape shell control metachars to Unicode private-use placeholders
before parse() and restore after; extract .pattern from glob ops.
Preserves existing quote/whitespace handling.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* ci: retrigger
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
* Add pr-stamp-sweep review workflow
Adds a reusable workflow script that fans out one reviewer agent per
candidate PR to judge whether it can be approved as-is, then runs an
adversarial verification pass on each approval recommendation before
including it in the final list.
Reviewers read pre-fetched PR dossiers from /tmp/claude/pr-sweep/<n>.md
(metadata, body, comments, full diff) rather than calling gh directly,
and verify diff claims against the local checkout. PR numbers are
passed via args; the script fails fast if none are provided.
* Sharpen security checklist in reviewer prompts
Replace the general security-history note with explicit attack classes
both reviewer stages must check: prompt injection (untrusted content
reaching Claude's context, including via tool output), code execution
(shell commands, eval/spawn, workflow expressions), path traversal,
and credential exposure.
GitHub serves pasted attachments from /user-attachments/assets/<uuid>
with no file extension, so getImageExtension() silently defaulted to
".png". When the bytes are actually JPEG/GIF/WebP the downloaded file is
mislabeled, and the Read tool then sends a base64 image whose declared
media_type doesn't match its magic bytes — which the Anthropic API
rejects with `400 invalid_request_error` ("image was specified using the
image/png media type, but the image appears to be a image/jpeg image").
Sniff the real format from the buffer's magic bytes after download and
only fall back to the URL-based extension when the signature is
unrecognized. Adds a regression test for a JPEG at an extensionless URL.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
stripHiddenAttributes used the pattern `["'][^"']*["']` for each quoted
attribute, which matches an opening quote of either type and stops at the
first quote of either type. When a value contained the other quote
character — e.g. an apostrophe inside a double-quoted attribute like
`title="We'll do it"` — the match terminated at the apostrophe, so the
wrong span was removed and the surrounding text was corrupted (e.g.
`<Tooltip title="We'll do it" placement="top">` became
`<Tooltipll do it" placement="top">`).
This surfaced via the github_inline_comment MCP tool: suggestion blocks
whose code lines contain quotes were mangled before posting (#1366).
Match each quoted form per quote type (`"[^"]*"` and `'[^']*'`),
mirroring stripMarkdownLinkTitles, so a value may freely contain the
other quote character. The unquoted fallback is unchanged.
Closes#1366
Co-authored-by: bymle <229636660+bymle@users.noreply.github.com>
The formatted_context block sent to the agent omitted labels for both
issues and pull requests, even though the GraphQL queries already
fetched them. This caused agents to incorrectly report "no labels"
when labels existed, breaking any workflow that routes on label state
(e.g., drift-fix routing on drift:* labels in a Drift Watcher pattern).
Changes:
- Add 'PR Labels:' line to PR context output
- Add 'Issue Labels:' line to issue context output
- Both emit 'none' when no labels are present (explicit > omitted)
- Bump labels(first: 1) → labels(first: 100) in both queries; the
previous cap meant only one label would appear even after the
formatter fix
- Update existing tests + add 'with labels' tests for both PR and
issue branches
Discovered while building a GitHub Actions workflow that uses this
action for drift-watcher routing in an internal seed framework
(joshpayne-joby/slim-routines#6). The agent self-diagnosed the gap
by inspecting this action's source — a satisfying full-loop.
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Previously, prompt files at `${RUNNER_TEMP}/claude-prompts/` from a prior
invocation could persist on non-ephemeral self-hosted runners (where the
documented `RUNNER_TEMP` cleanup contract is not reliably honored). In
particular, `claude-user-request.txt` is only written by `create-prompt`
when a user request exists; an `agent`-mode invocation does not overwrite
it, so a stale value left by an earlier mention-mode job in another repo
would leak into a later agent-mode job's effective context on the same
runner agent.
Fix: `rm -rf` the directory before `mkdir` in both write sites
(`src/create-prompt/index.ts`, `src/modes/agent/index.ts`). Idempotent,
safe on hosted runners (where the dir is already empty), and self-heals
on self-hosted runners.
Closes#1287
