* fix: fall back to repo default_branch instead of hardcoded "main"
When no explicit base_branch input is provided, the action previously
fell back to a hardcoded "main", which fails on repositories whose
default branch is named differently (e.g. "master", "develop").
This reads repository.default_branch from the GitHub event payload
(populated once in parseGitHubContext) and uses it as the fallback in
all three callsites: agent/index.ts, run.ts, and update-comment-link.ts.
Explicit env/input precedence is preserved; "main" remains only as a
last-resort defensive fallback if the payload somehow lacks the field.
* test: drop unused BASE_BRANCH env handling from default_branch test
agent/index.ts no longer reads process.env.BASE_BRANCH directly (it now
goes through context.inputs.baseBranch which is set on the mock context),
so saving/clearing/restoring that env var in the regression test is dead
code.
* Revert "chore: bump Claude Code to 2.1.89 and Agent SDK to 0.2.89"
This reverts commit bee87b3258c251f9279e5371b0cc3660f37f3f77.
* Revert "chore: bump Claude Code to 2.1.88 and Agent SDK to 0.2.88"
This reverts commit 7225f045c6219dd201504adc5534baf31024db31.
* Restore .claude/ and .mcp.json from PR base branch before CLI runs
The CLI's non-interactive mode trusts cwd: it reads .mcp.json and
.claude/settings{,.local}.json from the working directory and acts on
them before any tool-permission gating — executing hooks, setting env
vars (NODE_OPTIONS, LD_PRELOAD), running apiKeyHelper shell commands,
and auto-approving MCP servers. When this action checks out a PR head,
these files are attacker-controlled.
Rather than enumerate dangerous keys, replace the entire .claude/ tree
and .mcp.json with the versions from the PR base branch (which a
maintainer has reviewed). Paths absent on base are deleted. Uses local
git state, so no TOCTOU against the GitHub API.
* Read PR base ref from payload for config restore in agent mode
Agent mode's branchInfo.baseBranch defaults to "main" (or env/input
override) instead of the PR's actual target branch — it doesn't query
prData.baseRefName like tag mode does. This meant a PR targeting
develop would get .claude/ restored from main.
Fix by reading pull_request.base.ref directly from the webhook payload
for pull_request, pull_request_review, and pull_request_review_comment
events. For issue_comment on a PR (no base.ref in payload), fall back
to the mode-provided value — tag mode's value is correct (from GraphQL);
agent mode on issue_comment is an edge case that at worst restores from
the wrong trusted branch, which is still secure.
The payload value passes through validateBranchName for defense-in-depth
(GitHub enforces valid branch names server-side, but we validate anyway).
* Extend restored paths to .gitmodules, .ripgreprc, .claude.json
.gitmodules defines submodule URLs and paths; path-confusion attacks
against git submodule operations can write into .git/hooks. .ripgreprc
can set --pre (arbitrary command on each file) if RIPGREP_CONFIG_PATH
points at it. .claude.json is cheap defense-in-depth.
Documented why .git/ is excluded (not trackable in commits, and
restoring it would undo the PR checkout), along with .gitconfig
(git never reads it from cwd) and shell rc files (sourced from $HOME,
not cwd — checkout cannot reach $HOME).
* feat(inline-comment): add confirmed param + probe-pattern safety net
Subagents that inherit this tool sometimes probe it with test comments
('Test comment to see if I can create inline comments') after hitting
unrelated errors elsewhere. Recurring issue across customer PRs.
Adds two defenses:
1. confirmed param: set true to post (final review comments should pass
this). When false, buffers to a JSONL file instead of posting.
2. Probe-pattern safety net: when confirmed is omitted (backward compat
for existing prompts), the body is checked against obvious probe
patterns ('test comment', 'can i', 'does this work', etc.). Matching
calls are buffered instead of posted.
A post-run step in action.yml reports the buffered call count and bodies
as a workflow warning for diagnostics.
Backward compatibility:
- Existing single-agent prompts (no confirmed param) post normally unless
the body happens to start with a probe phrase (unlikely for real
review comments)
- The code-review skill is being updated to pass confirmed: true in its
final posting step
- Subagent probes that would previously post now harmlessly buffer
* refactor: replace probe-regex with Haiku classification in post-step
The regex approach was narrow and could miss creative probe phrasings.
Replaced with a batch Haiku classification that runs after the session
completes.
Flow:
- MCP server: confirmed !== true -> buffer to JSONL (no classification
in-band, no latency in the tool path)
- Post-step (src/entrypoints/post-buffered-inline-comments.ts): reads
buffer, sends all bodies to a single Haiku call, posts only those
classified as real review comments
- confirmed=false entries are never posted regardless of classification
Fail-open: if ANTHROPIC_API_KEY is unavailable (Bedrock/Vertex users)
or the classification call fails, posts all unconfirmed comments. This
matches pre-PR behavior where all calls posted immediately.
The post-step emits :⚠️: for each filtered comment so users can
see what was dropped and why.
* feat: add classify_inline_comments opt-out input
New action input classify_inline_comments (default 'true'). Setting to
'false' restores pre-buffering behavior: all inline comment calls post
immediately regardless of the confirmed param.
Threads through: action input -> CLASSIFY_INLINE_COMMENTS env ->
context.inputs.classifyInlineComments -> MCP server env ->
CLASSIFY_ENABLED module const.
Post-step is also gated on the input so it skips entirely when
classification is disabled.
* docs: document classify_inline_comments input and confirmed param
- usage.md: add classify_inline_comments to inputs table
- solutions.md: mention confirmed=true in the prompt example and explain
buffering/classification in the tool permissions section
Add a `display_report` input parameter (default: "true") that controls
whether the Claude Code Report is written to the GitHub Step Summary.
Setting it to "false" allows users with custom formatting solutions to
avoid duplicate output in the step summary.
Closes#206
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Ashwin Bhat <ashwin-ant@users.noreply.github.com>
Replace the over-engineered Mode interface/registry/detector pattern with
straightforward inline logic. There are only 2 modes (tag and agent) and
the complexity wasn't justified.
- Delete Mode interface, registry, and prepare pass-through modules
- Export prepareTagMode() and prepareAgentMode() as standalone functions
- Inline trigger checking and mode dispatch in run.ts/prepare.ts
- Change generatePrompt/createPrompt to take modeName string instead of Mode
- Remove dead code (extractGitHubContext, unused detector helpers)
- Update CLAUDE.md to reflect new architecture
* refactor: unify action into single composite step with run.ts entrypoint
Consolidate the prepare and base-action phases into a single composite
step that runs src/entrypoints/run.ts. This simplifies the action.yml
from multiple steps to one execution step, while keeping the same
behavior.
Key changes:
- Add src/entrypoints/run.ts as unified entrypoint
- Simplify action.yml to single 'Run Claude Code Action' step
- Pass all inputs via environment variables
- Update base-action to accept inputs via env vars
- Support agent mode auto-detection from prompt input
* refactor: keep SSH signing cleanup and token revocation as separate action steps
Move SSH signing key cleanup and app token revocation back to separate
composite action steps in action.yml with always() conditions, rather
than handling them inside run.ts. This keeps these cleanup concerns
as independently visible steps in the workflow.
* fix: address PR review feedback
- Use path.dirname() instead of manual string slicing for executable path
- Differentiate prepare vs execution errors in catch block so tracking
comment accurately reflects which phase failed
- Update CLAUDE.md architecture docs to reflect unified run.ts entrypoint
and four-phase design
* fix: address PR review feedback
- Use path.dirname() instead of manual string slicing for executable path
- Differentiate prepare vs execution errors in catch block so tracking
comment accurately reflects which phase failed
- Rewrite CLAUDE.md to focus on mental model, key concepts, and gotchas
instead of exhaustive file listings
* feat: add ssh_signing_key input for SSH commit signing
Add a new ssh_signing_key input that allows passing an SSH signing key
for commit signing, as an alternative to the existing use_commit_signing
(which uses GitHub API-based commits).
When ssh_signing_key is provided:
- Git is configured to use SSH signing (gpg.format=ssh, commit.gpgsign=true)
- The key is written to ~/.ssh/claude_signing_key with 0600 permissions
- Git CLI commands are used (not MCP file ops)
- The key is cleaned up in a post step for security
Behavior matrix:
| ssh_signing_key | use_commit_signing | Result |
|-----------------|-------------------|--------|
| not set | false | Regular git, no signing |
| not set | true | GitHub API (MCP), verified commits |
| set | false | Git CLI with SSH signing |
| set | true | Git CLI with SSH signing (ssh_signing_key takes precedence)
* docs: add SSH signing key documentation
- Update security.md with detailed setup instructions for both signing options
- Explain that ssh_signing_key enables full git CLI operations (rebasing, etc.)
- Add ssh_signing_key to inputs table in usage.md
- Update bot_id/bot_name descriptions to note they're needed for verified commits
* fix: address security review feedback for SSH signing
- Write SSH key atomically with mode 0o600 (fixes TOCTOU race condition)
- Create .ssh directory with mode 0o700 (SSH best practices)
- Add input validation for SSH key format
- Remove unused chmod import
- Add tests for validation logic
* feat: add Agent SDK support with USE_AGENT_SDK feature flag
Add a feature-flagged code path that uses the Agent SDK instead of
spawning the CLI as a subprocess. When USE_AGENT_SDK=true is set,
the new SDK path is used; otherwise, existing CLI behavior is unchanged.
Changes:
- Add parse-sdk-options.ts for parsing ClaudeOptions into SDK format
- Add run-claude-sdk.ts for SDK execution with query() function
- Update run-claude.ts with feature flag check at entry point
- Update update-comment-link.ts to handle both cost_usd and total_cost_usd
- Add @anthropic-ai/claude-agent-sdk dependency
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* refactor: simplify SDK types by using @anthropic-ai/claude-agent-sdk types directly
- Remove duplicate SdkRunOptions and McpStdioServerConfig types
- Use SDK's Options and McpStdioServerConfig types directly
- Return { sdkOptions, showFullOutput, hasJsonSchema } from parseSdkOptions
- Remove unnecessary convertMcpServers function
- Net reduction of ~70 lines
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* refactor: use extraArgs for claudeArgs pass-through to CLI
Simplify option parsing by converting claudeArgs to extraArgs record
and letting the SDK/CLI handle --mcp-config, --json-schema, etc.
- Remove extractJsonSchema and parseMcpConfigs functions
- Add parseClaudeArgsToExtraArgs for simple flag parsing
- CLI handles complex args like --mcp-config directly
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* ci
* refactor: remove hardcoded permission bypass flags
The SDK path should match CLI path behavior - permissions are handled
by the CLI itself, not hardcoded in the action.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* chore: add logging for SDK vs CLI path selection
---------
Co-authored-by: Claude <noreply@anthropic.com>
* chore: remove experimental allowed domains feature
Remove the experimental_allowed_domains feature which was used to
restrict network access via a Squid proxy. This removes:
- The input definition from action.yml
- The Network Restrictions workflow step
- The setup-network-restrictions.sh script
- Documentation from experimental.md, usage.md, and related files
- The input default from collect-inputs.ts
* chore: fix formatting with prettier
Co-authored-by: Ashwin Bhat <ashwin-ant@users.noreply.github.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Ashwin Bhat <ashwin-ant@users.noreply.github.com>
* Remove mcp_config input in favor of --mcp-config in claude_args
BREAKING CHANGE: The mcp_config input has been removed. Users should now use --mcp-config flag in claude_args instead.
This simplifies the action's input surface area and aligns better with the Claude Code CLI interface. Users can still add multiple MCP configurations by using multiple --mcp-config flags.
Migration:
- Before: mcp_config: '{"mcpServers": {...}}'
- After: claude_args: '--mcp-config {"mcpServers": {...}}'
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Add outer action MCP tests to workflow
- Add test-outer-action-inline-mcp job to test inline MCP config via claude_args
- Add test-outer-action-file-mcp job to test file-based MCP config via claude_args
- Keep base-action tests unchanged (they still use mcp_config parameter)
- Test that MCP tools are properly discovered and can be executed through the outer action
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Fix: Add Bun setup to outer action MCP test jobs
The test jobs for the outer action were failing because Bun wasn't installed.
Added Setup Bun step to both test-outer-action-inline-mcp and test-outer-action-file-mcp jobs.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Add id-token permission to outer action MCP test jobs
The outer action needs id-token: write permission for OIDC authentication
when using the GitHub App. Added full permissions block to both
test-outer-action-inline-mcp and test-outer-action-file-mcp jobs.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Use github_token parameter instead of id-token permission
Replace id-token: write permission with explicit github_token parameter
for both outer action MCP test jobs. This simplifies authentication by
using the provided GitHub token directly.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Use RUNNER_TEMP environment variable consistently
Changed from GitHub Actions expression syntax to environment variable
for consistency with the rest of the workflow file.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Use execution_file output from action instead of hardcoded path
Updated outer action test jobs to:
- Add step IDs (claude-inline-test, claude-file-test)
- Use the execution_file output from the action steps
- This is more reliable than hardcoding the output file path
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* tmp
* Fix MCP test assertions to match actual output format
Updated the test assertions to match the actual JSON structure:
- Tool calls are in assistant messages with type='tool_use'
- Tool results are in user messages with type='tool_result'
- The test tool returns 'Test tool response' not 'Hello from test tool'
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Make inline MCP test actually use the tool instead of just listing
Changed the inline MCP test to:
- Request that Claude uses the test tool (not just list it)
- Add --allowedTools to ensure the tool can be used
- Check that the tool was actually called and returned expected result
- Output the full JSON for debugging
This makes both tests (inline and file-based) consistent in their approach.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat: implement Claude Code GitHub Action v1.0 with auto-detection and slash commands
Major features:
- Mode auto-detection based on GitHub event type
- Unified prompt field replacing override_prompt and direct_prompt
- Slash command system with pre-built commands
- Full backward compatibility with v0.x
Key changes:
- Add mode detector for automatic mode selection
- Implement slash command loader with YAML frontmatter support
- Update action.yml with new prompt input
- Create pre-built slash commands for common tasks
- Update all tests for v1.0 compatibility
Breaking changes (with compatibility):
- Mode input now optional (auto-detected)
- override_prompt deprecated (use prompt)
- direct_prompt deprecated (use prompt)
* test + formatting fixes
* feat: simplify to two modes (tag and agent) for v1.0
BREAKING CHANGES:
- Remove review mode entirely - now handled via slash commands in agent mode
- Remove all deprecated backward compatibility fields (mode, anthropic_model, override_prompt, direct_prompt)
- Simplify mode detection: prompt overrides everything, then @claude mentions trigger tag mode, default is agent mode
- Remove slash command resolution from GitHub Action - Claude Code handles natively
- Remove variable substitution - prompts passed through as-is
Architecture changes:
- Only two modes now: tag (for @claude mentions) and agent (everything else)
- Agent mode is the default for all events including PRs
- Users configure behavior via prompts/slash commands (e.g. /review)
- GitHub Action is now a thin wrapper that passes prompts to Claude Code
- Mode names changed: 'experimental-review' → removed entirely
This aligns with the philosophy that the GitHub Action should do minimal work and delegate to Claude Code for all intelligent behavior.
* fix: address PR review comments for v1.0 simplification
- Remove duplicate prompt field spread (line 160)
- Remove async from generatePrompt since slash commands are handled by Claude Code
- Add detailed comment explaining why prompt → agent mode logic
- Remove entire slash-commands loader and directories as Claude Code handles natively
- Simplify prompt generation to just pass through to Claude Code
These changes align with v1.0 philosophy: GitHub Action is a thin wrapper
that delegates everything to Claude Code for native handling.
* chore: remove unused js-yaml dependencies
These were added for slash-command YAML parsing but are no longer
needed since we removed slash-command preprocessing entirely
* fix: remove experimental-review mode reference from MCP config
The inline comment server configuration was checking for deprecated
'mode' field. Since review mode is removed in v1.0, this conditional
block is no longer needed.
* prettify
* feat: add claudeArgs input for direct CLI argument passing
- Add claude_args input to both action.yml files
- Implement shell-style argument parsing with quote handling
- Pass arguments directly to Claude CLI for maximum flexibility
- Add comprehensive tests for argument parsing
- Log custom arguments for debugging
Users can now pass any Claude CLI arguments directly:
claude_args: '--max-turns 3 --mcp-config /path/to/config.json'
This provides power users full control over Claude's behavior without
waiting for specific inputs to be added to the action.
* refactor: use industry-standard shell-quote for argument parsing
- Replace custom parseShellArgs with battle-tested shell-quote package
- Simplify code by removing unnecessary -p filtering (Claude handles it)
- Update tests to use shell-quote directly
- Add example workflow showing claude_args usage
This provides more robust argument parsing while reducing code complexity.
* bun format
* feat: add claudeArgs input for direct CLI argument passing
- Add claude_args input to action.yml for flexible CLI control
- Parse arguments with industry-standard shell-quote library
- Maintain proper argument order: -p [claudeArgs] [legacy] [BASE_ARGS]
- Keep tag mode defaults (needed for functionality)
- Agent mode has no defaults (full user control)
- Add comprehensive tests for new functionality
- Add example workflow showing usage
* format
* refactor: complete v1.0 simplification by removing all legacy inputs
- Remove all backward compatibility for v1.0 simplification
- Remove 10 legacy inputs from base-action/action.yml
- Remove 9 legacy inputs from main action.yml
- Simplify ClaudeOptions type to just timeoutMinutes and claudeArgs
- Remove all legacy option handling from prepareRunConfig
- Update tests to remove references to deleted fields
- Remove obsolete test file github/context.test.ts
- Clean up types to remove customInstructions, allowedTools, disallowedTools
Users now use claudeArgs exclusively for CLI control.
* fix: update MCP server tests after removing additionalPermissions
- Change github_ci server logic to check for workflow token presence
- Update test names to reflect new behavior
- Fix test that was incorrectly setting workflow token
* model version update
* Update package json
* remove deprecated workflow file (tests features we no longer support)
* Simplify agent mode and re-add additional_permissions input
- Agent mode now only triggers when explicit prompt is provided
- Removed automatic triggering for workflow_dispatch/schedule without prompt
- Re-added additional_permissions input for requesting GitHub permissions
- Fixed TypeScript types for mock context helpers to properly handle partial inputs
- Updated documentation to reflect simplified mode behavior
* Fix MCP config not being passed to Claude CLI
The MCP servers (including github_comment server) were configured but not passed to Claude. This caused the "update_claude_comment" tool to be unavailable.
Changes:
- Write MCP config to a file at $RUNNER_TEMP/claude-mcp-config.json
- Add mcp_config_file output from prepare.ts
- Pass MCP config file via --mcp-config flag in claude_args
- Use fs/promises writeFile to match codebase conventions
* Fix MCP tool availability and shell escaping in tag mode
Pass MCP config and allowed tools through claude_args to ensure tools like
mcp__github_comment__update_claude_comment are properly available to Claude CLI.
Key changes:
- Tag mode outputs claude_args with MCP config (as JSON string) and allowed tools
- Fixed shell escaping vulnerability when JSON contains single quotes
- Agent mode passes through user-provided claude_args unchanged
- Re-added mcp_config input for users to provide custom MCP servers
- Cleaned up misleading comments and unused file operations
- Clarified test workflow is for fork testing
Security fix: Properly escape single quotes in MCP config JSON to prevent
shell injection vulnerabilities.
Co-Authored-By: Claude <noreply@anthropic.com>
* bun format
* tests, typecheck, format
* registry test update
* Update agent mode to have github server as a default
* Fix agent mode to include GitHub MCP server with proper token
* Simplify review workflow - prevent multiple submissions
- Rename workflow to avoid conflicts
- Remove review submission tools
- Keep only essential tools for reading and analyzing PR
* Add GitHub MCP server and context prefix to agent mode
- Include main GitHub MCP server (Docker-based) by default
- Fetch and prefix GitHub context to prompts when in PR/issue context
- Users no longer need to manually configure GitHub tools
* Delete .github/workflows/claude-auto-review-test.yml
* Remove github_comment and inline_comment servers from agent mode defaults
- Agent mode now only includes the main GitHub MCP server by default
- Users can add additional servers via mcp_config if needed
- Reduces unnecessary MCP server overhead
* Remove all default MCP servers from agent mode
Agent mode now starts with no default servers - users must explicitly configure any MCP servers they need via mcp_config input
* Remove GitHub context prefixing and clean up agent mode
- Remove automatic GitHub context fetching and prefixing
- Remove unused imports (fetcher, formatter, context checks)
- Clean up comments
- Agent mode now simply passes through the user's prompt as-is
* Add GitHub MCP support to agent mode
- Parse --allowedTools from claude_args to detect when user wants GitHub MCPs
- Wire up github_inline_comment server in prepareMcpConfig for PR contexts
- Update agent mode to use prepareMcpConfig instead of manual config
- Add comprehensive tests for parseAllowedTools edge cases
- Fix TypeScript types to support both entity and automation contexts
* Format code with prettier
* Fix agent mode test to expect branch values
* Fix agent test to handle dynamic branch names from environment
* Better fix: Control environment variables in agent test for predictable behavior
* minor formatting
* Simplify MCP configuration to use multiple --mcp-config flags
- Remove MCP config merging logic from prepareMcpConfig
- Update agent and tag modes to pass multiple --mcp-config flags
- Let Claude handle config merging natively through multiple flags
- Fix TypeScript errors in test file
This approach is cleaner and relies on Claude's built-in support for multiple --mcp-config flags instead of manual JSON merging.
* feat: Copy project subagents to Claude runtime environment
Enables custom subagents defined in .claude/agents/ to work in GitHub Actions by:
- Checking for project agents in GITHUB_WORKSPACE/.claude/agents/
- Creating ~/.claude/agents/ directory if needed
- Copying all .md agent files to Claude's runtime location
- Following same pattern as slash commands for consistency
Includes comprehensive test coverage for the new functionality.
* formatting
* Add auto-fix CI workflows with slash command and inline approaches
- Add /fix-ci slash command for programmatic CI failure fixing
- Create auto-fix-ci.yml workflow using slash command approach
- Create auto-fix-ci-inline.yml workflow with full inline prompt
- Both workflows automatically analyze CI failures and create fix branches
* Add workflow_run event support and auto-fix CI workflows
- Add support for workflow_run event type in GitHub context
- Create /fix-ci slash command for programmatic CI failure fixing
- Add auto-fix-ci.yml workflow using slash command approach
- Add auto-fix-ci-inline.yml workflow with full inline prompt
- Both workflows automatically analyze CI failures and create fix branches
- Fix workflow syntax issues with optional chaining operator
* Use proper WorkflowRunEvent type instead of any
* bun formatting
* Remove auto-fix workflows and commands from v1-dev
These files should only exist in km-anthropic fork:
- .github/workflows/auto-fix-ci.yml
- .github/workflows/auto-fix-ci-inline.yml
- slash-commands/fix-ci.md
- .claude/commands/fix-ci.md
The workflow_run event support remains as it's useful for general automation.
* feat: Expose GitHub token as action output for external use
This allows workflows to use the Claude App token obtained by the action
for posting comments as claude[bot] instead of github-actions[bot].
Changes:
- Add github_token output to action.yml
- Export token from prepare.ts after authentication
- Allows workflows to use the same token Claude uses internally
* Debug: Add logging and always output github_token in prepare step
* Fix: Add git authentication to agent mode
Agent mode now fetches the authenticated user (claude[bot] when using Claude App token)
and configures git identity properly, matching the behavior of tag mode.
This fixes the issue where commits in agent mode were failing due to missing git identity.
* minor bun format
* remove unnecessary file
* fix: Add branch environment variable support to agent mode for signed commits
- Read CLAUDE_BRANCH and BASE_BRANCH env vars in agent mode
- Pass correct branch info to MCP file ops server
- Enables signed auto-fix workflows to create branches via API
* feat: Add auto-fix CI workflow examples
- Add auto-fix-ci example with inline git commits
- Add auto-fix-ci-signed example with signed commits via MCP
- Include corresponding slash commands for both workflows
- Examples demonstrate automated CI failure detection and fixing
* fix: Fix TypeScript error in agent mode git config
- Remove dependency on configureGitAuth which expects ParsedGitHubContext
- Implement git configuration directly for automation contexts
- Properly handle git authentication for agent mode
* fix: Align agent mode git config with existing patterns
- Use GITHUB_SERVER_URL from config module consistently
- Remove existing headers before setting new ones
- Use remote URL with embedded token like git-config.ts does
- Match the existing git authentication pattern in the codebase
* refactor: Use shared configureGitAuth function in agent mode
- Update configureGitAuth to accept GitHubContext instead of ParsedGitHubContext
- This allows both tag mode and agent mode to use the same function
- Removes code duplication and ensures consistent git configuration
* feat: Improve error message for 403 permission errors when committing
When the github_file_ops MCP server gets a 403 error, it now shows a cleaner
message suggesting to rebase from main/master branch to fix the issue.
* docs: Update documentation for v1.0 release (#476)
* docs: Update documentation for v1.0 release
- Integrate breaking changes naturally without alarming users
- Replace deprecated inputs (direct_prompt, custom_instructions, mode) with new unified approach
- Update all examples to use prompt and claude_args instead of deprecated inputs
- Add migration guides to help users transition from v0.x to v1.0
- Emphasize automatic mode detection as a key feature
- Update all workflow examples to @v1 from @beta
- Document how claude_args provides direct CLI control
- Update FAQ with automatic mode detection explanation
- Convert all tool configuration to use claude_args format
* fix: Apply prettier formatting to documentation files
* fix: Update all Claude model versions to latest and improve documentation accuracy
- Update all model references to claude-4-0-sonnet-20250805 (latest Sonnet 4)
- Update Bedrock models to anthropic.claude-4-0-sonnet-20250805-v1:0
- Update Vertex models to claude-4-0-sonnet@20250805
- Fix cloud-providers.md to use claude_args instead of deprecated model input
- Ensure all examples use @v1 instead of @beta
- Keep claude-opus-4-1-20250805 in examples where Opus is demonstrated
- Align all documentation with v1.0 patterns consistently
* feat: Add dedicated migration guide as requested in PR feedback
- Create comprehensive migration-guide.md with step-by-step instructions
- Add prominent links to migration guide in README.md
- Update usage.md to reference the separate migration guide
- Include before/after examples for all common scenarios
- Add checklist for systematic migration
- Address Ashwin's feedback about having a separate, clearly linked migration guide
* feat: Add comprehensive examples for hero use cases
- Add dedicated issue deduplication workflow example
- Add issue triage example (moved from .github/workflows)
- Update all examples to use v1-dev branch consistently
- Enable MCP tools in claude-auto-review.yml
- Consolidate PR review examples into single comprehensive example
Hero use cases now covered:
1. Code reviews (claude-auto-review.yml)
2. Issue triaging (issue-triage.yml)
3. Issue deduplication (issue-deduplication.yml)
4. Auto-fix CI failures (auto-fix-ci/auto-fix-ci.yml)
All examples updated to follow v1-dev paradigm with proper prompt and claude_args configuration.
* refactor: Remove timeout_minutes parameter from action (#482)
This change removes the custom timeout_minutes parameter from the action in favor of using GitHub Actions' native timeout-minutes feature.
Changes:
- Removed timeout_minutes input from action.yml and base-action/action.yml
- Removed all timeout handling logic from base-action/src/run-claude.ts
- Updated base-action/src/index.ts to remove timeoutMinutes parameter
- Removed timeout-related tests from base-action/test/run-claude.test.ts
- Removed timeout_minutes from all example workflow files (19 files)
Rationale:
- Simplifies the codebase by removing custom timeout logic
- Users can use GitHub Actions' native timeout-minutes at the job/step level
- Reduces complexity and maintenance burden
- Follows GitHub Actions best practices
BREAKING CHANGE: The timeout_minutes parameter is no longer supported. Users should use GitHub Actions' native timeout-minutes instead.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Claude <noreply@anthropic.com>
* refactor: Remove unused slash commands and agents copying logic
Removes experimental file copying features that had no default content:
- Removed experimental_slash_commands_dir parameter and related logic
- Removed automatic project agents copying from .claude/agents/
- Eliminated flaky error-prone cp operations with stderr suppression
- Removed 175 lines of unused code and associated tests
These features were infrastructure without default content that used
problematic error handling patterns (2>/dev/null || true) which could
hide real filesystem errors.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* docs: Remove references to timeout_minutes parameter
The timeout_minutes parameter was removed in commit 986e40a but
documentation still referenced it. This updates:
- docs/usage.md: Removed timeout_minutes from inputs table
- base-action/README.md: Removed from inputs table and example
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: km-anthropic <km-anthropic@users.noreply.github.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Kashyap Murali <13315300+katchu11@users.noreply.github.com>
