Move the conditional PR template logic from SKILL.md instructions into
a deterministic bash script. This makes the template retrieval reliable
and testable rather than depending on the LLM to interpret empty/null
output from the gh command.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Fetch the repository's PR template using `gh repo view --json
pullRequestTemplates` before writing the PR description. This ensures
PRs follow the repo's expected format and fill in required sections.
Co-Authored-By: Claude <noreply@anthropic.com>
New skill that guides agents to create and maintain concise AGENTS.md
documentation with CLAUDE.md as a symlink. Enforces minimal style with
headers, bullets, and skill references instead of verbose prose.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Apply the same gh repo view pattern for consistency across all skills
that diff against the default branch.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Dynamically detect the default branch using git symbolic-ref instead of
hardcoding "main". Falls back to "main" if detection fails.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Document that commits with changes primarily generated by a coding agent
should include the Co-Authored-By attribution in the footer. Clarify that
this is the only indicator of AI involvement that should appear in commits.
Co-Authored-By: Claude <noreply@anthropic.com>
The skill was recommending a command-based MCP server using uvx, but
Sentry MCP is a hosted service using streaming HTTP. Update to use
the correct URL-based configuration with type "http".
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add workflow prerequisites to ensure proper git hygiene:
- commit skill: Check if on main/master and prompt to create feature branch
- create-pr skill: Check for uncommitted changes and invoke commit skill first
- AGENTS.md: Add step to include new skills in .claude/settings.json allowlist
These changes help enforce the expected workflow where changes are made on
feature branches and properly committed before creating PRs.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Bare `git push` may push all tracked branches depending on git config.
Specify origin and current branch explicitly to ensure only the
working branch is pushed.
- Add Skill() permissions for all sentry-skills to the recommendation
- Update AGENTS.md to remind contributors to update the skills list
when adding new skills
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Never suggest commands with user-specific absolute paths
- Never suggest custom scripts that may have side effects
- Only include the package manager actually used by the project
(e.g., if pnpm is detected, don't suggest npm/yarn commands)
- Reorganize stack-specific commands by detection criteria
- Add explicit "Package Manager Rules" table for clarity
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
New skill that guides PR creation following Sentry's engineering practices:
- PR description structure (what, why, alternatives, context)
- Title format following commit conventions
- Issue reference syntax (Fixes/Refs)
- Guidelines on PR scope and reviewer assignment
Notably excludes "Test plan" sections per Sentry conventions.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
New skill that automates the CI feedback loop:
- Checks CI status first, waits for bot checks (sentry/codecov/cursor/bugbot/seer)
- Gathers review feedback from humans and bots
- Investigates failures via gh run view --log-failed
- Validates feedback before acting on it
- Commits fixes and waits for CI to complete
- Repeats until all checks pass
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds a skill to identify and remove common AI code patterns like
unnecessary comments, defensive checks, any casts, and inline imports.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Daniel Szoke