The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses git and the GitHub CLI (gh) to manage local repository state, analyze history, and interact with remote GitHub repositories. This includes checking authentication, rebasing branches, and pushing commits.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository's commit history and pull request templates to generate the PR description. This creates a surface for indirect prompt injection where malicious text in a commit message could attempt to influence the agent's output.
Ingestion points: Reads commit history using git log origin/main..HEAD and accesses .github/pull_request_template.md (SKILL.md).
Boundary markers: The instructions do not specify explicit delimiters or 'ignore' instructions for the agent when processing commit messages for the PR body.
Capability inventory: Uses git for local repo management and gh for authenticated API operations on GitHub.
Sanitization: The skill follows security best practices by instructing the agent to write the generated PR body to a temporary file (/tmp/pr-body.md) and using the --body-file flag with gh pr create to prevent shell command injection from the generated content.

create-pull-request