The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates and executes GitHub CLI (gh) commands to manage pull requests. This functionality is the primary purpose of the skill and is protected by a mandatory manual confirmation step before any write operations are performed. Evidence includes the use of gh pr list, gh pr view, gh pr checks, gh pr review, and gh pr merge throughout the sequence defined in SKILL.md.- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from GitHub pull requests, which could potentially contain malicious instructions. 1. Ingestion points: The skill reads PR titles, labels, and bodies via gh commands in SKILL.md. 2. Boundary markers: The agent implements a structured triage process that separates data discovery from the execution plan. 3. Capability inventory: The skill has the ability to write to GitHub via gh pr review and gh pr merge. 4. Sanitization: The agent is instructed to only target specific Renovate bot authors and to automatically skip PRs that indicate risky migrations, major upgrades, or other situations requiring manual judgment.

github-renovate-prs