The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and follows instructions from untrusted files within the repository (e.g., CONTRIBUTING.md, AGENTS.md) and external tracker data. An attacker could potentially manipulate these sources to influence the agent's behavior during pull request and commit message generation. Ingestion points: CONTRIBUTING.md, .github/CONTRIBUTING.md, AGENTS.md, and ticket metadata via gh issue view. Boundary markers: No explicit delimiters or instructions to ignore embedded commands were identified. Capability inventory: The skill can perform git merge, git push, gh pr create, and gh issue close. Sanitization: No explicit sanitization or validation of the ingested text is mentioned.
[COMMAND_EXECUTION]: The skill uses git and the gh CLI to execute multiple shell commands for managing repository states and interactions with GitHub. Evidence: Commands like git checkout, git merge, git push, gh pr create, gh issue view, and gh issue close are defined in SKILL.md and references/trackers/github.md.

finish