The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub pull requests (diffs, titles, bodies, and existing comments) which could contain adversarial instructions intended to influence the agent's behavior or review outcome.
Ingestion points: The preflight stage uses gh pr view and gh pr diff to ingest PR metadata, comments, and code changes into the agent's context.
Boundary markers: The instructions do not define clear delimiters or use "ignore embedded instructions" warnings for the external data being processed.
Capability inventory: The skill possesses the ability to post comments back to the GitHub repository using gh api in the post stage.
Sanitization: The skill lacks evidence of sanitizing or escaping content from the PR before it is interpolated into the prompts for review stages like standards, spec, and bug-detector.
[COMMAND_EXECUTION]: The skill relies on the GitHub CLI (gh) to perform its primary functions. It executes commands such as gh pr view, gh pr diff, and gh api to read repository data and write review comments. This behavior is consistent with the skill's stated purpose of reviewing pull requests.

pr-reviewer