The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the gh (GitHub CLI) and git binaries to perform repository operations, including fetching PR metadata, downloading diffs, and making authenticated API calls to GitHub for posting reviews and comments.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks because it processes untrusted inputs from GitHub pull requests (such as PR descriptions and code comments) to guide its analysis and response generation. Ingestion points: The agent reads external data through gh pr view (PR body) and gh pr diff (code changes) in Phase 1 and Phase 2. Boundary markers: It uses specific parsing rules for author focus items and maintains an exclusion list of existing comments to manage context, though these are not cryptographic boundaries. Capability inventory: The skill can execute shell commands via gh and git, write to the local filesystem, and invoke other agent skills for specialized tasks. Sanitization: The instructions mandate filtering out non-review boilerplate from PR descriptions, but the skill lacks explicit sanitization for code content processed during deep investigation phases.

wk-pr-review