The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted external data in the form of code review feedback, PR comments, and reviewer notes.
Ingestion points: Feedback provided by the user or external reviewers is parsed in the "Parse Reviewer Notes" step in SKILL.md.
Boundary markers: The instructions lack specific boundary markers or delimiters to isolate the external feedback from the agent's system instructions.
Capability inventory: The agent has capabilities to search (Grep, Glob), modify files (Edit), create files (Write), and execute commands (uv run ruff check).
Sanitization: There are no explicit instructions to sanitize or validate the content of the review notes before creating actionable tasks or implementing changes.
[COMMAND_EXECUTION]: The skill instructs the agent to run uv run ruff check for linting. This is a standard and well-known development practice for code quality and is considered safe within this context.

review-implementing