The Agent Skills Directory

[SAFE]: The skill uses established command-line tools (git, gh, vercel) to manage repositories and analyze code. Its operations are transparent and consistent with its stated purpose as a development utility.\n- [PROMPT_INJECTION]: The skill inherently processes untrusted data as part of its code review function, creating an indirect prompt injection surface.\n
Ingestion points: The agent reads untrusted content from git diff output and GitHub issues/PRs using gh issue list and gh pr list in SKILL.md.\n
Boundary markers: There are no explicit delimiters instructed for use to separate untrusted diff text from the agent's core instructions.\n
Capability inventory: The skill possesses write access to GitHub (gh issue comment), the ability to modify local files via autofixes, and the capacity to execute shell scripts (scripts/run-tests.sh).\n
Sanitization: Input data is not sanitized or escaped before processing, which is a common characteristic of code review automation. This is documented as an inherent risk rather than a malicious defect.

check