The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing shell commands such as git, gh (GitHub CLI), and rg (ripgrep) to analyze code changes. These commands are used to fetch repository history, diffs, and PR metadata.
[DATA_EXPOSURE]: By design, the skill accesses and reads the entire contents of a git repository and external PR information from GitHub to perform its analysis. This involves reading source code, commit messages, and PR descriptions.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of GitHub PR content (titles, bodies, and code comments). There is a risk that a maliciously crafted PR could attempt to inject instructions to influence the agent's review. However, the skill provides explicit instructions to the agent to remain in a read-only mode and requires explicit user consent before performing any GitHub write operations (like submitting a review or pushing code).

pr-change-analysis