The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several git commands to manage the review process, including git merge-base, git rev-parse, git log, and git diff. These commands are used to identify the range of changes (SHAs) and retrieve the code content for analysis. These operations are essential to the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill acts on external data by reading git diffs and commit messages. This creates a potential surface where malicious instructions could be embedded in code comments or commit history to influence the agent's review.
Ingestion points: Content retrieved via git diff and git log (SKILL.md).
Boundary markers: The instructions mandate a structured "Review Findings" template and "Review Request" brief, which helps the agent distinguish between data being reviewed and its own operating instructions.
Capability inventory: The skill is primarily focused on analysis and reporting, with no dangerous file-writing or arbitrary network capabilities identified.
Sanitization: The skill relies on the agent's internal reasoning and the structured reporting template to maintain context boundaries.

code-review