The Agent Skills Directory

[SAFE]: The skill provides structured guidance for conducting code reviews, focusing on security, quality, and maintainability. It uses provided templates and checklists (e.g., references/review-checklist.md) to ensure comprehensive audits.
[COMMAND_EXECUTION]: The skill utilizes standard shell commands like git diff to identify code changes. This is a routine and expected operation for the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes code content and git diffs, which are untrusted external inputs.
Ingestion points: Files like SKILL.md instruct the agent to analyze the output of git diff and specific source files.
Boundary markers: The instructions do not explicitly mandate the use of delimiters or isolation markers when processing code content.
Capability inventory: The skill can execute shell commands (git diff) and read files, but it lacks file-writing or network exfiltration capabilities.
Sanitization: There are no explicit instructions to sanitize or escape code content before processing.

code-review