The Agent Skills Directory

[SAFE]: The skill performs standard development tasks such as running git diff, git log, and git rev-parse. These are read-only operations within the repository context.
[SAFE]: The process for identifying coding standards involves searching local files like CODING_STANDARDS.md and CONTRIBUTING.md. This is a best-practice behavior for code auditing agents.
[SAFE]: The 'smell baseline' included in the prompt is a static list of well-known software engineering patterns from industry-standard literature (Fowler's Refactoring). It does not contain executable code or malicious instructions.
[SAFE]: The skill uses sub-agents to parallelize review tasks. This is an architectural choice for context separation and does not introduce security vulnerabilities or privilege escalation risks.

code-review