The Agent Skills Directory

[SAFE]: The skill's instructions and reference materials promote secure coding practices and do not contain malicious code, patterns, or obfuscation.
[COMMAND_EXECUTION]: The skill executes git commands to resolve the scope of the audit and read code diffs. This is a standard and safe operation within a local development environment, necessary for the skill's primary function.
[PROMPT_INJECTION]: The skill includes specific instructions to analyze code for security defects (such as injection, improper input validation, and hardcoded secrets), which serves to strengthen the security posture of the development workflow rather than compromise it.

playbook