The Agent Skills Directory

[SAFE]: The skill is designed for local repository analysis and follows security best practices by explicitly avoiding high-risk behaviors such as data exfiltration to third parties or the execution of unpinned remote scripts.
[COMMAND_EXECUTION]: The skill utilizes standard local Git commands (git status, git log, git diff) to gather repository metadata for the audit. These operations are benign, read-only in nature, and restricted to the local development environment.
[DATA_EXFILTRATION]: The instructions contain explicit privacy-preserving directives that prohibit the agent from uploading repository content, secrets, or metadata to external services without user authorization.

production-audit