The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands to verify the presence of and interact with local analysis tools such as rg, ccc, ast-grep, semgrep, and graphify. This is the intended behavior for discovering and using existing search infrastructure on the user's machine.
[PROMPT_INJECTION]: The skill ingests and processes the content of local source files, which are treated as untrusted data. Although it does not implement explicit sanitization or boundary markers for the data returned to the user, the skill contains strict operating constraints to never fetch external URLs and to limit its scope entirely to the local codebase.

codebase-auditor