The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes grep to search the local filesystem for specific comment markers and git blame to identify authors. These tools are used within their standard parameters for codebase analysis and reporting.
[PROMPT_INJECTION]: The skill processes untrusted data from source code comments which could theoretically contain malicious content designed to influence the agent. 1. Ingestion points: Local source code files parsed by grep. 2. Boundary markers: The skill does not define delimiters to isolate the extracted comment content. 3. Capability inventory: The agent is limited to local read and metadata retrieval operations (grep, git blame). 4. Sanitization: No sanitization of the comment content is performed before it is formatted into the report.

honey-debt