The Agent Skills Directory

[SAFE]: The skill performs document processing, parsing raw text specifications into structured markdown files. It does not execute code or access network resources.
[DATA_EXPOSURE]: The skill reads project-specific files (test specifications and context) and writes to a designated documentation directory. It does not access sensitive system paths or credentials.
[PROMPT_INJECTION]: While the skill processes untrusted user input (plain-English test cases), its instructions are focused on verbatim extraction and structural mapping. It does not have access to tools or execution environments where instructions in the source text could trigger malicious actions.

ingest-spec