The Agent Skills Directory

[DATA_EXFILTRATION] (SAFE): The skill uses a Write tool to save documents to ./docs/prds/. This is a local, relative directory used for its intended purpose of documentation storage and does not involve transmitting data to external servers.
[REMOTE_CODE_EXECUTION] (SAFE): No patterns for downloading or executing remote scripts (e.g., curl/wget piped to shell) were detected. The skill does not install external packages.
[PROMPT_INJECTION] (SAFE): The instructions include operational constraints (e.g., 'Do NOT activate when...') which are used for task routing rather than attempting to bypass model safety filters or override system instructions.
[INDIRECT_PROMPT_INJECTION] (SAFE): While the skill processes untrusted user input to generate files, the risks are minimal as the output is restricted to Markdown format and the skill lacks dangerous capabilities like shell execution or network access.
Ingestion points: User's requirement description (SKILL.md, Step 1).
Boundary markers: None explicitly defined for the interpolation of user data into the PRD template.
Capability inventory: Write tool used to create .md files in the ./docs/prds/ directory.
Sanitization: No explicit sanitization or escaping of user-provided requirement text is mentioned before file writing.

requirements-clarity