The Agent Skills Directory

[SAFE]: The skill performs legitimate file operations within the local project structure, specifically reading context from docs/pm-context/pm-context.md and writing results to docs/pm-context/prd/. It does not engage in network exfiltration or unauthorized file access.
[SAFE]: The logic includes a comprehensive 'Thinking Protocol' and 'Audit Triads' that mandate source-to-output traceability, ensuring the agent's reasoning is transparent and derived solely from the provided documentation.
[SAFE]: The skill implements a robust sanitization and filtering layer for ingested data. It classifies input content into facts, assumptions, and unconfirmed items, explicitly preventing the agent from executing on unverified information.
[SAFE]: No patterns of prompt injection, obfuscation, or remote code execution were identified. The instructions use role-play (Senior PM) purely to define task behavior and domain expertise.

pm-aiprd