The Agent Skills Directory

[SAFE]: No malicious patterns or security risks detected. The skill is entirely focused on generating markdown documentation within a project repository and follows a structured logical workflow.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input during the product clarification workflow (ingestion points: SKILL.md). This input is interpolated into the final PRD output (capability: local file writing). The skill employs boundary markers by using a specific markdown structure (PRD Document Contract) to delimit data. While it lacks automated sanitization, it effectively mitigates risk by requiring explicit user confirmation for all 'Material Product Decisions' before they are considered canonical, ensuring the agent does not autonomously act on potentially malicious instructions embedded in user data.

clear-prd