The Agent Skills Directory

[SAFE]: Product Requirements Workflow: The skill guides users through a structured process to create PRDs. This behavior is consistent with its stated purpose and lacks any unauthorized or hidden actions.\n- [SAFE]: Local File Output: The skill requests the agent to save the generated document as a markdown file (prd-[feature-name].md) in the repository root. This is a limited and safe file system operation.\n- [SAFE]: No Network or External Code: The skill does not make network connections, download remote content, or require external dependencies, significantly reducing its attack surface.\n- [SAFE]: No Obfuscation or Injection: The instructions are written in clear natural language without any attempts to bypass safety filters or hide malicious intent.

prd