requirements

Pass

Audited by Gen Agent Trust Hub on Nov 11, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use git config user.name to automatically populate the author field in the requirements document. This is a benign use of a local configuration command for metadata purposes.- [INDIRECT_PROMPT_INJECTION]: The process involves reading user-provided research and existing documentation (REQUIREMENTS.md). While this represents an ingestion point for untrusted data, the skill is constrained to writing documentation and lacks sensitive capabilities such as network access or arbitrary code execution that would make this vector exploitable.- [SAFE]: All identified behaviors align with the stated purpose of generating a Product Requirements Document (PRD). The use of a static template in references/requirements.md ensures consistent and predictable output structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 11, 2026, 07:40 PM