write-a-prd

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill contains no malicious code, hidden instructions, or obfuscated content.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external repository data and user input. \n- Ingestion points: User-provided problem descriptions (SKILL.md, step 1) and repository content exploration (SKILL.md, step 2). \n- Boundary markers: None present. \n- Capability inventory: Reading repository files and writing GitHub issues via integrated tools. \n- Sanitization: None present. This surface is considered safe given the skill's limited capabilities and its primary function of text generation and documentation.- [DATA_EXFILTRATION]: The skill accesses codebase information and transmits user-approved PRD content to GitHub issues. This behavior aligns with the skill's stated purpose and does not involve unauthorized external communication.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 10:50 PM