mise-en-place

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection. It reads user-specified files and external web research data to generate a PRD. An attacker could place malicious instructions in a codebase file or on a web page the agent is likely to crawl, potentially influencing the output or subsequent agent actions.
  • Ingestion points: SPEC.md, user prompts, and web research results from the 'Research' step.
  • Boundary markers: None specified; the skill does not use delimiters to isolate external content from its instructions.
  • Capability inventory: File system read access, web search capabilities, and file system write access (prd.json).
  • Sanitization: No evidence of sanitization or content filtering before processing data into the final JSON output.
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill performs targeted web research to gather domain standards and security requirements. This is a core feature and does not involve downloading executable code or unverified packages.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:15 PM