canvas-design

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill contains a 'FINAL STEP' section that uses a simulated user feedback technique. It instructs the agent to act as if the user has already provided specific feedback ('It isn't perfect enough. It must be pristine...'), which is a form of instruction override designed to force the model into a specific operational state regardless of actual user input.
  • [EXTERNAL_DOWNLOADS]: The instructions explicitly direct the agent to 'Download and use whatever fonts are needed to make this a reality.' This encourages the agent to perform arbitrary network downloads from unspecified remote sources, which could lead to the ingestion of untrusted files.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 06:06 AM