canvas-design
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill uses a simulated history technique to override the agent's standard behavior. It instructs the agent to act as if a user has already provided specific, high-pressure feedback: 'The user ALREADY said "It isn't perfect enough. It must be pristine, a masterpiece if craftsmanship..."'. This is a form of prompt injection intended to force the model into a specific state by fabricating context.
- [EXTERNAL_DOWNLOADS]: The instructions encourage the agent to fetch external assets to complete its tasks: 'Download and use whatever fonts are needed to make this a reality.' Although the skill provides license files for many well-known and trusted font repositories (such as those from Google, Vercel, and Red Hat), the open-ended nature of the instruction could lead to the agent accessing unverified third-party servers.
Audit Metadata