canvas-design
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill utilizes a simulated user quote ("The user ALREADY said...") to override standard behavior and mandate a refinement process.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface. 1. Ingestion points: User conceptual instructions for art creation mentioned in SKILL.md. 2. Boundary markers: Absent; there are no delimiters or warnings to ignore embedded instructions. 3. Capability inventory: Generates and writes .md, .pdf, and .png files to the system. 4. Sanitization: None implemented for external user-supplied data.
- [EXTERNAL_DOWNLOADS]: The skill instructions suggest downloading fonts from external GitHub repositories. While many are from trusted organizations like Google, IBM, and Vercel, others belong to individual developer accounts.
Audit Metadata