canvas-design
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied input to derive a design philosophy and a subtle conceptual thread which are then used to generate documents.
- Ingestion points: User-provided creative prompts used as a foundation for the design manifesto.
- Boundary markers: None explicitly defined to separate user input from internal instruction logic.
- Capability inventory: Creation of markdown, PDF, and PNG files; searching local font directories.
- Sanitization: No specific filtering or escaping mechanisms are described in the instructions.
- [EXTERNAL_DOWNLOADS]: Instructions encourage the agent to obtain fonts to fulfill design requirements.
- Evidence: "Download and use whatever fonts are needed to make this a reality."
- Context: Included license files reference well-known and trusted organizations such as Google, Vercel, and IBM, indicating a reliance on reputable sources.
- [NO_CODE]: The skill consists entirely of markdown instructions and license text files. No executable scripts, binaries, or configuration files are present in the provided skill package.
Audit Metadata