The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill instructions focus on creative generation, aesthetic consistency, and the use of locally stored font assets.
[SAFE]: The skill uses local resources from the ./canvas-fonts directory to enhance visual output, which is a standard practice for design-focused tools.
[SAFE]: The mandatory evidence for Indirect Prompt Injection (Category 8) shows a very low risk profile: Ingestion points are limited to user creative prompts; Boundary markers are present via the two-step workflow (Philosophy then Expression); Capability inventory includes only file creation (.md, .pdf, .png) and font reading; Sanitization is handled by the model's inherent safety guidelines and the instruction to avoid copyright violations.

canvas-design