algorithmic-art

Warn

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADSSAFE
Full Analysis
  • [PROMPT_INJECTION]: The metadata in SKILL.md identifies the author as 'Anthropic', which is deceptive as the author context is 'zocomputer'.
  • [EXTERNAL_DOWNLOADS]: The skill fetches the p5.js library from Cloudflare's CDN (cdnjs.cloudflare.com), which is a trusted source.
  • [PROMPT_INJECTION]: The skill processes user-provided inputs to generate generative art. 1. Ingestion points: User requests for art themes in SKILL.md. 2. Boundary markers: Instructions specify using user input as a foundation without constraining creativity. 3. Capability inventory: The skill generates and renders HTML/JavaScript artifacts. 4. Sanitization: No explicit validation or escaping of user-provided conceptual seeds is defined before code generation.
  • [SAFE]: No other malicious code or persistence patterns were detected.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 13, 2026, 02:09 PM