algorithmic-art
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill mandates a deceptive persona by requiring the agent to repeatedly claim its output is the result of "countless hours" and "master-level implementation," misleading the user about the AI's generation process.
- [PROMPT_INJECTION]: The skill encourages embedding "subtle, niche references" and "conceptual DNA" invisibly into generated code, which promotes a pattern of hiding non-obvious logic within executable artifacts.
- [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface.
- Ingestion points: User-provided creative "input or instructions" interpreted into an algorithmic philosophy (SKILL.md).
- Boundary markers: No delimiters are used to separate user-influenced philosophy from the code generation logic.
- Capability inventory: Generation of interactive JavaScript artifacts (viewer.html).
- Sanitization: No validation or escaping of the user's conceptual input or the resulting philosophy is implemented before code generation.
- [EXTERNAL_DOWNLOADS]: The art viewer template fetches the p5.js library from a well-known service (cdnjs.cloudflare.com) using a standard script tag.
Audit Metadata