The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and review user-supplied generative art code, which establishes a potential surface for indirect prompt injection if malicious instructions are embedded in the data being analyzed.
Ingestion points: User-provided art code and requests for diagnosis/review as instructed in SKILL.md.
Boundary markers: Absent; there are no explicit delimiters to isolate user-provided code from the agent's instructions.
Capability inventory: The skill provides templates for browser-based JavaScript (p5.js) and fragment shaders (GLSL). No sensitive file system, network exfiltration, or shell execution capabilities were identified.
Sanitization: None identified; user input is processed based on the guidelines in references/validations.md without specific sanitization for prompt injection patterns.
[SAFE]: All provided code examples in references/patterns.md and references/sharp_edges.md utilize standard, non-malicious creative coding libraries and patterns. Mentions of external libraries like ethers and opentype.js in comments are for educational purposes and do not involve automated installation or execution.

generative-art