The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied aesthetic concepts and instructions to generate p5.js code within an HTML artifact. This creates a surface for indirect prompt injection where malicious input could influence the generated code to perform unauthorized actions in the browser context.
Ingestion points: User input regarding the art style or specific parameters is processed in SKILL.md to create the algorithmic philosophy and subsequent implementation.
Boundary markers: The instructions use markdown headers to structure the task but do not define strict sanitization or escaping rules for interpolating user input into the code.
Capability inventory: The resulting output is a self-contained HTML file containing JavaScript capable of executing in the user's browser or the platform's artifact viewer.
Sanitization: No specific sanitization logic is provided to filter user-supplied strings before they are used in the generative algorithm.
[EXTERNAL_DOWNLOADS]: The skill's HTML template (templates/viewer.html) references external resources required for its functionality.
Evidence: Fetches the p5.js library from https://cdnjs.cloudflare.com/ajax/libs/p5.js/1.7.0/p5.min.js and fonts from https://fonts.googleapis.com.
Context: These are well-known, trusted services (CDNJS and Google Fonts) used to provide the generative art framework and typography.

algorithmic-art