The Agent Skills Directory

[SAFE]: The skill instructions are focused entirely on visual design guidelines for creating a specific UI effect using procedural canvas rendering.
[PROMPT_INJECTION]: No attempts to bypass safety filters or override agent instructions were found. The language used is descriptive of the desired visual output.
[DATA_EXFILTRATION]: The skill and its demo do not include any network requests or code that accesses sensitive system information, credentials, or environment variables.
[REMOTE_CODE_EXECUTION]: The demo implementation uses standard, inline JavaScript for canvas animation. There are no external script downloads or execution of remote payloads.
[DYNAMIC_EXECUTION]: The animation logic uses standard requestAnimationFrame and sine-wave mathematics. No eval(), exec(), or other dangerous dynamic execution patterns are present.

atmosphere-background