The Agent Skills Directory

[PROMPT_INJECTION]: The instructions utilize a 'user simulation' technique in the 'FINAL STEP' section, where it provides a pre-written user complaint ('It isn't perfect enough...') to force the agent to perform additional refinement. This overrides the model's objective assessment of the task completion based on the actual user's input.
[NO_CODE]: The skill is composed solely of Markdown documentation and font license agreements. There are no Python scripts, Node.js modules, or shell commands included, which eliminates the possibility of technical vulnerabilities like remote code execution or privilege escalation.
[PROMPT_INJECTION]: The skill processes conceptual input from the user to deduce a 'subtle conceptual thread' for art creation. The instructions lack boundary markers (like delimiters) or sanitization steps for this ingested data, making the creative process susceptible to indirect prompt injection where instructions hidden in the user's thematic request could steer the agent's behavior.

canvas-design