The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided natural language to generate React components and layouts.
Ingestion point: User requests for specific UI components, data models, and page structures (SKILL.md).
Boundary markers: None explicitly defined to separate user input from system instructions.
Capability inventory: The skill has access to Write, Edit, and Bash tools, allowing it to modify files and execute system commands.
Sanitization: The skill relies on a 16-point self-validation checklist for design quality and correctness but does not describe any input sanitization or escaping mechanisms.
[COMMAND_EXECUTION]: The skill involves shell operations such as installing the @raydenui/ui package via npm install. While this is a standard developer workflow, the capability to execute bash commands based on skill logic represents a potential attack surface if misdirected by malicious input.

rayden-code