The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes functionality to analyze external project code, which creates a potential surface for indirect prompt injection.
Ingestion points: User-provided source code files, repository contents, and screenshots are processed during UI audits and reviews.
Boundary markers: The skill relies on natural language instructions rather than formal delimiters to separate analyzed content from the agent's core instructions.
Capability inventory: The agent maintains its standard capabilities, including reading and writing files and executing terminal commands, while performing design tasks.
Sanitization: The instructions do not specify sanitization or escaping of the ingested code prior to interpolation into the agent's context.
[EXTERNAL_DOWNLOADS]: The skill links to official documentation and trusted industry resources for design best practices, including domains such as web.dev, developer.mozilla.org, and various official library documentation sites.
[SAFE]: The skill's primary purpose is to provide UI/UX design guidance. All provided code snippets and external references align with this purpose and do not exhibit malicious behaviors.

critly