The Agent Skills Directory

[SAFE]: The skill's primary function is to analyze images and append structured design data to markdown files. It follows established best practices for workspace path resolution and uses the browser tool as configured in its metadata. No malicious patterns or security violations were detected across the analyzed files.
[EXTERNAL_DOWNLOADS]: The REFERENCES.md file contains links to technical documentation from well-known services and organizations, including the World Wide Web Consortium (W3C), Mozilla (MDN Web Docs), Google Fonts, and WebAIM. These references are used to ensure correct CSS syntax and compliance with accessibility standards.
[PROMPT_INJECTION]: The skill identifies a potential indirect prompt injection surface by processing untrusted images or screenshots. However, it incorporates a mitigation strategy by requiring the agent to output only concrete, implementable design values (e.g., hex codes, pixel measurements) rather than qualitative descriptions. This constraint limits the ability of an attacker to influence the skill's output through adversarial content embedded in images.

image-analysis