web-design-guidelines
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (LOW): The skill is configured to fetch guidelines from 'https://raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.md'. While this involves a remote download, the source belongs to a trusted organization (vercel-labs), which downgrades the risk.
- PROMPT_INJECTION (LOW): The skill exhibits an Indirect Prompt Injection surface by fetching markdown content from a remote URL and instructing the agent to follow the rules contained within that content.
- Ingestion points: Remote guideline URL specified in 'SKILL.md'.
- Boundary markers: Absent; the agent is simply told to 'Check against all rules in the fetched guidelines'.
- Capability inventory: File system read operations (reading user-specified UI code files).
- Sanitization: Absent; the fetched instructions are directly incorporated into the agent's task logic.
Audit Metadata