web-design-guidelines
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (LOW): The skill fetches dynamic instructions from a remote GitHub URL. While loading remote instructions is generally risky, the source belongs to a trusted organization.
- Evidence: Fetches from
https://raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.mdusingWebFetch. - Context: The repository
vercel-labs/web-interface-guidelinesis listed as a trusted source, which downgrades the severity per [TRUST-SCOPE-RULE]. - PROMPT_INJECTION (LOW): The skill is susceptible to Indirect Prompt Injection (Category 8) because it processes untrusted UI code provided by users.
- Ingestion points: Reads local files specified by the user or patterns.
- Boundary markers: Absent; the skill does not use delimiters to isolate the code being reviewed from the audit instructions.
- Capability inventory: Restricted to reading local files and fetching the guideline URL. No file-write, command execution, or sensitive data exfiltration capabilities were detected.
- Sanitization: No evidence of sanitizing or escaping content within the provided UI code before processing.
Audit Metadata