web-design-guidelines
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- EXTERNAL_DOWNLOADS (LOW): The skill fetches a markdown file containing guidelines from a remote URL (
raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.md). Since the source organization (vercel-labs) is on the trusted list, the severity is downgraded to LOW. - INDIRECT_PROMPT_INJECTION (LOW): (Category 8) The skill reads user-specified code files and applies remote rules to them.
- Ingestion points: Specified files (SKILL.md, line 12).
- Boundary markers: None specified in the prompt template.
- Capability inventory: File reading and remote fetching via WebFetch. No subprocess or shell execution capabilities were detected.
- Sanitization: No explicit sanitization or escaping of file content before processing.
- DATA_EXPOSURE (SAFE): The skill reads files for auditing purposes as part of its primary function. There is no evidence of accessing sensitive paths or exfiltrating data to untrusted domains.
Audit Metadata