web-design-guidelines
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (LOW): Fetches guidelines from raw.githubusercontent.com/vercel-labs/web-interface-guidelines. This is a trusted organization; the download finding is downgraded per [TRUST-SCOPE-RULE].
- [PROMPT_INJECTION] (HIGH): The skill is susceptible to Indirect Prompt Injection (Category 8). It dynamically fetches natural language instructions that the agent is commanded to follow. Evidence Chain: (1) Ingestion point: remote command.md file. (2) Boundary markers: Absent. (3) Capability inventory: Network access via WebFetch and local file system read access. (4) Sanitization: Absent.
Recommendations
- AI detected serious security threats
Audit Metadata