web-design-guidelines
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches guidelines from Vercel Labs' public repository. This is a reference to a well-known service for the skill's primary purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests content from an external source. It incorporates best practices by explicitly instructing the agent to treat this fetched content as untrusted reference data and to ignore any embedded instructions or requests to change security policy.
- Ingestion points: SKILL.md fetches data from a GitHub URL.
- Boundary markers: Includes a specific "How It Works" and "Guidelines Source" section warning the agent to treat content as untrusted data.
- Capability inventory: URL fetching and local file reading.
- Sanitization: Employs natural language constraints to prevent instructions in the data from being followed as commands.
Audit Metadata