web-design-guidelines
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches design guidelines from Vercel Labs' official GitHub repository (raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.md). This is a well-known service and the reference is handled neutrally.\n- [COMMAND_EXECUTION]: Instructs the agent to use
curlto retrieve the latest guidelines if standard web tools are unavailable, ensuring the most recent rules are applied during the review.\n- [INDIRECT_PROMPT_INJECTION]: The skill reviews user-provided UI source code, which serves as a vector for untrusted data ingestion.\n - Ingestion points: User-specified files or file patterns passed as arguments for UI review (identified in SKILL.md and references/command.md).\n
- Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish between guideline rules and instructions potentially embedded in the user code.\n
- Capability inventory: File reading capabilities for analysis and network access (curl) for fetching external guidelines.\n
- Sanitization: No mechanisms are defined to sanitize or filter the content of the UI files before analysis.\n- [METADATA_POISONING]: The skill's metadata lists 'vercel' as the author, which differs from the provided author information for this skill.
Audit Metadata