web-design-guidelines

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS] (LOW): The skill performs a GET request to 'https://raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.md' to retrieve guidelines. Because 'vercel-labs' is a trusted organization, the severity is downgraded per security protocols.
  • [REMOTE_CODE_EXECUTION] (MEDIUM): The skill implements remote instruction fetching. It retrieves a file that contains 'all the rules and output format instructions' and executes those instructions against local files. This effectively allows the remote file to control the agent's logic.
  • [PROMPT_INJECTION] (MEDIUM): Category 8 (Indirect Prompt Injection). The skill is vulnerable to instructions embedded in the external source it fetches.
  • Ingestion points: Remote URL fetch via WebFetch.
  • Boundary markers: Absent. The agent is instructed to 'Apply all rules from the fetched guidelines' without any validation or delimiters.
  • Capability inventory: The skill has the capability to read any local files or patterns provided by the user.
  • Sanitization: None. The skill assumes the remote content is safe and authoritative instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 10:17 AM