web-design-guidelines

Warn

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [METADATA_POISONING]: The skill metadata identifies the author as "vercel," while the platform attributes the skill to "Mrziadd0." This deceptive metadata could mislead users into trusting the skill based on a false association with a reputable organization.
  • [EXTERNAL_DOWNLOADS]: The skill fetches configuration and rules from the vercel-labs GitHub repository, which is an official repository of a well-known service provider.
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves external instructions that are used to analyze user code, creating a potential surface for indirect injection if the external source were compromised.
  • Ingestion points: Fetches content from https://raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.md at runtime.
  • Boundary markers: No explicit delimiters are used to isolate the fetched guidelines from the agent's core instructions.
  • Capability inventory: The skill has permissions to read local files and generate reports based on the fetched rules.
  • Sanitization: No validation or sanitization of the fetched guidelines is performed before they are integrated into the review process.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 21, 2026, 10:29 PM
Security Audit — agent-trust-hub — web-design-guidelines