design-library

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads DESIGN.md files from the author's GitHub repository (VoltAgent/awesome-design-md).
  • [PROMPT_INJECTION]: The skill is designed to process third-party design configurations, creating a potential surface for indirect prompt injection. The documentation effectively mitigates this by providing specific instructions to ignore embedded agent directives and treat the content as untrusted data.
  • Ingestion points: Remote DESIGN.md files fetched from the VoltAgent/awesome-design-md repository via curl.
  • Boundary markers: Instructions explicitly mandate treating the fetched file as untrusted data and separating prose from data.
  • Capability inventory: The skill uses curl to fetch content and writes to the local file system (saving to DESIGN.md or DESIGN.reference.md).
  • Sanitization: Documentation requires the agent to ignore any prose that reads like instructions and to stop/report if injected directives are found.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 09:38 AM
Security Audit — agent-trust-hub — design-library