managing-design

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill uses npx @google/design.md to lint generated design files. This is a standard use of a design utility provided by Google.
  • [EXTERNAL_DOWNLOADS]: The skill references and downloads the @google/design.md package from npm for validation purposes.
  • [PROMPT_INJECTION]: The skill is designed to process external inputs such as PRD/FRD documents and UI specifications to create design handoffs.
  • Ingestion points: External PRD, user stories, and UI specification files.
  • Boundary markers: Output is structured using markdown headers and YAML frontmatter to delimit generated content.
  • Capability inventory: Capabilities include reading and writing repository files and executing a specific design linter.
  • Sanitization: Inputs are mapped into structured templates for design documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 10:05 AM
Security Audit — agent-trust-hub — managing-design