skills/idaibin/skills/ui-spec/Gen Agent Trust Hub

ui-spec

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the @google/design.md package at pinned version 0.4.0 from the official npm registry for design contract validation.
  • [COMMAND_EXECUTION]: The validation script executes npm install and node via subprocess to run the design linter. The commands use fixed arguments and are not exposed to shell injection.
  • [DYNAMIC_EXECUTION]: Generates and executes a temporary JavaScript file (parse.mjs) to act as a bridge between the Python validation logic and the Node.js linter library.
  • [INDIRECT_PROMPT_INJECTION]: Processes external design documents and JSON evidence. (1) Ingestion points: Markdown DESIGN.md files and JSON visual evidence artifacts. (2) Boundary markers: Enforces a strict H2 section structure in DESIGN.md and uses JSON Schema validation for evidence files. (3) Capability inventory: Executes external linting tools via subprocess.run in scripts/validate-design-md-completeness.py. (4) Sanitization: Validates file structure and content against predefined schemas and canonical headings before use.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 07:41 AM
Security Audit — agent-trust-hub — ui-spec