design-md

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill retrieves HTML source code and visual screenshots from download URLs provided by the Stitch MCP server. These downloads are used for static analysis of design tokens, Tailwind classes, and layout structures.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for indirect instructions through the processing of untrusted project assets (HTML and metadata).
  • Ingestion points: Retrieval of external HTML code via web_fetch or read_url_content and screen metadata through the Stitch MCP get_screen tool.
  • Boundary markers: Absent. The skill does not define specific delimiters or instructions to ignore potential commands embedded within the design files.
  • Capability inventory: Includes listing projects/screens, fetching detailed screen metadata, downloading remote code assets, and writing content to a local DESIGN.md file.
  • Sanitization: Absent. The instructions do not mandate the filtering or sanitization of the retrieved HTML content prior to analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:18 PM
Security Audit — agent-trust-hub — design-md