pencil-design-from-stitch-html

Pass

Audited by Gen Agent Trust Hub on Apr 2, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted HTML data from external URLs or user input to drive automated design generation.
  • Ingestion points: HTML code is fetched from htmlCode.downloadUrl (provided via Stitch MCP tools) or pasted directly by the user, as outlined in the Retrieval and HTML section of SKILL.md.
  • Boundary markers: The instructions lack explicit delimiters or instructions to ignore potential commands embedded within the HTML content being processed.
  • Capability inventory: The skill has access to powerful tools including mcp_pencil:batch_design for modifying design documents, Bash for system operations, and Write for local file storage.
  • Sanitization: No instructions are provided for sanitizing or validating the DOM tree or Tailwind classes before they are used to generate Pencil design operations.
  • [EXTERNAL_DOWNLOADS]: Fetches HTML source code and visual assets from URLs provided by the Stitch MCP server (e.g., stitch.withgoogle.com) to support the conversion process.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 2, 2026, 06:28 AM
Security Audit — agent-trust-hub — pencil-design-from-stitch-html