organize
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided information architecture data (such as product names, category labels, and site structures) and interpolates this content into generated HTML templates and design tool command parameters.
- Ingestion points: User input describing navigation structures, site maps, and taxonomies in the context of the
/organizecommand. - Boundary markers: The skill specifies rigid structural templates for output and defines constraints (e.g., "no external JS"), but it does not provide explicit boundary markers or instructions to ignore potential commands embedded within the user-provided labels.
- Capability inventory: The skill generates self-contained HTML/CSS code blocks and interacts with Figma and pencil.dev via platform MCP tools (
mcp__claude_ai_Figma__use_figma,mcp__pencil__open_document,mcp__pencil__set_variables,mcp__pencil__batch_design). - Sanitization: There are no instructions or patterns for sanitizing, escaping, or validating user-provided strings before they are rendered in HTML or sent to design APIs.
Audit Metadata