style-guide-builder

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill analyzes reference images provided by the user (SKILL.md). Ingestion point: Reference images. Capability inventory: File-writing to project directories and Vision model invocation. Boundary markers: Explicit stage 0 questions for purpose, audience, and exclusions (audience.md). Sanitization: The visual analysis process focuses on stylistic consistency rather than textual content.
  • [COMMAND_EXECUTION]: The skill documentation refers to the optional use of the 'gen-ai' command-line interface and Picsart's MCP servers (mcp.picsart.io) for vision-based analysis. These are recognized tools within the vendor's ecosystem used for image analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 10:59 PM
Security Audit — agent-trust-hub — style-guide-builder