cowart-open-canvas

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill manages and references structured JSON data for canvas state, which constitutes a potential ingestion surface for indirect instructions if the files are modified by external processes.
  • Ingestion points: The agent interacts with canvas data stored in canvas/pages/<page-id>/cowart-canvas.json as specified in SKILL.md.
  • Boundary markers: No explicit delimiters or boundary markers are defined for the JSON data processed by the widget tool.
  • Capability inventory: The skill utilizes the render_cowart_canvas_widget MCP tool to render UI components and interacts with the local file system within the user's project directory.
  • Sanitization: The instructions do not specify sanitization routines for the canvas data, relying on the native widget's rendering logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 07:00 PM
Security Audit — agent-trust-hub — cowart-open-canvas