configure-canvas-mcp

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [Data Ingestion Surface]: The skill requests a Power Apps studio URL from the user to extract session parameters (environment_id, app_id, environment_category).
  • Ingestion Point: The user input URL is requested in SKILL.md step 1.
  • Boundary Markers: No specific structural boundaries are defined for the text extraction, though explicit formatting expectations for the URL are provided to the user.
  • Capability Inventory: Uses the mcp__canvas-authoring__connect tool and the Bash tool.
  • Sanitization: The skill relies on string and path segment parsing to isolate IDs before passing them to the connection tool.
  • [Local Command Execution]: The skill employs a local command line instruction (dotnet --list-sdks) via the Bash tool to verify whether the required .NET environment is available if the MCP server fails to launch. This is a routine diagnostic check to assist the user with troubleshooting.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 10:15 AM
Security Audit — agent-trust-hub — configure-canvas-mcp