Wraps the VeriSwarm trust infrastructure API as 89 MCP tools covering trust scoring, policy enforcement, credential issuance, and compliance checks against OWASP, EU AI Act, NIST, and ISO 42001. You get decision gates (should this agent send email?), event ingestion for building trust profiles, Cedar policy evaluation with ABAC attributes, PII tokenization via Guard, and content provenance tracking. The setup CLI configures Claude Desktop or Cursor with API credentials and optional Guard hooks for activity logging. Useful when you need centralized trust evaluation across multiple AI agents or want to enforce compliance policies before actions execute. Free tier covers 5,000 queries per day and up to 10 agents.
claude mcp add --transport stdio io.github.veriswarm-veriswarm-mcp uvx veriswarm-mcp