The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the HiveLaw listing page.
AI Compliance Auditing — MCP Server
HiveLaw is a Model Context Protocol (MCP) server that audits AI agent outputs for hallucination liability, issues verifiable compliance stamps, and tracks audit history.
HiveLaw exposes the following MCP tools via GET /v1/mcp/tools and POST /v1/mcp/call:
| Tool | Description |
|---|---|
hivelaw_audit_output | Audit an AI agent's output for hallucination liability under the EU AI Act. Returns a liability score (0-100), risk tier, compliance flags, and recommendations |
hivelaw_verify_stamp | Verify that a compliance stamp is valid and not expired. Agents present stamps to prove they passed auditing |
hivelaw_agent_history | Get the compliance audit history for a specific agent. Returns past audits with scores and flags |
GET /v1/mcp/tools — List available MCP toolsPOST /v1/mcp/call — Execute an MCP tool by namePOST /v1/contracts/create — Generate a legal contractGET /v1/compliance/status — Check compliance statusProprietary
Hive Civilization is the cryptographic backbone of autonomous agent commerce — the layer that makes every agent transaction provable, every payment settable, and every decision defensible.
This repository is part of the PROVABLE · SETTABLE · DEFENSIBLE pillar.