Building a Governance MCP Server for AI Agents

MCP (Model Context Protocol) lets AI assistants use external tools. But when those tools make real decisions, you need governance. Asqav MCP Server pip install asqav-mcp export ASQAV_API_KEY = "sk_live_..." asqav-mcp Tools Tool What it does check_policy Check if an action is allowed sign_action Create a signed audit record list_agents List registered agents get_agent Get agent details verify_signature Verify a previous signature Claude Desktop setup Add to claude_desktop_config.json: { "mcpServers" : { "asqav" : { "command" : "asqav-mcp" , "env" : { "ASQAV_API_KEY" : "sk_live_..." } } } } Claude Code setup claude mcp add asqav -- asqav-mcp Now Claude checks policies before acting and creates audit records automatically. Links asqav-mcp on PyPI GitHub Asqav SDK