agentfence

Policy firewall for AI agents and MCP tools. Intercepts tool calls, enforces allow/deny/ask rules, redacts secrets, validates arguments, and writes auditable logs before agents touch your filesystem, GitHub, browser, or APIs.