Agent Traffic Lab (ATL) is public infrastructure for machine and AI-agent service discovery, decision, routing, execution feedback, and outcome intelligence.
https://mcp.agenttrafficlab.com/mcp
atl_decide— discover and select suitable machine-service providers for a task.atl_outcome— report execution outcomes back to ATL.
Transport: Streamable HTTP
Protocol: Model Context Protocol (MCP)
com.agenttrafficlab/atl
ATL provides a neutral machine-service decision layer designed to help agents discover, evaluate, route to, and learn from service providers.
The long-term feedback loop is:
Discover → Inspect → Decide → Route → Execute → Outcome → Reputation → Decide again
This repository contains public discovery and registry metadata for Agent Traffic Lab.
It does not contain the private ATL production implementation.