mission-control

Why we included this project

Running more than a couple of AI agents across different tools quickly raises questions about task ownership, what actually executed, and where failures and spending landed. Mission Control answers those from a self-hosted dashboard that sits above runtimes such as OpenClaw, Claude Code, and Codex without interfering with their reasoning or tool loops. It offers a task board with assignment and review stages, an agent registry that tracks presence and heartbeats, live activity streams, and cost and token views, all stored in SQLite so your data stays on your own infrastructure. You can also drive it headless via a REST API, CLI, or MCP server, which makes it straightforward to integrate with existing agent tooling. The project is alpha software, so expect schema and API changes, but teams that have outgrown CLI visibility will find it a practical way to get a single window into a multi-agent fleet.