mission-control: an open-source dashboard for orchestrating AI agent fleets with built-in security auditing and cost tracking

What it solves

Mission Control provides a centralized, open-source dashboard for managing fleets of AI agents. It eliminates the need to stitch together multiple tools by offering a single interface to dispatch tasks, track costs, monitor security telemetry, and coordinate complex multi-agent workflows.

How it works

It operates as a self-hosted SPA (Single Page Application) powered by Next.js and SQLite, requiring no external dependencies like Redis or Postgres. The system uses a set of framework adapters to normalize communication with various agent frameworks (such as CrewAI, LangGraph, and AutoGen). It features a Kanban-style task board for workflow orchestration, a memory knowledge graph for exploring agent data, and a built-in security scanner to vet agent skills before installation.

Who it’s for

It is designed for teams and developers operating multi-agent systems who need predictable orchestration, real-time observability, and a secure, local-first deployment model for their AI agent infrastructure.

Highlights

• Comprehensive Control Center: Includes 32 feature panels covering tasks, logs, tokens, memory, and security.
• Multi-Framework Support: Adapters for OpenClaw, CrewAI, LangGraph, AutoGen, and the Claude SDK.
• Security-First Design: Features RBAC, secret detection, trust scoring, and a four-layer evaluation framework for agent performance.
• Local-First Architecture: SQLite-backed storage with a simple one-command installation process.
• Claude Code Integration: Read-only bridge to surface Claude Code sessions and team tasks on the dashboard.
• Natural Language Scheduling: Ability to create recurring tasks using plain English (e.g., "every morning at 9am").