CUGA

Open-source enterprise agent **harness**: a code-act loop in which the model writes Python that calls MCP, OpenAPI, and LangChain tools as functions inside a restricted sandbox, with results kept as variables, pre-execution verification, reflection, and tool shortlisting; CugaSupervisor coordinates multi-agent teams locally or over A2A, and runtime-enforced policies (intent guards, tool approvals, ToolGuard checks) gate what agents may do. Previously ranked #1 on the AppWorld and WebArena leaderboards; runs on open-weight or hosted models, including air-gapped deployments. Maintained by IBM and open to community contributions.

Stars

1.1k

Adoption surface

mostly simple

Autonomy

bounded

Recovery

resumable

License

✅ open-source

Category

Multi-agent and orchestration

Repository ↗ Example: Example apps ↗

Related in Multi-agent and orchestration

• MetaGPT
• autogen
• crewAI
• OpenManus
• ChatDev
• openai-agents-python

Maintain CUGA? Add the badge

Paste this into your README so readers can find the list and see how CUGA compares:

[![Best of Agent Harnesses](https://img.shields.io/badge/%F0%9F%8F%86_Best_of-Agent_Harnesses-5ac4bf)](https://github.com/RyanAlberts/best-of-Agent-Harnesses)

Copy