Q00/ouroboros

Overview

Q00/ouroboros is an Agent OS that enables agents to improve themselves through an interview-gated, staged evaluation process with a budgeted evolution loop. It operates as an MCP server and supports 14 runtimes including Claude Code, Codex CLI, Gemini CLI, OpenCode, Copilot, and Kiro. Written in Python, it is open-source with over 5,700 stars.

Best for

Best for
Developers building self-improving AI agents across multiple CLI environments.

Use cases

• Automate agent self-improvement pipelines
• Evaluate agent performance across multiple CLI runtimes
• Integrate MCP-based agent orchestration into existing workflows

How to use

Install

pip install 'ouroboros-ai[mcp,tui]' && ouroboros setup --runtime claude-cli  # recommended MCP v2 default

Tested with

Claude Code, Continue, ChatGPT

Notes

Q00/ouroboros is an Agent OS that enables agents to improve themselves through an interview-gated, staged evaluation process with a budgeted evolution loop. It operates as an MCP server and supports 14 runtimes including Claude Code, Codex CLI, Gemini CLI, OpenCode, Copilot, and Kiro. Written in Python, it is open-source with over 5,700 stars.

5,799 stars on GitHub. Last updated 2026-09-08. Licensed MIT.

Use cases

• Automate agent self-improvement pipelines
• Evaluate agent performance across multiple CLI runtimes
• Integrate MCP-based agent orchestration into existing workflows

Pros

• Broad runtime compatibility with 14 supported tools
• Self-improving architecture with staged evaluation
• Active open-source project with 5.8k stars

Cons

• Requires Python environment and MCP setup
• Evolution loop may need careful budget tuning
• Not a standalone agent, depends on external runtimes

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.