Ouroboros

Ouroboros is a locally installed workflow engine for developers who want to define requirements and check an AI coding agent's output. It runs through a terminal CLI, a Claude Code plugin or an MCP server. Your chosen coding runtime supplies the agent and model access. When used with Claude through Claude Code's standard cloud connection, model requests go to that external service.

The requirements interview identifies assumptions and scores ambiguity before execution. Its answers become a fixed specification with acceptance criteria and constraints, so the agent has a defined target throughout the run.

Evaluation combines mechanical checks with semantic assessment and consensus across models. The coding agent doesn't receive the grading command or expected result in its success contract. This separates the task it's asked to complete from the checks used to judge its work.

Evaluation also feeds further attempts. Stored run records preserve the history needed to resume the loop across sessions and machine restarts, which matters for coding tasks that take several iterations.

Supported runtimes include Claude Code, Codex CLI, OpenCode, Gemini CLI and GitHub Copilot CLI. Developers can keep the same requirements and evaluation process while changing the agent that writes the code. The Python core is open source under the MIT license; the Claude Code plugin requires Claude Code but doesn’t require a manual Python installation.