#4646th of 149 in Skills, plugins and rules
ouroboros
Spec-first workflow engine that interviews you, then runs and evaluates coding agents
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 6.2k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- Python
1,750 on 25 Mar 20266,227 on 14 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Ouroboros adds an `ooo` command set to coding-agent CLIs. It runs a Socratic interview that produces an immutable Seed spec with an ambiguity score, then executes, evaluates through a 3-stage gate, and evolves the result, recording actions in a replayable ledger. It registers as an MCP server and supports Claude Code, Codex CLI, OpenCode, Hermes, Gemini, Kiro, Copilot and others.
Who it is for: Developers wanting spec-first, replayable workflows on their coding agent CLI
Strengths
- Works across about 14 agent hosts, including Claude Code, Codex CLI, Kiro and Copilot CLI
- Interview step yields a locked Seed spec with a measurable ambiguity score
- Ships as a Claude Code and Codex plugin, pip package, Homebrew tap or one-line installer
- MIT licensed, with recent commits and releases
Weaknesses
- Native Windows is experimental, and Codex CLI on Windows needs WSL 2
- Python 3.12 or newer required; LiteLLM profiles limited to 3.12-3.13
- Many install profiles and MCP v1/v2 splits make setup choices confusing
- Needs an existing coding agent or LLM backend; it does not run models itself
What it needs
- no GPU
- Needs Python >= 3.12, uv or pipx, A supported coding-agent CLI or LLM backend
- Models: Claude Code, Codex CLI, OpenCode, Hermes, Gemini
Also in Skills, plugins and rules
See all 149| Rank | Project | Score |
|---|---|---|
| 1 | cavemanCuts token use in coding agents with terse replies and compressed inputs | 95 out of 100 |
| 2 | skillsAgent skills for planning, TDD, debugging and code review | 94 out of 100 |
| 3 | eccAgents, skills, hooks and rules bundle for Claude Code, Codex and other harnesses | 94 out of 100 |
| #4 | impeccableDesign skill with 24 commands and a detector for AI-generated frontends | 94 out of 100 |
| #5 | ponytailSkill prompt that makes coding agents write less code and add tests | 92 out of 100 |