nyldn/claude-octopus

Run multiple AI models against the same research, design, or coding task. Surface disagreements before you ship.

Claude Octopus – a multi‑LLM orchestration layer for Claude Code

What it is – Claude Octopus is a plugin for Claude Code (Anthropic’s IDE‑style coding assistant). It stays idle until you explicitly invoke an /octo:* command, at which point it can call up to twelve external LLM providers (e.g. Codex, Antigravity CLI, GitHub Copilot, Qwen, Ollama, Perplexity, OpenRouter, OrcaRouter, OpenCode, Grok, Kimi Code) and run them in parallel or in a structured “council”. The built‑in Claude model remains the default; Octopus is only used for escalation when you want multiple opinions, adversarial review, or stricter quality gates.

Key capabilities

  • Consensus gating – a 75 % agreement threshold flags disagreements before code is shipped.
  • Persistent memory – integrates with claude‑mem and agentmemory so decisions and research survive across sessions.
  • “Dark Factory” mode – give a specification and Octopus runs the full pipeline (research → define → develop → deliver) automatically; you only review the final output.
  • Method‑aware workflows – pre‑packaged flows for architecture, TDD, debugging, etc., each selecting the appropriate set of personas.
  • Multi‑LLM council – /octo:council can convene 3‑7 personas across different providers with configurable goals (advice, decision, plan, implement, review) and styles (balanced, adversarial, red‑team, executive).
  • Rich persona library – 31 specialized AI agents (security auditor, backend architect, etc.) and 53 slash commands that map to 63 reusable workflow modules.
  • Zero‑provider start – you can install Octopus and begin using Claude alone; external providers are added one‑by‑one and become available only inside explicit workflows.
  • Cost‑aware – five providers (Codex, Antigravity CLI, Copilot, Cursor CLI, Ollama) cost nothing extra if you already have a subscription or run locally.

How you use it

  1. Install via Claude Code’s plugin marketplace (/plugin → marketplace → install octo).
  2. Run /octo:setup inside a Claude session – the command detects which external providers are installed and guides you through any needed configuration.
  3. Invoke a workflow, e.g.:
    /octo:embrace   # full Discover → Define → Develop → Deliver pipeline
/octo:council --goal decision --style adversarial "Should we split this monolith?"
/octo:review   # independent code review by a second model
  4. Optional auto‑router mode can suggest or automatically invoke Octopus for ordinary prompts (OCTOPUS_AUTO_ROUTER_MODE=suggest|invoke).

Supported environments – works natively in Claude Code, but the repo also ships:

  • A Codex CLI plugin (codex plugin add claude‑octopus).
  • An MCP server for the Cursor IDE (Node.js server exposing tools like octopus_discover).
  • Simple install scripts for OpenCode and other Claude‑compatible hosts.

Version 11.2.0 highlights

  • New engineering methods derived from Matt Pocock’s skill set.
  • Default model roster: Claude Opus 5 (architecture & security), GPT‑5.6 Sol (implementation/review), Claude Sonnet 5 (standard seat), optional Claude Fable 5.1.
  • Environment variables for fine‑grained control (e.g., OCTOPUS_OPUS5_AUTO_XHIGH, OCTOPUS_CONTEXT_AWARENESS).

Safety & updates

  • All Octopus commands are gated behind explicit slash‑command invocation; the model cannot auto‑call Octopus without you.
  • Built‑in diagnostics (octopus doctor config, octopus doctor skills, octopus doctor updates).
  • Auto‑update is disabled by default; you can enable it through the Claude UI or run the provided orchestrate.sh script.

Who it’s for – developers who already use Claude Code and want a controlled, multi‑model review layer for higher‑risk code, architecture decisions, or autonomous spec‑to‑code pipelines, without having to manually switch between different AI services.

Bottom line – Claude Octopus is an open‑source orchestration plugin that lets Claude Code call out to a dozen other LLM providers, run consensus checks, and automate end‑to‑end development workflows, all behind explicit user commands.

Related

  • Project
    hex/claude-council

    A Claude Code plugin that queries multiple AI coding agents in parallel to provide side-by-side comparisons and a synthesis of consensus and divergence for technical decisions.

  • Project
    Yeachan-Heo/oh-my-claudecode

    A multi-agent orchestration framework for Claude Code that enables natural-language task distribution across specialized agents and multiple AI providers.

  • Project
    fengshao1227/ccg-workflow

    CCG‑Workflow is a Node‑JS CLI that makes Claude Code the central orchestrator for multiple code‑generation models (Codex, Gemini, etc.). It injects JavaScript hooks to keep context, creates persistent task folders, runs built‑in strategies (e.g., full‑collaborate), applies security/quality gates, and can even deploy to Baota panels. Install with `npx ccg-workflow` (full mode) or as a Claude plugin for just the reusable skills.

  • Project
    CronusL-1141/AI-company

    AI Team OS is an open‑source platform that turns a Claude‑style LLM into an autonomous “CEO”. It continuously runs without prompting, orchestrating specialist agents, a task wall, research pipelines, and a two‑layer memory system. The FastAPI + React dashboard shows live workflows, decision traces, and a unified search over all memos. It’s aimed at technical founders or research teams who want an AI‑first, self‑driving company‑like workflow.

  • Project
    Prorise-cool/Claude-Code-Multi-Agent

    Claude Code Multi‑Agent is an open‑source framework that gives Anthropic’s Claude Code project awareness and expert‑agent capabilities. It uses a Python Hook system and a locally‑run Ollama model to automatically detect the codebase type, analyse user intent, load 300+ domain‑specific “Skills” (backend, frontend, testing, security, etc.), and enforce automatic documentation (DEVELOPMENT.md, KNOWLEDGE.md, CHANGELOG.md). After cloning, installing Ollama, a model, and uv, you drop your project into the repo and Claude Code instantly knows its context, suggests appropriate tools, and can be driven with slash commands like `/backend‑specialist` or workflow commands such as `/kiro/spec`. Key points: - Zero‑config start; all configuration lives in `.claude/settings.json` and `prompts.json`. - Hooks run on session start, user prompt, tool use, etc., handling project detection, intent analysis, and doc updates. - Skills are defined by `SKILL.md` files and are auto‑discovered. - Works entirely locally (no external API), integrates with Git, and is MIT‑licensed. Ideal for developers or small teams who want a reproducible, on‑premise AI coding assistant.