Claude Octopus (github.com)

Claude Octopus is an open-source extension for Claude Code that runs multiple AI models against the same research, design, or coding task to surface disagreements before shipping. It supports twelve external provider integrations alongside the built-in Claude Code host, including Codex, Copilot, Qwen, Ollama, Perplexity, and Grok. It is aimed at developers who want a second opinion, adversarial review, or stricter multi-LLM workflows, and the project states it is not affiliated with, endorsed by, or sponsored by Anthropic.

Octopus remains dormant by default and does not route ordinary prompts, launch provider workflows, or delegate to Octopus agents; every shipped command and skill uses Claude Code's native manual-invocation gate, so a user must explicitly type a slash command starting with /octo: to activate it. Setup detects installed providers, shows what is missing, and walks the user through configuration, and zero external providers are required to start since Claude is built in.

Claude Octopus includes 31 specialized personas such as security-auditor and backend-architect, 54 slash commands, and 63 reusable workflow skills. Its multi-LLM council feature runs a structured three, five, or seven-persona deliberation across Claude, Codex, Antigravity, and OpenCode with goal modes such as advice, decision, plan, implement, and review, plus quorum and critical-veto gates. A 75% consensus gate catches disagreements before they reach production, and Dark Factory mode takes a spec and autonomously runs research, define, develop, and deliver phases.

Claude Octopus runs natively on Linux and macOS; Windows users must run the Claude Code CLI inside WSL or a desktop SSH session, and native Git Bash, MSYS2, and Cygwin are unsupported. Five providers cost nothing extra when a user already has access, since Codex, Antigravity CLI, Copilot, and Cursor CLI use existing subscriptions or local auth and Ollama runs locally for free; Qwen now requires API-key or Coding-Plan authentication since its free OAuth tier ended.

Use cases

• Getting a second opinion on architecture or security decisions
• Comparing decisions with weighted criteria and scenario stress tests