dsh-claude-octopus
Run the following command in DeepSeek Harness:
dsh plugin install GongYuanCaiJi/dsh-claude-octopus
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install GongYuanCaiJi/dsh-claude-octopus in DeepSeek Harness to install; source code is available at https://github.com/GongYuanCaiJi/dsh-claude-octopus
About this plugin
Every AI model has blind spots. Architecture judgment, security review, and implementation detail rarely align from a single perspective. dsh-claude-octopus puts one /octo command in front of up to ten external providers — Codex, Antigravity, Copilot, Qwen, Ollama, Grok and more — each seated with a distinct persona to research, build, and review independently, then a 75 percent consensus gate blocks delivery until disagreements are resolved. Zero external providers also works; the host itself is a seat.
Core capabilities span the full delivery pipeline. Dark Factory mode takes a spec and autonomously runs the Discover, Define, Develop, Deliver sequence with quality gates between every phase. Cross-session memory keeps decisions and research context alive across sessions. Thirty-two personas, fifty-four commands, and sixty-three skills activate only on explicit workflow invocation, leaving everyday requests untouched. The /octo debate subcommand adds a structured AI Debate Hub for multi-model discourse on the same question.
Best for developers already juggling several AI tools who want cross-validation before critical deliverables, and for teams that want a multi-model collaborative workflow behind a single command surface.
Use Cases
- Cross-validate critical architecture decisions by having multiple AI models review independently before shipping
- Hand over a spec and let the pipeline autonomously run Discover, Define, Develop, Deliver with quality gates between phases
- Flag and resolve multi-model disagreements with a 75 percent consensus threshold before any output reaches the deliverable
Best For
- Developers already running several AI coding tools who want cross-validation on critical outputs
- Engineering teams that require multi-model review as a quality gate before delivery
- AI engineers who want a single command surface to orchestrate a multi-model collaborative workflow
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.