Deliberationv4.6.1

Install

Pick your host and follow its steps. Then run init once and check the result. You need at least one provider; Requirements below lists what each one needs.

Your host
  1. Add the marketplace

    /plugin marketplace add antonbabenko/agent-plugins
  2. Install the plugin

    /plugin install deliberation@antonbabenko

    The plugin manifest registers the MCP servers, so they load on start and stay current across updates.

  3. Run setup

    /deliberation:setup

    A one-time step. It runs init (below) and finds the providers you already have.

  4. Update later

    /plugin marketplace update antonbabenko
    /deliberation:reload-mcp
    /reload-plugins

    Pulls the new version, recycles the dashboard and reconnects the MCP servers. Restarting Claude Code works too.

  1. Add the marketplace

    codex plugin marketplace add antonbabenko/deliberation
  2. Install the plugin

    Open /plugins inside Codex and install Deliberation. The plugin starts the server with npx -y @antonbabenko/deliberation-mcp.

More for Codex CLI: other ways to install, provider keys, updating

  1. Add the MCP server

    Use the Install in Cursor link. It cannot carry provider keys, so set them in your environment.

  2. Add the rule

    Copy .cursor/rules/deliberation.mdc from the repository into your project's .cursor/rules/. The rule tells Cursor when to call the experts.

More for Cursor: other ways to install, provider keys, updating

  1. Choose "Add power from GitHub" and paste the repository URL

    https://github.com/antonbabenko/deliberation

    Kiro reads plugin.json, registers the MCP server from mcp.json and loads the steering files.

More for Kiro: other ways to install, provider keys, updating

  1. Install the extension

    gemini extensions install https://github.com/antonbabenko/deliberation

    Add --ref v<version> to pin a release, then restart Gemini CLI. The extension needs node on PATH.

  2. Check the server

    gemini extensions list
    gemini mcp list

    The deliberation server should show Connected. Gemini CLI starts stdio MCP servers only in trusted folders.

More for Gemini CLI: other ways to install, provider keys, updating

  1. Copy the skills and the MCP config into your project

    git clone --depth 1 https://github.com/antonbabenko/deliberation /tmp/deliberation
    mkdir -p .agents && cp -R /tmp/deliberation/.agents/skills /tmp/deliberation/.agents/mcp.json .agents/

    Run it in your project root. It puts .agents/skills/ and .agents/mcp.json into your project's .agents/.

More for Antigravity: other ways to install, provider keys, updating

  1. Register the MCP server in opencode.json

    {
      "$schema": "https://opencode.ai/config.json",
      "mcp": {
        "deliberation": {
          "type": "local",
          "command": ["npx", "-y", "@antonbabenko/deliberation-mcp"],
          "enabled": true,
          "environment": {}
        }
      }
    }

    The file sits at the project root, not inside .opencode/. Provider keys come from the environment OpenCode starts the server in.

  2. Copy the commands and agents

    Copy .opencode/commands/ and .opencode/agents/ from the repository into your project. The file name becomes the command or agent name.

More for OpenCode: other ways to install, provider keys, updating

  1. Add the server to your host's MCP config

    {
      "mcpServers": {
        "deliberation": {
          "command": "npx",
          "args": ["-y", "@antonbabenko/deliberation-mcp"]
        }
      }
    }

    Most hosts use the mcpServers key. VS Code uses servers and needs "type": "stdio", Zed uses context_servers, and Codex CLI reads TOML. The README lists the config file of each host and how to run the server in Docker.

First run, on any host

npx -y @antonbabenko/deliberation-mcp init

It finds the keys and CLI logins you already have, writes the config, pings each provider and runs one short demo consensus. Run it again any time: a second run changes nothing and says so. In Claude Code, /deliberation:setup runs it for you. The quickstart has every step.

Check that it works

Run /deliberation:doctor in Claude Code, or call the doctor tool on any other host. It is read-only: it checks the config, provider CLIs and keys, sessions and path drift, and suggests a fix for each problem it finds.

Requirements

Node.js 18 or newer runs the plugin, the bridges and the npm server.

You need at least one provider:

  • Codex CLI (GPT): npm install -g @openai/codex, then codex login (ChatGPT subscription; codex login --device-auth on a remote machine, one login per machine). A ChatGPT Business/Enterprise workspace can set CODEX_ACCESS_TOKEN instead. With neither, codex uses CODEX_API_KEY; OPENAI_API_KEY is never used.
  • Gemini (API key, the default): no CLI to install. Set GEMINI_API_KEY to a Google AI Studio key (get one at https://aistudio.google.com/apikey (external)). Advisory calls go to the Gemini API, model providers.gemini.apiModel (default gemini-flash-latest).
  • Antigravity CLI (agy, opt-in): needed only for Gemini implementation (workspace-write) calls, or if you set providers.gemini.transport to "agy"; with no key set, Gemini also falls back to it. Getting Started with Antigravity CLI (external), then run agy and log in. Read the Antigravity terms (external) first: section 6 lists "Using third party software, tools, or services to access the Service (e.g. using OpenClaw with Antigravity OAuth)" as a breach that "may be grounds for suspension or termination of your Antigravity and/or Gemini CLI accounts". Deliberation drives agy as third-party software.
  • Grok (xAI): no CLI to install; the bridge ships with the plugin (needs Node 18+). Set XAI_API_KEY (get a key at https://console.x.ai (external)).
  • OpenRouter: no CLI; the bridge ships with the plugin (needs Node 18+). Set OPENROUTER_API_KEY (get a key at https://openrouter.ai/keys (external)). With no OpenRouter model records, a built-in panel of z-ai/glm-5.3-flashx (default), openai/gpt-6.1-sol and google/gemini-3.8-flash joins ask-all and consensus. To pick your own models, declare them in ~/.config/deliberation/config.json (Windows: %APPDATA%\deliberation\config.json; override with DELIBERATION_CONFIG); they replace the built-in panel. providers.openrouter.enabled: false turns OpenRouter off. Works with other OpenAI-compatible endpoints (vLLM, HuggingFace Inference) through providers.openrouter.apiBase - auth is skipped automatically when the key env var is empty. Supports custom OpenRouter presets (e.g. cheapest-first) to optimize inference cost and provider routing.
  • Ollama and LM Studio: keyless by default (set apiKeyEnv if your server needs a key). Each has its own provider block, providers.ollama (default http://localhost:11434/v1) and providers.lmstudio (default http://localhost:1234/v1); declare models with provider: "ollama" or provider: "lmstudio". See SETUP.md.

Windows note (CLI providers). npm install -g on Windows installs a codex.cmd shim, not an executable, and Node cannot launch a .cmd directly - so GPT calls used to fail with nothing but a spawn error. Deliberation now resolves the CLI itself (PATHEXT, falling back to the package's own entry point) and keeps it out of a shell. If resolution ever misses on your machine, point CODEX_BIN (GPT) or AGY_BIN (Gemini) at the real executable. Grok and OpenRouter are unaffected - they are HTTP bridges and spawn nothing.

Claude Code on the web note. The web host caps every MCP tool call at 60s (MCP_TOOL_TIMEOUT=60000). The plugin manifest overrides that per server: each Deliberation server declares "timeout": 1800000 (30 min, which Claude Code applies ahead of MCP_TOOL_TIMEOUT) and mirrors it into the server's env so the plugin's own clamp tracks the real cap. Nothing to set in the environment. A timeout result whose message still names MCP_TOOL_TIMEOUT=60000 means an older plugin install: claude plugin update deliberation@antonbabenko and start a new session. /deliberation:doctor reports the cap, the codex credential, and a missing agy (set GEMINI_API_KEY and Gemini needs no agy for advisory calls); the panel skips providers that cannot answer (panel.unavailable) instead of waiting on them. For GPT, do not copy your laptop's ~/.codex/auth.json into the container: a ChatGPT login's refresh token works once, so both the other copy stops working as soon as one of them refreshes. On Plus/Pro, the first GPT call in a session starts codex login --device-auth and shows its link and code (a dialog when the host supports it, else in the result); approve it and GPT answers. On Business/Enterprise, set CODEX_ACCESS_TOKEN. See SETUP.md.

Rendered from README.md in the repository.

Type to search the site.