analyze
Analyze recent runs - per-model latency, tokens, and verdict agreement - and suggest…
Start the local read-only dashboard (live and past deliberation runs as state graphs, config, health, stats) and print its URL. Needs a browser on the same machine.
> /plugin marketplace add antonbabenko/deliberation > /plugin install deliberation@antonbabenko-deliberation
How it fires
How this command gets triggered: by you, by Claude, or both.
/dashboardContext preview
What this command does when you run it.
Start the local read-only dashboard (live and past deliberation runs as state graphs, config, health, stats) and print its URL. Needs a browser on the same machine.
name: dashboard description: Start the local read-only dashboard (live and past deliberation runs as state graphs, config, health, stats) and print its URL. Needs a browser on the same machine. allowed-tools: Bash timeout: 30000
Starts the local dashboard: a read-only web page on `127.0.0.1` that draws each `/consensus`, `/ask-all` and `/ask-*` run as a state graph while it runs, plus run history, the effective config, provider health, and usage stats. It never changes deliberation state. Nothing is recorded until `dashboard.enabled` is `true` in the config.
Run it with the **sandbox disabled**: it binds a local port and writes its pidfile under `~/.cache`, which a sandbox blocks. Pass `--port N` only if the user asked for a port.
bash "${CLAUDE_PLUGIN_ROOT}/scripts/commands/dashboard.sh"If your host does not set `CLAUDE_PLUGIN_ROOT`, pass the plugin directory instead: `bash <plugin-root>/scripts/commands/dashboard.sh`.
The script needs a POSIX shell (Git Bash on Windows). Without one, tell the user to run `deliberation-mcp dashboard` (or `node <plugin-root>/server/mcp/index.js dashboard`) in a terminal instead.
The script starts `node server/mcp/index.js dashboard` detached, so the dashboard keeps running after this session ends, and waits for its first output line. A second run finds the live dashboard and prints the same URL.
Print the script's output **as-is**, then act on its exit code:
browser opens on it; tell the user to open that URL if it did not. The token in it is the only credential: say it is for this machine only and should not be pasted anywhere.
`"dashboard": { "enabled": true }` to that config file (it hot-reloads) and run `/deliberation:dashboard` again. Runs made before that are not recorded.
most likely a dashboard started before an update. Tell the user to check that pid and end it if it is the old dashboard, or rerun with `--port <another port>`. Do not kill it yourself.
predates the dashboard; tell the user to update it as the message says.
same machine, and `localhost` here is the remote container. Stop.
Get a second opinion in Claude Code from GPT, Gemini, and Grok - plus 400+ more models through OpenRouter, including Qwen, Kimi, and DeepSeek.
Repo: antonbabenko/deliberation
Analyze recent runs - per-model latency, tokens, and verdict agreement - and suggest…
Ask GPT, Gemini, Grok, and any configured OpenRouter models in parallel for independent…
Get Gemini second opinion on a question or current work. Single-shot, advisory, no…
Get GPT (Codex) second opinion on a question or current work. Single-shot, advisory, no…
Get Grok (xAI) second opinion on a question or current work. Single-shot, advisory, no…
Ask a single configured OpenRouter model for a second opinion. Advisory only. Single-shot or…