Tab · Manage
Dashboard quickstart
Wire the dashboard into a fresh repo in four steps. No JSON to hand-edit, no env-var hunt. Clone, copy the example config, start the server, open localhost — you're reading your brain in a browser in under three minutes.
See it in motion
Where to find it
- Localhost:
/dashboard.README.html?repo=<id> - On disk:
.claude/brain/dashboard.README.md - Keyboard: ⌘ K then
quickstart - Sidebar: Manage → Help → Dashboard quickstart
What it does for you
repo-dashboard · cp repos.config.example.json repos.config.json · npm start · open http://localhost:4000. That's it — the config is committable, secrets stay in env vars.repos.config.json. The aggregator mirrors it as a worktree, indexes its brain, and the tier-0 picker shows both.npm ci finishes in seconds; the server starts in under a second. Hack any tab in plain JavaScript.Bootstrap a brain in any tracked repo that doesn’t have one
The aggregator works on day one for security, env doctor, vulnerabilities, and settings — even with no .claude/brain/. To unlock the brain-powered tabs (Sessions, Decisions, Errors, Lessons, Brain reflection, Link graph) in a given tracked repo, open Repo bootstrap in the sidebar (Manage → Help → Repo bootstrap), copy the prompt, and paste it into Claude Code, Cursor, or Codex pointed at that repo. The LLM creates .claude/brain/, CLAUDE.md, the skills, and the CI gates. Refresh the dashboard — the brain tabs light up. Plan for 2–3 minutes end-to-end including the paste.
Configure
Per-repo entries in repos.config.json support ${env:VAR} substitution for any secret (telemetry tokens, GitHub PATs). Document the var in .env.example + .claude/brain/config.md so the next dev knows what to set.
Use it well
Run the dashboard locally; don't deploy it. The whole point is your brain stays on your laptop, in your repo, on disk you control. When a teammate needs the same view, they clone the same repo-dashboard and add their repos.config.json. The hosted Team SKU exists for the rollup view, not as a replacement.