Agent Dashboard

agent-dashboard

A tmux-integrated orchestrator and dispatcher for AI coding agents — Claude Code, Codex, and more.

`agent-dashboard` runs coding agents across tmux panes, dispatches your input to whichever one needs you, and gates each session through workflow skills (TDD, conventional commits, branch policy) enforced by hooks. The TUI is built with [Bubble Tea](https://github.com/charmbracelet/bubbletea); a companion PWA exposes the same orchestration surface from your phone over your local network.

https://github.com/user-attachments/assets/01aa0f85-cfd4-4dc3-ac46-651bcfc03f99

**Documentation: [bjornjee.github.io/agent-dashboard](https://bjornjee.github.io/agent-dashboard/)** — getting started, guides, keybindings, settings, and architecture.

Why agent-dashboard?

• One pane for many agents. Stop alt-tabbing between tmux panes — see every agent grouped by state (blocked, waiting, running, review, PR, merged) in a single TUI, with live pane capture so you don't lose context when you switch.
• Workflow gates, not vibes. Skills (feature, fix, chore, refactor, investigate, implement, pr, rca) plus hooks (commit-lint, test-gate, no-commits-to-main, destructive-op warnings) enforce TDD and conventional commits at the harness level — agents can't merge if tests fail.
• Phone-first remote control. A companion PWA exposes the same orchestration surface over your local network: approve permissions, reply to questions, open/merge PRs, and get browser notifications when an agent needs you — all without opening your laptop.
• tmux-native, not tmux-replacing. Agents stay where they live; the dashboard adds a control plane on top of tmux capture-pane. No new pane manager, no daemon competing with tmux. If tmux dies, orphaned sessions survive as resumable and continue in a fresh pane.
• Multi-backend. Claude Code and Codex CLI are supported directly, with per-spawn harness, model, and thinking-effort selection in the New Agent flow.

Features