AI Skill Hub 强烈推荐:AI代理仪表盘 是一款优质的Agent工作流。AI 综合评分 8.0 分,在同类工具中表现稳健。如果你正在寻找可靠的Agent工作流解决方案,这是一个值得深入了解的选择。
AI代理仪表盘 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
AI代理仪表盘 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:go install(推荐) go install github.com/bjornjee/agent-dashboard@latest # 方式二:从源码编译 git clone https://github.com/bjornjee/agent-dashboard cd agent-dashboard go build -o agent-dashboard . # 方式三:下载预编译二进制 # 访问 Releases 页面下载对应平台二进制文件 # https://github.com/bjornjee/agent-dashboard/releases
# 查看帮助 agent-dashboard --help # 基本运行 agent-dashboard [options] <input> # 详细使用说明请查阅文档 # https://github.com/bjornjee/agent-dashboard
# agent-dashboard 配置说明 # 查看配置选项 agent-dashboard --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export AGENT_DASHBOARD_CONFIG="/path/to/config.yml"
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; 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 — getting started, guides, keybindings, settings, and architecture.
TUI — real-time agent monitoring with live tmux pane capture, subagent tree, conversation history, GitHub-style split-pane diff viewer, plan and mermaid-diagram viewers, per-agent token/cost dashboard with live rate-limit bars, session creation with frecency-ranked paths and skill/model/effort selection, a search palette (/) that also resumes orphaned sessions, quick replies and numbered-option sends, and a full GitHub PR workflow (create, review diff, merge via gh).
Mobile companion — an installable PWA with the same state grouping, full remote control (approve/reject, reply, stop, PR create/merge/close), session creation, browser notifications, usage dashboard, live updates over Server-Sent Events, and optional single-user Google OAuth.
See the guides for walkthroughs and the reference for the full keybinding and settings tables.
| Dependency | Required | Purpose |
|---|---|---|
| [tmux](https://github.com/tmux/tmux) | Yes | Agent pane management and live capture |
| [Claude Code](https://claude.com/claude-code) | Yes | The agents this dashboard monitors |
| [Node.js 18+](https://nodejs.org/) | Yes | Claude Code and Codex adapter hooks |
| [git](https://git-scm.com/) | Yes | Diff viewer, branch detection |
[GitHub CLI (gh)](https://cli.github.com/) | No | PR detection and merge |
| [Codex CLI](https://developers.openai.com/codex/) 0.130+ | No | Show Codex sessions in the dashboard |
| [z (zsh plugin)](https://github.com/agkozak/zsh-z) | No | Frecency-ranked directory suggestions |
curl -fsSL https://raw.githubusercontent.com/bjornjee/agent-dashboard/main/install.sh | sh
The installer downloads the binary for your platform from the latest GitHub Release, verifies its SHA256 checksum, and installs it to ~/.local/bin/agent-dashboard. No Go toolchain required.
Or build from source (requires Go 1.26+):
git clone https://github.com/bjornjee/agent-dashboard
cd agent-dashboard
./install.sh --build
In any Claude Code session, run:
/marketplace add bjornjee/agent-dashboard
/plugin install agent-dashboard@agent-dashboard
/plugin enable agent-dashboard@agent-dashboard
Then restart Claude Code sessions for hooks and skills to take effect.
In any Claude Code session, run /plugin uninstall agent-dashboard@agent-dashboard and /marketplace remove agent-dashboard. Then remove the binary and state:
curl -fsSL https://raw.githubusercontent.com/bjornjee/agent-dashboard/main/uninstall.sh | sh
Or from a repo checkout: make uninstall.
agent-dashboard
Or press prefix + D if you set up the tmux keybinding. The keys you'll use most:
| Key | Action |
|---|---|
j/k | Navigate agent list |
Enter | Jump to agent's tmux pane |
r | Reply to agent (free-text) |
y/n | Quick approve/reject |
a | Create new agent session |
/ | Search palette (also resumes orphaned sessions) |
d | Show git diff |
g / m | Open/create PR / merge PR |
h | Help overlay with all keybindings |
The full tables — including the diff viewer and search palette keys — are in the keybindings reference.
The included agent-dashboard.tmux script binds prefix + D to switch to a dedicated dashboard session:
```bash
tmux source-file ~/.tmux.conf ```
The dashboard reads a TOML file at ~/.agent-dashboard/settings.toml — banner, notifications, usage polling, thinking-effort defaults, and the harness picker (claude or codex) all live there. Missing keys fall back to sensible defaults; see settings.example.toml and the settings reference for every key, default, and environment variable.
Register the marketplace entry with Codex:
codex plugin marketplace add bjornjee/agent-dashboard
Then enable the plugin by appending the following to ~/.codex/config.toml:
[plugins."agent-dashboard@agent-dashboard"]
enabled = true
Restart Codex sessions and approve the agent-dashboard hooks prompt. Once approved, the dashboard sees Codex sessions just like Claude sessions — same state files, same conversation panel, same cost dashboard. From a repo checkout, make install-codex-adapter performs the registration and prints the config snippet.
agent-dashboard works best when paired with bjornjee/skills — a plugin of workflow skills (TDD guide, language-specific strict reviewers, refactor cleaner, codex delegation, terminal ops) that the dashboard's session-creation flow expects:
/marketplace add bjornjee/skills
/plugin install skills@bjornjee-skills
Without it, skill-gated session types (feature, fix, chore, refactor, investigate, implement, pr, rca) will not function as intended.
Do I need tmux? Yes. agent-dashboard reads live pane content via tmux capture-pane and spawns agent sessions in tmux panes. Without tmux there are no panes to monitor.
Which agents are supported? Claude Code is first-class via the adapter in adapters/claude-code/. Codex CLI is supported directly via the adapter in adapters/codex/. The architecture supports additional backends via the domain.Harness interface.
How do I use codex / gpt-5.x models? Pick codex in the New Agent harness step (TUI wizard or web form), with per-spawn flags from [harness.codex] in ~/.agent-dashboard/settings.toml. Codex sessions appear in the dashboard once the Codex plugin is installed and its hooks approved (see Codex CLI support).
Does this require a paid Claude account? No — it uses whatever Claude Code itself requires (Pro, Max, or API). agent-dashboard does not call the Anthropic API directly; it reads the JSONL transcripts Claude Code writes locally.
Can I use the dashboard without the mobile companion? Yes. The TUI is the primary interface. The PWA is optional and runs separately via make web.
How is this different from a generic tmux session manager? A session manager creates and switches panes. agent-dashboard understands what's running in each pane — it parses the harness's transcripts to detect state (blocked, waiting, running, review, PR, merged), captures plans and Mermaid diagrams, tracks token usage, and integrates the GitHub PR workflow.
Is this related to Claude Code's official UI? No. agent-dashboard is an unofficial third-party plugin. It builds on top of Claude Code's hooks system and JSONL transcripts but is not affiliated with Anthropic.
Does it work on Windows? Native Windows is unsupported — the project targets macOS and Linux because tmux is required. WSL with tmux installed should work but is untested.
agent-dashboard 是一个专为 AI 编程智能体(如 Claude Code、Codex 等)设计的集成式编排与调度器。它深度集成了 tmux,能够通过 tmux pane 管理多个智能体会话,并实现输入分发功能,确保指令能精准发送至所需的智能体。此外,它通过 hooks 机制引入了工作流技能(如 TDD、conventional commits 和分支策略),为 AI 辅助编程过程提供规范化约束。其 TUI 界面基于 Bubble Tea 构建,提���流畅的终端交互体验。
本项目核心功能包括:基于 tmux 的智能体会话管理与实时内容捕获;智能指令调度,支持将用户输入分发至特定的智能体 pane;通过 hooks 实现工作流技能门控(Workflow Skills),强制执行 TDD 或提交规范等开发流程;以及基于 Bubble Tea 构建的高质量终端用户界面(TUI),确保开发者在命令行环境下也能获得直观的操作体验。
运行本项目需要满足以下环境要求:核心依赖为 tmux,用于智能体 pane 的管理及实时内容捕获;必须安装 Claude Code,作为仪表盘监控的核心智能体对象。此外,TUI 界面依赖于 bubbletea、bubbles 和 lipgloss 等 Go 语言组件。请确保您的开发环境已正确配置上述工具,以保证 agent-dashboard 的各项功能正常运作。
安装分为两个主要步骤:首先,通过官方提供的安装脚本下载并安装预构建的二进制文件,该脚本会自动处理平台识别、SHA256 校验并将其安装至 `~/.local/bin/agent-dashboard`。其次,需要在 Claude Code 会话中通过 `/marketplace` 命令注册并安装插件,随后执行 `/plugin install` 与 `/plugin enable`。完成插件安装后,请重启 Claude Code 会话以使 hooks 和 skills 生效。
您可以直接在终端运行 `agent-dashboard` 命令来启动仪表盘。为了提升效率,建议配置 tmux 快捷键绑定:通过运行项目提供的 `agent-dashboard.tmux` 脚本并重新加载 tmux 配置(`tmux source-file ~/.tmux.conf`),您可以实现通过按下 `prefix + D` 快速切换到专属的仪表盘会话,实现智能体监控与常规开发的无缝转换。
用户可以通过修改 `~/.agent-dashboard/settings.toml` 文件进行个性化配置(若设置了 `$AGENT_DASHBOARD_DIR` 环境变量,则会优先读取该目录下的配置文件)。安装程序在首次运行时会从 `settings.example.toml` 复制模板文件。对于配置文件中缺失的键值,程序将自动回退至合理的默认设置,确保系统在灵活配置的同时保持稳定运行。
agent-dashboard 提供了对 Codex CLI 的扩展支持。通过在 `adapters/codex/` 目录下内置的适配器,开发者可以将 Codex 接入调度体系。您只需在 Codex 中通过 `codex plugin marketplace add` 注册插件,并在 `~/.codex/config.toml` 中添加相应的插件配置项,即可实现类似 Claude Code 的集成化管理体验。
为了获得最佳体验,强烈建议配合 `bjornjee/skills` 插件使用。该插件提供了一系列工作流技能,包括 TDD 指引、特定语言的严格审查器、重构清理工具、Codex 委派��及终端操作等。这些技能能够与 agent-dashboard 的会话创建流程完美契合,通过 `/marketplace` 安装后,可显著提升 AI 编程过程中的规范性与自动化程度。
常见问题解答:1. 是否必须使用 tmux?是的,agent-dashboard 依赖 `tmux capture-pane` 读取实时内容并在 tmux pane 中生成智能体会话,缺少 tmux 将无法实现监控功能。2. 支持哪些智能体?目前已原生支持 Claude Code,并通过适配器支持 Codex CLI。3. 如何卸载?可以通过 `/plugin uninstall` 命令移除插件并从 marketplace 中删除相关条目。
高质量的AI工作流管理工具
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
总体来看,AI代理仪表盘 是一款质量优秀的Agent工作流,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | agent-dashboard |
| 原始描述 | 开源AI工作流:Real-time tmux dashboard to monitor, manage, and orchestrate AI coding agents — 。⭐10 · Go |
| Topics | AI工作流代理管理实时监控 |
| GitHub | https://github.com/bjornjee/agent-dashboard |
| License | MIT |
| 语言 | Go |
收录时间:2026-05-26 · 更新时间:2026-05-30 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端