能力标签
⚙️
Agent工作流

AI工作流

基于 TypeScript · 无代码搭建完整 AI 自动化流程
英文名:agentboard
⭐ 371 Stars 🍴 43 Forks 💻 TypeScript 📄 MIT 🏷 AI 7.5分
7.5AI 综合评分
AI工作流TypeScript
✦ AI Skill Hub 推荐

经 AI Skill Hub 精选评估,AI工作流 获评「推荐使用」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 7.5 分,适合有一定技术背景的用户使用。

📚 深度解析
AI工作流 是一套完整的 AI Agent 自动化工作流方案。随着 AI 能力的不断提升,基于 Agent 的自动化工作流正在成为提升个人和团队效率的核心方式。区别于传统的 RPA 自动化(模拟鼠标键盘操作),AI Agent 工作流通过理解任务意图、动态规划执行路径,能够处理更复杂的非结构化任务。

AI工作流 工作流的设计遵循"最小配置,最大复用"原则:核心逻辑已经封装好,用户只需配置自己的 API Key 和业务参数即可快速上手。工作流内置错误处理和重试机制,在网络波动或 API 限速等情况下仍能稳定运行,适合作为生产环境的自动化基础设施。

在实际部署时,建议先在测试环境中运行 3-5 次,验证各个环节的输出结果符合预期,再部署到生产环境。AI Skill Hub 评分 7.5 分,是同类 Agent 工作流中的精选推荐。
📋 工具概览

优化的AI工作流Web GUI,支持iOS Safari和移动设备

AI工作流 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。

GitHub Stars
⭐ 371
开发语言
TypeScript
支持平台
Windows / macOS / Linux
维护状态
轻量级项目,按需更新
开源协议
MIT
AI 综合评分
7.5 分
工具类型
Agent工作流
Forks
43
📖 中文文档
以下内容由 AI Skill Hub 根据项目信息自动整理,如需查看完整原始文档请访问底部「原始来源」。

优化的AI工作流Web GUI,支持iOS Safari和移动设备

AI工作流 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。

📌 核心特色
  • 可视化 Agent 工作流编排,无需编写复杂代码
  • 支持多步骤自动化任务链,实现全流程无人值守
  • 与外部 API、数据库和第三方服务无缝集成
  • 内置错误处理与自动重试机制,保障稳定运行
  • 提供可复用的自动化模板,快速在同类场景部署
🎯 主要使用场景
  • 自动化日常重复性工作,将精力集中于创造性任务
  • 构建数据采集 → 处理 → 输出的完整自动化管线
  • 实现跨平台、跨系统的数据流转和业务协同
以下安装命令基于项目开发语言和类型自动生成,实际以官方 README 为准。
安装命令
# 方式一:npm 全局安装
npm install -g agentboard

# 方式二:npx 直接运行(无需安装)
npx agentboard --help

# 方式三:项目依赖安装
npm install agentboard

# 方式四:从源码运行
git clone https://github.com/gbasin/agentboard
cd agentboard
npm install
npm start
📋 安装步骤说明
  1. 访问 GitHub 仓库获取工作流文件
  2. 在对应平台(Dify / Flowise / Make 等)中找到「导入工作流」功能
  3. 上传工作流文件
  4. 按照提示配置必要的环境变量和 API Key
  5. 运行测试确认流程正常后投入使用
以下用法示例由 AI Skill Hub 整理,涵盖最常见的使用场景。
常用命令 / 代码示例
# 命令行使用
agentboard --help

# 基本用法
agentboard [options] <input>

# Node.js 代码中使用
const agentboard = require('agentboard');

const result = await agentboard.run(options);
console.log(result);
以下配置示例基于典型使用场景生成,具体参数请参照官方文档调整。
配置示例
# agentboard 配置说明
# 查看配置选项
agentboard --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 AGENTBOARD_CONFIG="/path/to/config.yml"
📑 README 深度解析 真实文档 完整度 46/100 查看 GitHub 原文 →
以下内容由系统直接从 GitHub README 解析整理,保留代码块、表格与列表结构。

Agentboard

CI npm License: MIT TypeScript Ask DeepWiki

Agentboard is a Web GUI for tmux that's optimized for agent TUI's (claude, codex, etc). A lighter-weight, agent-optimized alternative to Blink, Termux, etc.

Run your desktop/server, then connect from your phone or laptop over Tailscale/LAN. You get a shared workspace across devices.

  • iOS Safari mobile experience with:
  • Paste support (including images)
  • Touch scrolling
  • Virtual arrow keys / d-pad
  • Quick keys toolbar (ctrl, esc, etc.)
  • Out-of-the-box log tracking and matching for Claude, Codex, and Pi — auto-matches sessions to active tmux windows, with one-click Wake for Hibernating and History sessions.
  • Shows the last user prompt for each session, so you can remember what each agent is working on
  • Hibernate sessions to close their tmux window while keeping them visible across restarts for manual Wake

Requirements

  • tmux (brew install tmux / apt install tmux)
  • Tailscale (recommended) or any network path to your machine

Dependency Risk Scanner

Use the built-in scanner to report security and maintenance risk for direct dependencies:

bun run deps:risk

Machine-readable output:

bun run deps:risk:json

Policy:

  • Security risk comes from bun audit --json findings and is aggregated by severity (low, moderate, high, critical).
  • Maintenance risk comes from bun outdated and classifies version lag as major, minor, or patch behind latest.
  • Local default threshold is high (high + critical fail).
  • CI enforces security threshold breaches at critical only (bun run deps:risk:ci) while existing upstream high advisories are tracked; maintenance findings are warnings for prioritization.

You can override the security threshold with --threshold (or DEPENDENCY_RISK_FAIL_ON):

bun run deps:risk -- --threshold moderate

Install

Environment

PORT=4040
HOSTNAME=127.0.0.1
TMUX_SESSION=agentboard
REFRESH_INTERVAL_MS=5000
DISCOVER_PREFIXES=work,external
PRUNE_WS_SESSIONS=true
AGENTBOARD_PREFER_WINDOW_NAME=false
TERMINAL_MODE=pty
TERMINAL_MONITOR_TARGETS=true
VITE_ALLOWED_HOSTS=nuc,myserver
AGENTBOARD_DB_PATH=~/.agentboard/agentboard.db
AGENTBOARD_INACTIVE_MAX_AGE_HOURS=24
AGENTBOARD_EXCLUDE_PROJECTS=<empty>,/workspace
AGENTBOARD_HOST=blade
AGENTBOARD_REMOTE_HOSTS=mba,carbon,worm
AGENTBOARD_REMOTE_POLL_MS=2000
AGENTBOARD_REMOTE_TIMEOUT_MS=4000
AGENTBOARD_REMOTE_STALE_MS=45000
AGENTBOARD_REMOTE_SSH_OPTS=-o BatchMode=yes -o ConnectTimeout=3
AGENTBOARD_REMOTE_ALLOW_ATTACH=false
AGENTBOARD_REMOTE_ALLOW_CONTROL=false
AGENTBOARD_LOG_WATCH_MODE=watch

HOSTNAME controls which interfaces the server binds to (default 127.0.0.1 for localhost-only). With the default localhost binding, if Tailscale is detected the server also binds to your Tailscale IP automatically. Set to 0.0.0.0 to listen on all interfaces.

Security note: Agentboard has no built-in authentication. Anyone who can reach the server has full access to your terminal sessions, including the ability to run commands as your user. The default localhost binding is safe. Tailscale provides network-level auth for remote access. Avoid setting HOSTNAME=0.0.0.0 on untrusted networks (public WiFi, shared LANs) without an additional access control layer.

DISCOVER_PREFIXES lets you discover and control windows from other tmux sessions. If unset, all sessions except the managed one are discovered.

PRUNE_WS_SESSIONS removes orphaned agentboard-ws-* tmux sessions on startup (set to false to disable).

AGENTBOARD_PREFER_WINDOW_NAME (default false) controls how externally-discovered sessions are labeled. When false, the tmux session name is used (more meaningful than auto-renamed window names that follow the running process under tmux automatic-rename on). Set to true to use the tmux window name when it is non-empty and distinct from the session name — useful when you keep one shared session (e.g. dev) with one explicitly-named window per project (myapp, infra, ...).

TERMINAL_MODE selects terminal I/O strategy: pty (default, grouped session) or pipe-pane (PTY-less, works in daemon/systemd/docker without -t).

TERMINAL_MONITOR_TARGETS (pipe-pane only) polls tmux to detect closed targets (set to false to disable).

VITE_ALLOWED_HOSTS allows access to the Vite dev server from other hostnames. Useful with Tailscale MagicDNS - add your machine name (e.g., nuc) to access the dev server at http://nuc:5173 from other devices on your tailnet.

All persistent data is stored in ~/.agentboard/: session database (agentboard.db) and logs (agentboard.log). Override paths with AGENTBOARD_DB_PATH and LOG_FILE.

AGENTBOARD_INACTIVE_MAX_AGE_HOURS limits History sessions shown in the UI to those with recent activity (default: 24 hours). Older sessions remain in the database but are not displayed or processed for orphan rematch. The env var name is kept for compatibility.

AGENTBOARD_EXCLUDE_PROJECTS filters out sessions from specific project directories (comma-separated). Use <empty> to exclude sessions with no project path. Useful for hiding automated/spam sessions.

AGENTBOARD_SKIP_MATCHING_PATTERNS controls which orphan sessions skip expensive window matching (comma-separated). Defaults: <codex-exec> (headless Codex exec sessions), /private/tmp/*, /private/var/folders/*, /var/folders/*, /tmp/*. Patterns support trailing * for prefix matching. Set to empty string to disable skip matching entirely.

AGENTBOARD_HOST sets the host label for local sessions (default: hostname).

AGENTBOARD_REMOTE_HOSTS enables remote tmux polling over SSH. Provide a comma-separated list of hosts (e.g., mba,carbon,worm). Remote sessions show live status (working/waiting/permission) via pane content capture over SSH.

AGENTBOARD_REMOTE_POLL_MS, AGENTBOARD_REMOTE_TIMEOUT_MS, and AGENTBOARD_REMOTE_STALE_MS control remote poll cadence, SSH timeout, and stale host cutoff.

AGENTBOARD_REMOTE_SSH_OPTS appends extra SSH options (space-separated).

AGENTBOARD_REMOTE_ALLOW_ATTACH enables interactive terminal attach to remote sessions (input, resize, copy-mode). When false (default), remote sessions are view-only.

AGENTBOARD_REMOTE_ALLOW_CONTROL enables destructive remote session management (create, kill, rename) via the UI. Setting this to true implies REMOTE_ALLOW_ATTACH=true. Kill and rename are restricted to agentboard-managed sessions — externally-discovered remote sessions cannot be killed or renamed even with control enabled.

AGENTBOARD_LOG_WATCH_MODE selects the log detection strategy: watch (default) uses fs.watch for instant file-change detection, poll falls back to periodic directory scanning. Use poll if you experience issues with filesystem notifications (e.g., on network-mounted home directories). On Linux, watch mode automatically includes a 15-second fallback poll since fs.watch({ recursive: true }) has known platform bugs.

SSH multiplexing (recommended): Each poll cycle opens SSH connections to every remote host. Enable SSH connection multiplexing to reuse connections and reduce overhead from ~200-500ms to ~5ms per poll. Add to your ~/.ssh/config:

Host *
    ControlMaster auto
    ControlPath ~/.ssh/sockets/%r@%h-%p
    ControlPersist 600

Then create the sockets directory: mkdir -p ~/.ssh/sockets && chmod 700 ~/.ssh/sockets

Troubleshooting

🎯 aiskill88 AI 点评 A 级 2026-05-26

优化的AI工作流工具,易于使用和扩展

⚡ 核心功能
👥 适合人群
自动化工程师和运维人员项目经理和业务分析师希望减少重复性工作的专业人士数字化转型团队
🎯 使用场景
  • 自动化日常重复性工作,将精力集中于创造性任务
  • 构建数据采集 → 处理 → 输出的完整自动化管线
  • 实现跨平台、跨系统的数据流转和业务协同
⚖️ 优点与不足
✅ 优点
  • +MIT 协议,可免费商用
  • +大幅减少重复性人工操作
  • +可视化流程,清晰直观
  • +可扩展性强,支持复杂场景
⚠️ 不足
  • 初始配置和调试需投入一定时间
  • 强依赖外部服务的稳定性
  • 复杂场景需具备一定技术基础
⚠️ 使用须知

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。

📄 License 说明

✅ 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。

🔗 相关工具推荐
🧩 你可能还需要
基于当前 Skill 的能力图谱,自动补全的工具组合
❓ 常见问题 FAQ
参考项目文档和示例
💡 AI Skill Hub 点评

AI Skill Hub 点评:AI工作流 的核心功能完整,质量良好。对于自动化工程师和运维人员来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。

⬇️ 获取与下载
⬇ 下载源码 ZIP

✅ MIT 协议 · 可免费商用 · 直接从 aiskill88 服务器下载,无需跳转 GitHub

📚 深入学习 AI工作流
查看分步骤安装教程和完整使用指南,快速上手这款工具
🌐 原始信息
原始名称 agentboard
Topics AI工作流TypeScript
GitHub https://github.com/gbasin/agentboard
License MIT
语言 TypeScript
🔗 原始来源
🐙 GitHub 仓库  https://github.com/gbasin/agentboard

收录时间:2026-05-26 · 更新时间:2026-05-26 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。