经 AI Skill Hub 精选评估,好帮手 获评「强烈推荐」。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 8.0 分,适合有一定技术背景的用户使用。
好帮手 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。
好帮手 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:npm 全局安装 npm install -g goodboy # 方式二:npx 直接运行(无需安装) npx goodboy --help # 方式三:项目依赖安装 npm install goodboy # 方式四:从源码运行 git clone https://github.com/akhayam99/goodboy cd goodboy npm install npm start
# 命令行使用
goodboy --help
# 基本用法
goodboy [options] <input>
# Node.js 代码中使用
const goodboy = require('goodboy');
const result = await goodboy.run(options);
console.log(result);
# goodboy 配置说明 # 查看配置选项 goodboy --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 GOODBOY_CONFIG="/path/to/config.yml"
Stop re-explaining yourself.
You have a repo. You have a goal. You also have four CLIs open in four windows, each holding a slightly different version of the same task. By evening you've spent more time pasting the goal back into the next chat than actually building.
Goodboy is a desktop app that holds the goal, the plan and the context once, then hands them to whichever agent you want to run next. Same brief, same memory, different model. Conversation, plans, decisions and PR state stay in a local SQLite on your machine. Your keys, your data, your bandwidth.
<img width="3972" height="2234" alt="CleanShot 2026-06-02 at 04 59 56@2x" src="https://github.com/user-attachments/assets/45ed6f2d-f22a-4c6a-9278-ee4cddc7533c" />
Shared context, not vendor sessions. Goal, decisions, last summary, open questions. A summarizer keeps it fresh after every turn. Edit any field by hand when the agents get it wrong. The next agent shows up already briefed.
Provider swap mid-task, without amnesia. Each turn is rebuilt from the shared context, never resumed from a vendor's session blob. Drop Claude halfway, hand the same task to Cursor, Codex or Gemini, watch it pick up clean.
Workflows for the multi-step stuff. Refactor incoming? Line up a sequence: a cheap model to scout the area, a smart one to plan it, a mid one to implement, another to review, a cheap one to open the PR. Each step picks its own provider and model, so you're never paying Opus prices to run a grep.
Plans as artifacts, not transcript scrollback. Agents write the plan before they touch your code, and it stays put: something you can read, edit and hand to whichever model implements it. Not a message that scrolls away.
GitHub Studio. Every pull request you're involved in, in one inbox, bucketed by state (draft, in review, approved, merged). Open one and you've got the body, the lifecycle controls and the unresolved comments in a single view. Reply yourself, or hand a comment to an agent to resolve.
Linear Studio. Every open issue assigned to you, bucketed by Linear state. Pick one and the goal is already written, the branch is named, the linked PR is recognized. Hit launch and a session is on it, with the issue tagged in the rail above. Already shipped a PR for that issue? Pick "Continue on PR" instead of "Start fresh" and the same branch comes back, ready for the next round.
Cost meter that taps your shoulder. Every session shows what it's costing as it runs. Goodboy nudges you before you burn Opus on a one-liner.
macOS, Intel and Apple Silicon, one universal build. Signed and notarized by Apple, so it opens with no security prompts.
Homebrew (recommended).
brew install --cask akhayam99/tap/goodboy
Direct download. Grab the .dmg from the latest release and drag Goodboy to Applications.
Updates are automatic. When a new release ships, a "Restart to update" control appears in the status bar and next to the sidebar logo. One click downloads it and relaunches; Homebrew users can also brew upgrade --cask goodboy.
高质量的AI工作流项目
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。
AI Skill Hub 点评:好帮手 的核心功能完整,质量优秀。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。
| 原始名称 | goodboy |
| Topics | ai-agentsautomationtypescript |
| GitHub | https://github.com/akhayam99/goodboy |
| License | MIT |
| 语言 | TypeScript |
收录时间:2026-06-05 · 更新时间:2026-06-05 · 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