经 AI Skill Hub 精选评估,伊格德拉斯 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。
伊格德拉斯 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。
伊格德拉斯 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 yggdrasil # 方式二:npx 直接运行(无需安装) npx yggdrasil --help # 方式三:项目依赖安装 npm install yggdrasil # 方式四:从源码运行 git clone https://github.com/krzysztofdudek/Yggdrasil cd Yggdrasil npm install npm start
# 命令行使用
yggdrasil --help
# 基本用法
yggdrasil [options] <input>
# Node.js 代码中使用
const yggdrasil = require('yggdrasil');
const result = await yggdrasil.run(options);
console.log(result);
# yggdrasil 配置说明 # 查看配置选项 yggdrasil --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 YGGDRASIL_CONFIG="/path/to/config.yml"
<p align="center"> <img src="docs/public/demo.gif" alt="Yggdrasil review loop" width="900" /> </p>
It enforces structure, not runtime behaviour. It can require that you call the audit utility. It cannot prove the audit fired in production.
A green check is only as good as the rule behind it. A shallow rule passes shallow code. The enforcement is real. Deciding what is worth enforcing stays yours.
examples/ has seven runnable projects, five of them keyless. This repository uses Yggdrasil on itself, so .yggdrasil/ is a live graph you can read. Full docs at krzysztofdudek.github.io/Yggdrasil.
progressive: reference: origin/main ```
Nothing is hidden and no rule is switched off. yg check --full answers for the whole project whenever you want the plain picture, and a recording run pays to review the rules your change reached instead of the whole backlog. Leave the key out and nothing changes at all.
Progressive mode has the rest.
How is this different from a rules file? A rules file is flat text dumped into every prompt, with no scoping and no verification. Here the agent gets only the rules that touch the file it is editing, and the output is checked against them.
How is this different from a pre-commit or agent hook? A hook is a real gate and you should use one. Point it at yg check and you have wired this in. What a bare hook has no notion of is which rule applies to which file, rules that need judgment rather than a script, and a lock that lets CI re-prove a model verdict for free.
How is this different from an AI review bot? Review bots hunt for bugs against their own idea of good code, and they re-run and re-bill on every pull request. This checks your specific rules, the ones only your team knows, and records a durable proof of each verdict.
What if I want to stop? Delete .yggdrasil/ and the rules file. No runtime dependencies, no build hooks, nothing left behind.
高质量的AI工作流项目,值得关注
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:伊格德拉斯 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 | Yggdrasil |
| 原始描述 | 开源AI工作流:Architecture rules your AI coding agent can't ignore. Every change gets checked 。⭐26 · TypeScript |
| Topics | AI工作流TypeScript |
| GitHub | https://github.com/krzysztofdudek/Yggdrasil |
| License | MIT |
| 语言 | TypeScript |
收录时间:2026-06-13 · 更新时间:2026-06-16 · 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