能力标签

🤖 Agent 🔄 工作流 💻 CLI 🔗 REST API 🧠 Claude ✨ GPT ⌨️ Cursor 🖥 本地 LLM

⚙️

Agent工作流

伊格德拉斯

基于 TypeScript · 无代码搭建完整 AI 自动化流程

英文名：Yggdrasil

⭐ 26 Stars 🍴 6 Forks 💻 TypeScript 📄 MIT 🏷 AI 8.0分

8.0AI 综合评分

AI工作流TypeScript

⬇ 下载源码 ZIP ⚙️ 配置说明

✦ AI Skill Hub 推荐

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。

📚 深度解析

伊格德拉斯是一套完整的 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。

伊格德拉斯工作流的设计遵循"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 8.0 分，是同类 Agent 工作流中的精选推荐。

📋 工具概览

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。

GitHub Stars

⭐ 26

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

8.0 分

工具类型

Agent工作流

Forks

📖 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」。

📌 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 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

📋 安装步骤说明

访问 GitHub 仓库获取工作流文件
在对应平台（Dify / Flowise / Make 等）中找到「导入工作流」功能
上传工作流文件
按照提示配置必要的环境变量和 API Key
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
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);

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

# 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"

📑 README 深度解析真实文档完整度 40/100 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。

简介

Getting started

1. Install and init. Requires Node.js 22+.

npm install -g @chrisdudek/yg
cd your-project
yg init

The wizard walks you through platform selection and reviewer setup (provider, model, and where keys live in yg-config.yaml / yg-secrets.yaml).

2. Tell the agent what matters.

You:    "All payment operations must emit audit events."
Agent:  Creates rule, applies it to payment code.

You:    "All API endpoints must validate input with zod."
Agent:  Creates rule, applies it to endpoint handlers.

The agent manages the structure. Which rules apply where, which files are mapped, how components relate. You say what should be enforced.

3. Work normally.

The agent verifies its own code as it works. When it violates a rule, it gets feedback and fixes it.

4. Enforce in CI.

- run: npx @chrisdudek/yg check

yg check is the deterministic gate: it makes no LLM calls and needs no provider config or keys. It recomputes the input hash of every expected pair and compares it against the verdict recorded in the lock, and also validates structure, schema, coverage, and completeness. If code changed without being verified, the pair no longer matches its recorded hash and check fails.

Examples

examples/ has two projects you can run. One passes, one has a deliberate violation for the reviewer to catch.

This repo uses Yggdrasil on itself. Browse .yggdrasil/ for a real, live graph, or run yg check in a clone to see the current node and aspect coverage for yourself.

FAQ

How is this different from CLAUDE.md or .cursorrules? Rules files are flat text dumped into every prompt. No scoping, no verification. Yggdrasil delivers only the rules relevant to each file and reviews the output against them.

How is this different from linters? Linters check syntax and patterns. "Rate limiting required" isn't a lint rule. "No direct DB access from this layer" isn't in any AST. "All mutations must emit audit events" can't be checked with regex. Yggdrasil reviews against rules that only exist in your head until you write them down — and where a rule is mechanically checkable, you can write it as a script that runs for free.

How is this different from a PR review? PR review happens after the code is written. By then the agent has moved on, context is lost, and you're catching up. Yggdrasil reviews while the agent is working, so violations get fixed in the same session.

Does it work? Locally, yg check --approve sends LLM aspects to the reviewer and runs script aspects on your machine, then records each verdict in a single committed lock file. yg check in CI makes no LLM calls — it recomputes the input hash of every expected pair and compares it against the verdict the lock recorded (and validates structure and coverage). If a PR has unverified changes, the hashes no longer match and CI catches it.

What if I want to stop? Delete .yggdrasil/ and the rules file. No runtime dependencies, no build hooks, nothing left behind.

Is this just another AI code review bot? No. AI code review bots scan diffs for bugs after the fact. Yggdrasil runs a reviewer against specific rules you wrote, inside the agent's loop, so violations get fixed in the same session rather than piling up in a PR that nobody has time to read. The rules are yours; the enforcement is what Yggdrasil adds.

🎯 aiskill88 AI 点评 A 级 2026-06-13

高质量的AI工作流项目，值得关注