AGI-is-going-to-arrive

AGI-is-going-to-arrive / workflow-cookbook

Public

《织经》Claude Code Workflows (CLAUDE_CODE_WORKFLOWS) 深度剖析 · 29章+6附录 · 每个配方真跑过（附Run ID）| Loom: A hands-on deep dive into Claude Code Workflows — 29 chapters, real runs with traceable Run IDs

32
3
89% credibility
Found May 28, 2026 at 32 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
HTML
AI Summary

This is a comprehensive, bilingual (Chinese/English) cookbook that teaches users how to master Claude Code's Workflow feature for orchestrating multiple AI agents. The book contains 29 chapters plus 6 appendices, with 23 real verified workflow runs documented. It covers everything from basic concepts to advanced patterns like adversarial verification, loop-until-dry, and budget management. The guide uses a weaving metaphor (warp and weft) to explain how Workflows work, and provides hands-on recipes for common tasks like code review, bug hunting, and deep research. Everything is documented with actual execution logs, making it a trustworthy educational resource for developers learning to automate multi-agent workflows.

How It Works

1
💡 You hear about a new way to run AI agents

A friend tells you about Claude Code's Workflow feature that can automatically coordinate multiple AI assistants working together.

2
📖 You find the cookbook

You discover this comprehensive guide that teaches you everything about Workflows through real examples and hands-on recipes.

3
🧵 You learn the weaving metaphor

The guide explains that Workflows work like weaving cloth—the structure stays fixed while intelligent units shuttle through to get things done.

4
🚀 You run your first workflow

Following the cookbook, you ask your assistant to run a simple workflow and watch it execute step by step with visible progress.

5
You pick a recipe that fits your needs
🔍
Code review recipe

Automatically scan, review, verify, and summarize code changes across your entire project

🐛
Bug hunting recipe

Let multiple agents work together to find and verify bugs, with critics checking each other's work

📚
Deep research recipe

Have agents retrieve information and verify facts version by version until results are complete

6
🔧 You build your own workflow library

Using patterns from the cookbook, you create reusable workflows tailored to your specific projects and share them with others.

🎉 Your AI agents work together automatically

Instead of manually directing each step, your assistants now form up on their own and complete complex tasks as a coordinated team.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 32 to 32 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is workflow-cookbook?

This is a hands-on guide for Claude Code's Workflow feature (community nickname: "ultrawork") -- a JavaScript-based engine for deterministically orchestrating multiple AI agents. The book walks you through 29 chapters and 6 appendices, teaching you how to write reusable, testable pipelines using APIs like agent(), parallel(), pipeline(), and schema validation. It comes in both Chinese and English, with every recipe tied to a real execution and traceable Run ID.

Why is it gaining traction?

Most Claude Code tutorials show you prompts. This one shows you pipelines. The hook: every recipe claims to be actually run, with raw logs you can verify yourself. It covers advanced patterns developers actually need -- adversarial verification, loop-until-dry, budget management, resume from failure -- that you won't find in official docs. The bilingual format and static HTML site make it easy to read without installing anything.

Who should use this?

Devs already using Claude Code who want to automate multi-agent workflows beyond single-prompt sessions. Teams evaluating Claude Code for structured tasks like code review, bug hunting, or research pipelines. Anyone building reusable agent pipelines who needs patterns beyond "ask Claude to do X."

Verdict

The credibility score of 0.8999999761581421% reflects a very new project with only 32 stars -- this is experimental territory. That said, the documentation is thorough, the real-run verification is a genuinely useful claim, and the bilingual coverage fills a gap for Chinese-speaking developers. If you're serious about Claude Code Workflows, this is currently the most comprehensive resource available -- but treat it as a community guide, not official documentation. Wait for more community validation before betting production workflows on it.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.