testdouble

testdouble / han

Public

Han: AI skills and agents for "Solo" product engineers and small teams

46
3
100% credibility
Found May 17, 2026 at 46 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

Han is a plugin that gives solo product engineers access to 15 specialized AI skills and 21 AI agents that work together like a small team. It runs inside Claude Code, an AI coding assistant. You can use it to plan features through evidence-based interviews, review code with adversarial eyes, investigate bugs with root-cause analysis, maintain documentation, and post detailed pull request reviews—all by calling simple named commands. The skills are designed to compose, so planning feeds into implementation, which feeds into review, which feeds into documentation. It's built by Test Double, a software consulting company, and is open source under the MIT license.

How It Works

1
💡 You realize you need a team

As a solo engineer, you wish you had specialists to bounce ideas off—someone to review your code, help you plan features, and keep your docs up to date.

2
🧩 You add the plugin to your assistant

You connect Han to your AI coding assistant with two simple commands, unlocking a whole suite of specialist skills you can call on whenever you need them.

3
📋 You plan a new feature together

You ask your assistant to plan a feature, and it brings in a project manager agent who interviews you, challenges your assumptions, and builds a solid plan.

4
You pick your next move
🔍
Investigate a bug

Run an investigation skill that digs through your code, finds the root cause, and validates any proposed fix before you commit to it.

👀
Review your code

Have a code review skill examine your changes from multiple angles—security, architecture, testing coverage—and post feedback directly to your pull request.

📝
Update your docs

Keep your documentation fresh with a skill that creates and maintains reference material as your project evolves.

5
🤝 Your specialists work together

The agents hand off to each other seamlessly—planning feeds into review, review feeds into documentation—so nothing falls through the cracks.

🎉 You ship with confidence

You have a feature spec, tested code, thorough review, and updated docs—all from a team of AI specialists that fits in your pocket.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 46 to 46 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 han?

Han is a Claude Code plugin that gives solo engineers and small teams access to a suite of AI-powered skills for planning, reviewing, and documenting their work. Instead of juggling multiple tools or doing heavy planning and review work solo, you invoke named skills like `/plan-a-feature` or `/code-review` that spin up specialist agents to handle judgment-heavy tasks. The system is built in Shell and integrates directly with Claude Code's plugin architecture, letting you chain skills together so a feature spec can flow into an implementation plan, which then flows into code review and finally a PR description.

Why is it gaining traction?

The hook here is the "solo team" fantasy made real. Most AI coding tools are either solo chatbots or require a full team setup. Han brings the team structure to one developer by dispatching multiple agents (project managers, reviewers, architects) that hand off to each other through a shared artifact system. The YAGNI-first philosophy is also refreshing -- it actively pushes back against over-engineering before you commit to a plan. For developers already living in Claude Code, adding `/plugin install han@han` is a low-friction way to get structured workflows without leaving the terminal.

Who should use this?

Solo product engineers who find themselves doing all the planning, review, and documentation work alone will get the most value. Small teams without dedicated PMs or QA will also benefit -- the `/gh-pr-review` skill is particularly useful for teams wanting automated PR feedback without a second pair of eyes. If you're a frontend developer tired of context-switching between planning docs and code review, or a backend dev who wants architectural analysis without calling in a consultant, this is worth a look. Teams already invested in Claude Code will adopt fastest.

Verdict

Han is a well-conceptualized plugin with solid documentation and a coherent skill model, but the 46 stars and 1.0% credibility score signal it's early and unproven at scale. The Test Double backing adds credibility, and the MIT license lowers adoption risk, but you'll be an early adopter. If you're deep in Claude Code and want structured workflows without building them yourself, install it and try `/project-discovery` first to see how it maps your codebase. Hold off on mission-critical use until the community and star count grow.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.