lawrencewzen

lawrencewzen / hero-coding

Public

Minimal autonomous coding agent MVP — drop user stories in inbox/, get git commits out. Built on top of pi-coding-agent.

17
1
100% credibility
Found May 01, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

hero-coding is an autonomous agent that takes markdown user stories from an inbox folder, implements them via AI-driven code changes and git commits in a target repository, and uses an AI judge to verify and retry until acceptance criteria are met.

How It Works

1
🔍 Discover Hero Coding

You find this handy tool on GitHub that turns simple task descriptions into real code improvements in your project.

2
📥 Set It Up

Download and prepare the tool on your computer so it's ready to help with your coding tasks.

3
🔗 Connect Your Project

Tell the tool where your code project lives and link it to a smart AI service that does the thinking.

4
📝 Write a Task Story

Create a short note describing what you want done, like adding a feature or fixing something, with clear goals.

5
📂 Drop In Your Task

Place your task note into the special incoming folder and start the watcher to see the magic begin.

6
⚙️ Watch It Work

The tool reads your task, makes careful changes to your code, checks everything works, and tries again if needed until it's perfect.

✅ Task Complete

Your project now has the new code safely added, with a record of what happened, and the task moves to the done folder.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 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 hero-coding?

Hero-coding is a minimal TypeScript autonomous coding agent that turns markdown user stories dropped into an inbox folder into atomic git commits on a target repo. You write stories with frontmatter specifying goals, acceptance criteria, constraints, and retry limits; it processes them via `npm run watch` or `hero story `, using OpenAI-compatible models for coding and review. All state lives in git and filesystem, solving the pain of manual bug fixes or small features by automating the edit-commit-judge-retry loop.

Why is it gaining traction?

Its dead-simple setup—no complex servers, just .env tweaks for worker and judge models—stands out in the hero coding space, especially as a lightweight MVP built atop pi-coding-agent. Developers dig the stateless design, real-time inbox watching, and detailed run stats in JSON logs, making it a quick way to test AI-driven coding without bloat. Hooks like auto-rescue commits and loop detection keep runs reliable, perfect for minimal GitHub actions workflows or hero coding experiments.

Who should use it?

Solo devs or small teams handling repetitive TypeScript fixes, like adding params to utils or writing tests, without full CI pipelines. Indie hackers prototyping features in a target repo, or educators running campus coding hero workshops with student stories. Avoid if you need production-scale; it's for quick wins on personal repos.

Verdict

At 17 stars and 1.0% credibility, this early-stage gem has solid docs and examples but lacks tests—try it for hero coding 2025 proofs-of-concept if you're into minimal autonomous agents. Worth forking for custom tweaks, but wait for v2 stability before betting the farm.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.