How to Use RepoHamster
Follow the step-by-step visual road below to pick your challenge, code in GitHub Codespaces, consult the Hamster AI Mentor, and get automatically evaluated.
Select Challenge & Claim Handle
Authenticate with GitHub and select Still & Form
Welcome coder! Sign in with your GitHub account, get your unique developer handle (@yourname), and pick our featured challenge Still & Form!
What You Need To Do:
- 1.Click Start Debugging on the top right to sign in via GitHub or preview with a Project Review Passcode.
- 2.Confirm your assigned developer handle (e.g.
@coder) in the header. - 3.Open Still & Form from the challenge list.
Still & Form: Commerce API & React Storefront
A realistic home-goods storefront built with React 19 and Express 5. Track down cart deduction race conditions, tax calculation discrepancies, and checkout flow bugs.
Start Your First Challenge in Under 60 Seconds
Jump directly into Still & Form. Launch in GitHub Codespaces with zero setup, reproduce the bugs, and ship your solution.
The Complete Beginner's Playbook
Everything you need to know about navigating the platform, running tests, avoiding common beginner pitfalls, and understanding the automated grading sandbox.
1. Prerequisites & Codespaces vs Local Setup
Zero machine requirements with GitHub Codespaces
RepoHamster is designed so you can code directly from any device—even a Chromebook or tablet—using GitHub Codespaces. Every repository includes a .devcontainer configuration that automatically installs the required Node.js LTS, dependencies, and VS Code extensions.
Click “Open in GitHub Codespaces” from the challenge page. No Git CLI or Node install is needed. Your work is saved automatically in the cloud container.
Clone your fork locally:git clone https://github.com/repohamster/Still-Form.gitEnsure you have Node.js 20+ installed, run npm install and npm test.
2. Anatomy of the Still & Form Challenge
What bugs are you tasked with fixing?
Still & Form is a modern React 19 storefront paired with an Express 5 commerce API. It simulates a small-batch home goods business. The repository contains 3 critical problem areas:
When rapid decrement requests are fired concurrently, in-memory cart quantities can drop below 1 or become negative numbers.
Subtotal calculations fail to correctly apply tiered coupon codes when state sales tax rules are evaluated in boundary edge cases.
Submitting the checkout form multiple times with the same cart token should be idempotent and prevent double-charging or duplicate orders.
Pro Hamster Tip: Run Targeted Test Suites
Vitest CommandInstead of running all tests simultaneously every time, you can run `npm run test:backend` to inspect the API only, or `npm run test:business` to inspect discount calculation logic quickly.
3. How the Automated Evaluator Works
Zero human intervention • HMAC-signed webhooks • Isolated sandbox
When you open a Pull Request targeting repohamster/Still-Form:main, here is what happens behind the scenes:
- GitHub Webhook Dispatch: GitHub signs the payload with our secret key and pings RepoHamster at
/api/webhooks/github. - Sandbox Worker Spawn: An isolated test worker clones your exact PR commit SHA in a clean container with stripped environment variables and network egress restrictions.
- Public Tests Execution: The runner executes all 4 public test suites to verify that standard behavior remains intact.
- Authoritative Hidden Tests Execution: The runner injects our proprietary hidden assertions (verifying atomic cart bounds, concurrent checkout tokens, and coupon invariants).
- Live Scorecard & Profile Sync: If your code satisfies 100% of the assertions, the PR status is marked Passed, your scorecard updates with Grade A+, and the badge appears on your profile.
4. Frequently Asked Questions by Beginners
Common questions and troubleshooting
Q: What if I don't have a GitHub Codespaces allowance?
GitHub provides 60 free Codespaces hours every month to all personal GitHub accounts! If you ever run low, you can simply clone the repository with standard git and run it locally.
Q: Can I push multiple commits to improve my score?
Yes! Every new commit pushed to your Pull Request triggers a fresh evaluation run. If your first attempt scores 70/100, read the public test feedback, fix the remaining edge case, push again, and your score will update to 100/100.
Q: Why did the Hamster AI Mentor refuse to write the code for me?
The Hamster AI Mentor is deliberately programmed as a Socratic engineering coach. It asks probing questions, points you to the exact source lines, and explains distributed systems concepts, but it will never generate copy-paste answers. This ensures your evaluation is authentic and respected by hiring engineering teams.