Résumés don't ship. You do.

Take a real challenge in a realistic codebase, use any agent, and earn a score for what you actually understand.

No résumé required. Sign up with GitHub, preview any challenge for free.

Use the agent. Own the outcome.

We don't watch your keystrokes or police your prompts. We care about one thing: can you ship real work in our stack — and explain it.

01

Use any agent you like

We don't watch your keystrokes or police your prompts. Claude, Copilot, whatever ships your best work.

instead of Lockdown browsers and AI-detection

02

Own the outcome

We care about one thing: can you ship real work in our stack — and explain it.

instead of Whiteboard trivia that proves nothing

03

Keep the full report

Every candidate gets their complete score and an evidence-linked report — always.

instead of A silent rejection with no signal back

The gap is the signal.

A passing, AI-assisted submission and a real understanding of it look identical on the surface. The reasoning defense is where they diverge — and an ownership multiplier discounts the score when the answers don't hold up.

  • Why you built it that way — trade-offs, failure modes, the road not taken
  • Asked unaided, one question per turn, time-boxed and scored live
  • Own what you shipped and the score is yours to keep

A real signal you can take anywhere.

Push to your own repo, get scored against a reference expert, and walk away with the full, evidence-linked report — every time.

Your own private repo

Clone the bootstrapped repo and push to your own GitHub — nothing public unless you want it.

GitHub-native feedback as you work

Visible tests guide you while you build; the hidden gateway suite streams its verdict live.

Deterministic pass/fail

Pass/fail is reproducible — no vibes, no recruiter guesswork.

A deep, per-competency analysis

Every point evidence-linked to your own code and anchored to a reference expert — a real comparison, not a curve.

Banded tiers over a raw number

Strong / Promising / Borderline / Weak — the honest signal instead of false precision.

GitHub corroborates, never condemns

Profile analysis can only help your score — it never penalizes you.

Our transparency promise

You always get your full score and a detailed, evidence-linked report — including how you compare to our reference expert. Every rationale is defensible. We grade you the way we'd want to be graded.

Show what you can actually do.

Take a challenge in your stack. Use any agent. Earn a signal that travels.