Hire Engineers,
Not Prompt Passers.

CraftScore analyzes GitHub metadata to separate true software architects from candidates who just copy-paste AI generated code. Reduce false positives in your technical hiring.

Test the Engine (Demo)

Paste a public GitHub repo URL to see our AI Collaboration Score in action. Free and instant.

Far beyond the final code. We read the developer's "fingerprint".

ChatGPT can generate clean architecture in seconds. What it cannot do is fake human behavior. CraftScore analyzes Git metadata to separate those who use AI as a tool from those who use it as a crutch.

Impossible Velocity Detector

We cross-reference complex logic volume with typing time. If the candidate built 800 lines of architecture in 2 minutes, our radar lights up.

Trial and Error Radar

We measure the mutation rate (Churn). If the candidate repeatedly deletes and rewrites the same file, we know they are fighting AI hallucinations.

Engineering Maturity

We evaluate traceability. Does the candidate deliver atomic commits with clear messages, or dump everything in one giant commit at the end of the day?

True Collaboration

We detect the terrible blind 'LGTM' (Looks Good To Me), penalizing summary PR approvals with no inline discussion.

🚨 The Vibecoding vs. Copilot X-Ray

Our engine doesn't hate AI, it hates blind outsourcing. We issue automatic alerts on candidate posture:

  • Vibecoding DetectedMassive volumes of logic in very few commits. AI did the work entirely.
  • Push DumpColossal package pushed in a single commit, breaking best practices.
  • Senior with Copilot (Bonus)High velocity, but with granular and atomic commits. AI was used as an assistant, not a crutch.

How it Works

Stop guessing. Get objective data on how candidates interact with AI.

1. Connect

Connect the technical test repository with 1 click using our official GitHub App.

2. Analyze

The candidate takes the test normally (with or without Copilot/Claude). We monitor GitHub metadata in real time.

3. Decide

Receive the automatic AI Collaboration Score, separating architects from those who just copy-pasted generated code.

Simple, Transparent Pricing

Evaluate engineering talent efficiently. Upgrade when you need deeper behavioral insights.

Free

$0/mo

Perfect for individual recruiters or small startups.

  • Base AI Metrics
  • Code Churn & Velocity
  • 5 Evaluations per month
RECOMMENDED

Pro

$49.90/mo

For growing engineering teams and recruitment agencies.

  • Everything in Free
  • Unlimited Evaluations
  • Behavioral Risk Tags (Vibecoding)
  • Invite Team Members
  • Multi-Organization (Workspaces)