I'm hikariming, the developer behind ghfind. The backstory: A few months ago I was evaluating candidates for a project and kept running into GitHub profiles that looked incredible on the surface — thousands of stars, dark-green contribution graphs, hundreds of followers. But when I actually looked at the repos, most of the stars were on forks, the contributions were one-line typo fixes across repos they didn't own, and the followers had the unmistakable curve of a follow-back loop. I realized GitHub's built-in metrics — stars, followers, contribution count — measure popularity, not contribution value. And popularity is cheap to manufacture. So I built ghfind. It scores any GitHub account from 0–100 across six weighted dimensions: account maturity, original project quality, contribution quality, ecosystem impact, community influence, and activity authenticity. The engine is fully deterministic — same inputs, same score, every time. No AI touches the number. The scoring core is open source under AGPL, ported line-for-line from Python to TypeScript with unit tests locking the outputs in parity. A few things you might find interesting: Out of ~19,000 scored accounts, fewer than 1% tripped the strongest red flags. Most developers are genuine. But the ones that aren't are spectacularly not. Stars have almost zero correlation with contribution quality. I've seen accounts with 8,000+ stars and contribution scores in the bottom quartile — all forks, no original work. The optional AI roast is the part everyone tries first. It's brutal. You've been warned. You can bring your own LLM key (stays in your browser localStorage, never hits the server), use the REST API, the CLI (ghfind scan), or plug it into any MCP-compatible agent. What I'd love feedback on: Scoring edge cases — if your score feels off, tell me. The engine gets smarter with every real-world data point. Language leaderboards — which languages should I add next? Anything broken, confusing, or just not fun. Try it on yourself first (everyone does): ghfind.com Then try some usernames you admire — and some you're suspicious of. That's where it gets fun. Happy to answer any questions. Really excited to hear what you all think! 🔥

ghfind

Discover the best developers — and become one.

Discover the best developers — and become one.