@ha7ch/ainative-rank

How AI-native are you, really? Paste one prompt into Claude Code or Codex and get graded S / A / B / C / D in under 5 minutes.

全球首个 AI native 开发者水平测试 — 粘一条 prompt 到 Claude Code 或 Codex，5 分钟拿到你的等级。

How

Paste this prompt into your IDE of choice:

Run `npx @ha7ch/ainative-rank start` and walk me through it.

Your agent runs the CLI, the CLI orchestrates four short staged prompts (~10–15 minutes), and the server scores how you delegated the work. You get a letter grade — S, A, B, C, or D.

Only S tier (top ~5%) gets invited to Hatch's weekend builder meetups.

Commands

ainative-rank start          # create an anonymous session and emit stage 1
ainative-rank usage          # local 7-day Claude/Codex usage summary for stage 1
ainative-rank next           # advance to the next stage
ainative-rank finish         # upload the transcript, get scored, see the grade
ainative-rank city <name>    # (S-tier only) pick a meetup city

city <name> accepts English ("Shenzhen"), Chinese ("深圳"), or a short id ("sf"). It's only invoked after finish for S-tier candidates.

You don't run these by hand — the prompt above tells your agent which command to run at each step. finish auto-detects your GitHub handle from your local gh CLI auth (set up in stage 4); there's no interactive prompt for it.

Links

• Web: https://rank.ha7ch.com
• Leaderboard: https://github.com/HA7CH/ainative-rank-leaderboard
• Hatch: https://ha7ch.com

License

MIT