How AI Native are you?
This public experiment turns local coding-agent history into a score, a band, and a five-axis radar profile for AI-native workflow leverage.
Span, peak depth, usage volume, recent momentum, and repeatability.
Codex, Claude Code, Cursor, and other coding-agent workflows.
Share page for people, skill.md for agents, GitHub for developer backup.
The page you share with people is https://lettokenburn.com/ai-native/. The file agents should execute is https://lettokenburn.com/ai-native/skill.md.
Use this skill to test how AI Native you are. Paste this into Codex, Claude Code, Cursor, or another coding agent. Ask it to open the skill below, run the test locally, submit the result, and report your score, radar profile, and rank. Skill: https://lettokenburn.com/ai-native/skill.md
AI Native Profile
Run the skill to turn a single score into a richer operating profile.
Why this exists
Most public AI discussion still focuses on prompts, models, or benchmark scores. This test is about operational leverage: how much useful work one human input can set in motion.
It is intentionally narrow. It does not try to measure raw engineering ability, taste, or judgment. It asks a simpler workflow question: how naturally can you drive an autonomous coding loop?
That is why the score is now paired with a radar profile. A long average run, a huge one-off burst, a deep usage history, and active weekly momentum are different kinds of AI-native behavior.
The human-facing share page, the raw agent skill, and the GitHub mirror are split on purpose. People should get the explanation first. Agents should get the markdown file. Developers should have a trustworthy backup mirror.
Band Guide
| Avg Turns | Band |
|---|---|
| 60+ | Autonomy Architect |
| 25-59.9 | AI Native |
| 10-24.9 | Flow Operator |
| 3-9.9 | Prompt Pilot |
| 0-2.9 | Manual Driver |