DEV Community

J Now
J Now

Posted on

I didn't know which AI collaboration habits I was skipping

After months of daily Claude Code sessions, I noticed I couldn't answer a basic question: am I actually improving, or just moving faster through the same four or five patterns? Faster isn't better if you're optimizing a rut.

Anthropic published a study in February (9,830 Claude conversations, 11 observable collaboration behaviors) that gives a population baseline. I wanted to see where my own sessions landed against it — and more importantly, which behaviors I never touched.

So I built skill-tree: a Claude Code plugin that reads your session history, classifies all 11 behaviors from the AI Fluency Index against your actual messages, and returns a stable URL with your results.

The output is a tarot-style archetype card — seven possible archetypes (Illuminator, Navigator, Alchemist, and four others), each rendered with curated museum art alongside a skill radar showing your behavior distribution across three axes from Dakan & Feller's 4D AI Fluency Framework: Description, Discernment, and Delegation. The fourth axis (Diligence) isn't recoverable from chat logs alone, so it's excluded rather than fabricated.

The growth quest piece is what I use it for day-to-day. After classification, the tool picks one behavior from the 11 that you haven't exercised recently and surfaces it as a prompt for your next session. It persists via a SessionStart hook in ~/.skill-tree/ so Claude Code surfaces it automatically when you open a new session.

Installation:

claude plugin marketplace add robertnowell/ai-fluency-skill-cards
claude plugin install skill-tree-ai@ai-fluency-skill-cards
Enter fullscreen mode Exit fullscreen mode

You can see a live fixture at skill-tree-ai.fly.dev/fixture/illuminator before running it on your own sessions. The full classification pipeline runs in about 30–60 seconds end-to-end: find session files, extract user messages, remote classifier (Claude Haiku on Fly.io), archetype assignment, narrative synthesis, render, return URL.

Also works in Cowork via skill-tree-ai.zip, and there's an MCP server (npm install skill-tree-ai) if you're on Cursor, VS Code, or Windsurf.

https://github.com/robertnowell/ai-fluency-skill-cards

Top comments (0)