I've been using Claude Code daily for months and realized I had no idea whether I was actually improving or just running the same three prompting habits faster. Anthropic published a study in February 2026 measuring 11 observable collaboration behaviors across 9,830 Claude conversations. I wanted to know what mine looked like against that baseline.
skill-tree analyzes your Claude Code or Cowork session history, classifies those same 11 behaviors, and assigns one of seven archetype cards — rendered as tarot cards with curated museum art — plus a skill radar showing where your behavior clusters. A live example: skill-tree-ai.fly.dev/fixture/illuminator.
The behavior taxonomy comes from Dakan & Feller's 4D AI Fluency Framework, which breaks collaboration down across Description, Discernment, Delegation, and Diligence axes. Diligence doesn't show up in chat logs, so the classifier works across the other three. The archetype assignment isn't cosmetic: it reflects which of those axes dominate your actual session data, not a quiz.
The part I found most useful: after classification, it picks one behavior you haven't touched and assigns it as a growth quest persisted via a SessionStart hook — so the next time you open Claude Code, it surfaces that quest at the start of the session. The full pipeline (find session files → extract user messages → remote classifier → archetype assignment → narrative synthesis → render → return URL) runs in 30–60 seconds.
Install in Claude Code:
claude plugin marketplace add robertnowell/ai-fluency-skill-cards
claude plugin install skill-tree-ai@ai-fluency-skill-cards
Also available as an MCP server (npm install skill-tree-ai) for Cursor, VS Code, and Windsurf.
Top comments (0)