Months of daily Claude Code sessions, and I had no reliable signal on whether my collaboration patterns were improving or just calcifying. I was faster, probably. But faster at what?
Anthropic published a study in February 2026 measuring 11 observable collaboration behaviors across 9,830 Claude conversations. I wanted to run that same classification on my own session history and see which behaviors I never touch.
So I built skill-tree: it reads your Claude Code or Cowork session files, classifies those 11 behaviors from Anthropic's AI Fluency Index, assigns one of seven archetype cards (rendered as tarot cards with curated museum art, see live example at skill-tree-ai.fly.dev/fixture/illuminator), and surfaces one behavior you haven't tried as a growth quest for your next session.
The orchestration is 7 steps — find session files, extract user messages, remote classifier (Claude Haiku on Fly.io), archetype assignment, narrative synthesis, render, return URL — and runs in 30–60 seconds end-to-end.
The interesting design problem was persistence. Growth quests need to survive across sessions, which is straightforward in Claude Code: write state to ~/.skill-tree/ and it's there next time. Cowork is different. Its $HOME is ephemeral — anything you write there is gone when the session ends. So the tool splits paths: Claude Code uses ~/.skill-tree/, Cowork uses $CLAUDE_PLUGIN_ROOT/.user-state/ because that volume persists across sessions even though $HOME doesn't.
One path per client, both resolve to the same data structure. The SessionStart hook reads from whichever path is populated and injects the active quest into the session context before you type anything.
Install in Claude Code:
claude plugin marketplace add robertnowell/ai-fluency-skill-cards
claude plugin install skill-tree-ai@ai-fluency-skill-cards
For Cowork, install via skill-tree-ai.zip. Also available as an MCP server (npm install skill-tree-ai) for Cursor, VS Code, and Windsurf.
The behavior taxonomy comes from Dakan & Feller's 4D AI Fluency Framework — Description, Discernment, Delegation are the three axes visible in chat logs. Diligence is the fourth axis but doesn't surface in conversation data.
Top comments (0)