I kept noticing I'd ask Claude Code to write things, then rewrite them, then ask again with slightly different wording. Months in, I couldn't tell whether that loop was refinement or just repetition dressed up as iteration.
Anthropomorphic published a study in February across 9,830 Claude conversations, classifying 11 observable collaboration behaviors. I wanted to know which ones showed up in my sessions — and which ones I never touched.
So I built skill-tree: it reads your Claude Code or Cowork session history, classifies those same 11 behaviors against Anthropic's AI Fluency Index baseline, and assigns you one of seven archetype cards rendered as tarot cards with curated museum art. There's also a growth quest — one behavior you haven't used, surfaced at the start of your next session via a SessionStart hook.
The archetypes (Illuminator, Navigator, Alchemist, and four others) aren't cosmetic. They map to three axes from Dakan and Feller's 4D AI Fluency Framework: Description (how well you frame problems), Discernment (how you evaluate outputs), Delegation (how you structure work handoffs). The fourth axis, Diligence, doesn't appear in chat logs so skill-tree leaves it unscored.
The card renders live — see the Illuminator at skill-tree-ai.fly.dev/fixture/illuminator. Each analysis runs a 7-step orchestration (find session files, extract messages, remote classifier on Claude Haiku via Fly.io, archetype assignment, narrative synthesis, render, return URL) in roughly 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
For Cowork, drop in skill-tree-ai.zip. If you're on Cursor, VS Code, or Windsurf, the MCP server is npm install skill-tree-ai.
Growth quest state persists in ~/.skill-tree/ for Claude Code (Cowork uses $CLAUDE_PLUGIN_ROOT/.user-state/ because its $HOME is ephemeral).
Top comments (0)