skill-tree-ai
v1.0.1
Published
Your AI collaboration style — skill tree visualization with character archetype cards and growth recommendations, grounded in the AI Fluency Framework. MCP server for Claude Code.
Maintainers
Readme
Skill Tree — AI Fluency Profile for Claude
Analyze your Claude collaboration style. Get a character archetype card, skill tree visualization, and personalized growth recommendations — grounded in Anthropic's AI Fluency Framework.
Install
# From this directory
claude plugin install .
# Or from GitHub
claude plugin install github:robertnowell/skill-treeUse
In any Claude Code session:
/skill-treeThis analyzes your conversation history (Claude Code + Cowork), classifies 11 AI collaboration behaviors, and opens a beautiful visualization in your browser.
What You Get
Character Card — Your archetype (e.g., "The Architect"), tagline, signature strengths, and vibe description
Skill Tree — 11 behaviors across 4 branches, with animated bars showing your rate vs. population average:
- Planning: Goal clarity, approach consultation
- Craft: Iteration, examples, format, tone, audience
- Judgment: Context gap detection, reasoning challenges
- Rigor: Fact verification
Growth Quest — One specific, actionable challenge for your next session, injected via SessionStart hook
Requirements
- Python 3.10+
anthropicPython SDK (pip3 install anthropic)ANTHROPIC_API_KEYenvironment variable
How It Works
- Scans
~/.claude/projects/and Cowork local sessions for conversation history - Extracts clean user messages (filters out tool results, pasted output, system messages)
- Classifies 11 behaviors per session using Claude Haiku (~$0.005/session, cached)
- Aggregates into a skill profile with population baseline comparisons
- Generates a character archetype card
- Renders a self-contained HTML visualization
Research Basis
The 11 behaviors come from Anthropic's AI Fluency Index (Feb 2026), which analyzed 9,830 conversations using the 4D AI Fluency Framework (Delegation, Description, Discernment, Diligence). Population baselines are from that study.
License
MIT
