catalan-adobe/ai-fluency-assessment
Assess your AI fluency using Anthropic's 4D framework (Dakan, Feller & Anthropic, 2025). Scans Claude Code sessions, runs LLM-based behavior classification on all messages, asks a self-assessment questionnaire for 6 unobservable behaviors, and generates a visual HTML report with scores and actionable feedback. Use when "assess fluency", "AI fluency", "fluency report", "fluency assessment", "4D framework", or "how AI fluent am I".
100
100%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
REPORT-SPEC.mdreferences/
Report Specification
Design system, layout, and formatting for the AI Fluency HTML report.
Design System
- Font: Inter (Google Fonts import) with system fallback
- Background:
#FAFAF9, Surface:#FFFFFF, Border:#E7E5E4 - Competency colors:
- Delegation:
#2563EB(blue), light:#EFF6FF - Description:
#7C3AED(purple), light:#F5F3FF - Discernment:
#0891B2(cyan), light:#ECFEFF - Diligence:
#059669(green), light:#ECFDF5
- Delegation:
- Score colors: 1=
#EF4444, 2=#F97316, 3=#EAB308, 4=#22C55E, 5=#10B981
Report Structure
-
Header: Title, date, overall score (large number with level label)
-
Overall Score Card: Large score with Novice/Emerging/Developing/Proficient/Expert label. Compute as average of 4 competency scores.
-
Key Takeaways: Profile summary ("Your Profile: [type]") with two columns — "What You Do Well" (4 strengths with evidence) and "Where You Can Level Up" (4 growth areas with actionable advice). Include a callout box with one concrete habit to try.
-
4 Competency Summary Cards: Each shows competency name, color-coded score, progress bar, and sub-competency breakdown. Score = average of sub-competency scores. Sub-competency score = average of behavior scores within it.
-
Strengths & Growth Areas: Side-by-side cards showing top 3 strengths (highest-scored behaviors) and top 3 growth areas (lowest-scored behaviors with specific recommendations).
-
Behavior Heatmap: Horizontal bar chart ranking all 18 classified behaviors by LLM match count. Scale bars relative to the highest behavior (highest = 100% width). Show absolute message counts. Color bars by competency color.
-
Top Projects Breakdown: Horizontal bar chart showing message volume per project (top 10), scaled relative to the busiest project.
-
Footer: Framework attribution, generation date, classification method
Heuristic Bar Formatting
- Label: "Detected in" (NOT "Detection rate")
- Value: "N messages" (NOT "N (X%)")
- Bar width: Scale relative to the max behavior count
- Projects: Label as "Strongest in" with just project names (NO percentages)