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
Quality
100%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Securityby
Passed
No known issues
FRAMEWORK.mdreferences/
AI Fluency Framework Reference
24 behaviors across 4 competencies, from Anthropic's 4D AI Fluency Framework (Dakan, Feller & Anthropic, 2025).
| # | Behavior | Competency | Classification |
|---|---|---|---|
| 1 | Clarifies goal before asking for help | Delegation | LLM |
| 2 | Understands problem scope and nature | Delegation | LLM |
| 3 | Assesses AI fit | Delegation | Questionnaire |
| 4 | Selects platform | Delegation | Questionnaire |
| 5 | Consults AI on approach before execution | Delegation | LLM |
| 6 | Distributes work strategically | Delegation | LLM |
| 7 | Specifies format and structure needed | Description | LLM |
| 8 | Defines audience for the output | Description | LLM |
| 9 | Provides examples of what good looks like | Description | LLM |
| 10 | Iterates and refines | Description | LLM |
| 11 | Sets interaction mode | Description | LLM |
| 12 | Communicates tone and style preferences | Description | LLM |
| 13 | Checks facts and claims that matter | Discernment | LLM |
| 14 | Identifies when AI might be missing context | Discernment | LLM |
| 15 | Questions when AI reasoning doesn't hold up | Discernment | LLM |
| 16 | Detects hallucination | Discernment | LLM |
| 17 | Evaluates tone and communication fit | Discernment | LLM |
| 18 | Recognizes bias in AI outputs | Discernment | Questionnaire |
| 19 | Chooses AI tools ethically | Diligence | Questionnaire |
| 20 | Maintains ethical awareness during interaction | Diligence | Questionnaire |
| 21 | Discloses AI involvement to stakeholders | Diligence | LLM |
| 22 | Represents AI contribution accurately | Diligence | LLM |
| 23 | Verifies and tests outputs before sharing | Diligence | LLM |
| 24 | Takes ongoing accountability | Diligence | Questionnaire |
Troubleshooting
| Problem | Fix |
|---|---|
No module named 'anthropic' | Install: uv pip install anthropic. Or use --regex-only for free approximation. |
No such file or directory: assess.py | Use the full path to assess.py in the skills directory |
| 0 Claude Code messages found | Check that the sessions directory exists and has JSONL files |
| API authentication error | Set ANTHROPIC_API_KEY or ANTHROPIC_FOUNDRY_API_KEY env var |
| Questionnaire hangs | Don't run interactive mode; ask questions in chat and save programmatically |
| Bars look like bad scores | Use relative scaling (max behavior = 100% bar width) and absolute counts |