About

Method (MVP)

This is an educational self-assessment inspired by the 4D framework (Description, Delegation, Discernment, Diligence) and themes from the Anthropic AI Fluency Index report.

It is an independent coaching tool, not a replication of Anthropic's research methodology or benchmark scores. The MVP uses heuristic signal detection over your prompts, extracts evidence quotes, and scores dimensions with Discernment weighted higher.

Privacy

Transcript mode stores runs in your browser (localStorage). Challenge mode uses server API routes and stores scored runs in .runtime/runs.json so results pages can be reopened.

If an LLM provider key is configured, challenge turn text is sent to that provider for assistant replies. Without a key, the app uses heuristic fallback responses.

What this is NOT

Not a scientifically validated psychometric test
Not a guarantee of work quality or professional performance
Not a replacement for domain expertise and review