Part of an experiment
This is the TDD-style pipeline build — test-first, separated personas (red → green → review). The methodical, ceremony-heavy approach. See the video, blog post, and source code for the full study.
Blog post Video Source

CLAUDE.md Scorer

Evaluate your AI instruction files for quality, completeness, and best practices

Upload Your CLAUDE.md File

📄
Drop your CLAUDE.md file here
or click to browse

Try a Sample

2 lines
Minimal Rules
A basic CLAUDE.md with minimal structure and conventions. Good for seeing how lack of detail affects scoring.
57 lines
Agent Guidelines
Well-structured agent conventions with clear sections and actionable rules. Shows good practices in action.
459 lines
Comprehensive Setup
Detailed agent configuration with extensive conventions. Demonstrates comprehensive but potentially complex documentation.

Scoring Report

A
Overall Score
9.2

Improvement Suggestions

Live Evaluation

Built by Dory Zidon — AI Will Replace You
doryzidon.com YouTube (@aiwillreplaceyou) LinkedIn (/in/doryzidon)