Matt Beane

Bidirectional claim-evidence traceability for qualitative research

Understudy: an AI that learns to write in your voice. Method + reference implementation, no private data.

Turn dormant research data into submittable papers using AI-assisted analysis

Prototype: Web-based visualization for Beads issue tracking

Anti-skills for Claude. You already have these skills. Now they're documented.

The 4F Method: Four-Function Fractal software development. Recursive specification → bottom-up implementation.

Auto-click permission prompts in Claude Code. Stop babysitting your AI assistant.

Extract defensible value from professional backlogs

Qualitative content evaluation toolkit - discover, measure, and calibrate quality dimensions

When Claude tics, Claude pays. Public chalkboard for overused Claude phrases — Stop hook detects, forces Claude to write lines, posts to a Bart-style chalkboard with a leaderboard.

Monte Carlo LLM workflow analyzer — the prototype from the Measure Twice, Spend Once workshop

LLM family attribution for academic text — identify which frontier model (OpenAI, Anthropic, Google, Meta) generated a given document

Analysis of publicly available AI policies for classes