Work
Case Study
RitualAI
TypeScript CLI that scans local Claude/Codex prompt history and turns repeated workflows into reusable skills.
Overview
RitualAI mines local Claude and Codex prompt history for repeated workflows, then helps convert approved candidates into reusable Agent Skills.
Problem
Developers often repeat useful agent prompts and workflows without noticing which patterns are stable enough to turn into reusable automation.
My role
I built the local scan workflow, clustering-oriented candidate flow, approval step, and SKILL.md generation path.
Technical details
- Scans local prompt history to find recurring work patterns without requiring hosted ingestion.
- Groups candidate workflows for user review before generating reusable skill instructions.
- Guides approved candidates into SKILL.md workflows that can be refined and reused across future agent sessions.
Hard parts and tradeoffs
- Local prompt analysis needs to surface useful patterns without exposing sensitive development history to a remote service.
- The workflow keeps user approval in the loop because not every repeated prompt deserves to become a persistent skill.
Current state
The project is public on GitHub.