Andon
View on GitHubA Lean quality system for AI-assisted work — every defect becomes a permanent countermeasure. Built by a manufacturing operator.
A dependency-free operating kit for AI-assisted development, with project memory templates, agent roles, commands, and Python hooks. Its claim-check hook prompts for verification when an agent says work is done; it is a forcing function, not proof that work was verified.
Use Cases
Load durable project memory at the start of an agent sessionFlag unverified completion claimsRecord recurring agent defects and track countermeasuresCoordinate parallel agent sessions with lanes and shared work protocolsProvide reusable agent roles and slash commands
Built With
- Language
- Python
- Frameworks
- Claude Code
Tags
AI coding agents · agent memory · Claude Code · hooks · verification · defect tracking · prompt engineering · Lean · agent workflows