This skill provides Claude with specific guidance on implementing 'belt-and-suspenders' logic without the risks of data corruption, churn, or infinite loops. It focuses on ensuring that fallback paths, reconciliation loops, and repair hooks are safe to execute multiple times by enforcing patterns like atomic writes, existence checks, and specific integration flags. By following these implementation patterns, developers can build highly reliable systems that recover from failures gracefully while maintaining data integrity across distributed environments.
Características Principales
01Prevention of recursive repair triggers
02Safe fallback path implementation
03Braintrust integration via merge flags
04Atomic file operation patterns
05Reconciliation loop safety checks
060 GitHub stars
