Use when designing or reviewing terse, token-efficient communication skills, compressed response modes, or persona skills that compress output while preserving technical accuracy.
