wcz234

Reduces LLM token consumption by 60-90% on common commands. Filters and compresses command outputs before they reach LLM context. Use when executing shell commands, reading files, running tests, or any operation that produces verbose output. Inspired by RTK…