prompt-compressor

Compress verbose prompts & context before LLM processing. This skill should be used when input exceeds 1500 tokens, contains redundant phrasing, or includes unnecessary context. Reduces tokens by 40-60%.