prompt-compressor

name	prompt-compressor
description	Compress long prompts and contexts while preserving semantic meaning — reduce token count without losing information.
version	0.1.0
publisher	acmecorp
handler_language	typescript
handler_file	handler.ts
when-to-use	Use when the user has a long prompt or context that needs to be shorter, or when hitting token limits.
inputs_schema	{"type":"object","properties":{"action":{"type":"string","enum":["compress","decompress_plan","benchmark"]},"text":{"type":"string"},"target_ratio":{"type":"number","description":"Target compression ratio (0.0-1.0). E.g. 0.5 means halve the token count."},"preserve_code":{"type":"boolean"}},"required":["action","text"]}
outputs_schema	{"type":"object","properties":{"result":{"type":"object"}}}

prompt-compressor

Compresses long contexts while preserving semantic information. Uses multiple techniques: abbreviation, structure extraction, key-phrase retention, and whitespace optimization.