Agentskin

AgentSkin is an open-source Model Context Protocol (MCP) server built for AI assistants and agents. It exposes tools for recursive, high-efficiency data pruning: transforming bloated or inconsistent API/web payloads into token-optimized, context-rich 'skins' using the Semantic Shorthand Standard (SSS). The two main tools are 'fetch_optimized_data' (fetch and convert API/web data into a dense, unified Markdown skin with custom key mapping) and 'skin_reasoning' (removes linguistic noise from text for optimal LLM input). All processing is local to the host, maintaining privacy and performance. Ideal users include developers building or orchestrating LLM-powered agents and any organization seeking to optimize LLM token usage with clean, unified data feeds.