Clone_Context_Compressor

Audit a Mind's stream, scratchpad, and boards to identify old or low-salience entries, then non-destructively compress and archive them to reduce LLM token context usage, with hard-math validation of token savings and explicit proof of margin saved.