Scores conversational turns and tool outputs by semantic salience to prune low-value tokens within context budgets
mcp cost-optimization python-standard-library prompt-compression context-compression agent-skills llm-efficiency memory-pruning salience-pruner token-budget-optimizer context-window-manager
-
Updated
Oct 12, 2026 - Python