Context window optimization for AI coding agents. Sandboxes tool output, 98% reduction. 15 platforms
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).