headroom-litellm-proxy Changelog

Updated	Change
2026-07-07 16:07	v1.1 — Removed --lossless (blocks ML compression via OpenAI endpoints); added --target-ratio 0.5 and --intercept-tool-results for effective compression; added Compression Tuning section; added Flags NOT to Use section (lossless, memory, learn); expanded health verification checks; expanded stats key fields; added compression-specific troubleshooting entries; updated memory range
2026-07-06 21:46	v1.0 — Initial skill; split from goose-headroom-provider