Make prune-first compression cache-aware by only accepting prune-only compaction when it gets comfortably below threshold. If pruning merely dips under threshold, fall through to the existing summary compaction so we avoid frequent near-threshold recompressions. Tests cover both the conservative fallback and the prune-only fast path. |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| anthropic_adapter.py | ||
| auxiliary_client.py | ||
| context_compressor.py | ||
| display.py | ||
| insights.py | ||
| model_metadata.py | ||
| prompt_builder.py | ||
| prompt_caching.py | ||
| redact.py | ||
| skill_commands.py | ||
| trajectory.py | ||