Cache-friendly context compaction for OpenCode (V1 and V2): append a summary in place, then cut what the model sees down to it.
Version 0.2.0, 14 weekly downloads, rank 1059. Judged from its declared dependencies.
| @opencode-ai/plugin | >=1.17.9 |
|---|---|
| @opencode/plugin | >=2.0.0 |
[](https://rafal-one.github.io/opencode-v2-ready/p/opencode-cache-compact.html)
It ships both implementations.