Preserve semantic reader after failed generation rebuild

SKILLS.md

@@ -21,8 +21,10 @@ searchable content or hide slow first inputs behind warm percentiles.
 ## Runtime and setup
 
 Use tools/embeddinggemma2/model-manifest.json and its pinned runtime builder.
-Build/stage CPU workers for VM tests. Production acceleration requires separate
-hardware evidence and reliable CPU fallback. Verify size/SHA-256, cache reuse,
+Normal distribution builds and staging use `auto`, with GPU support and CPU
+fallback in the same runtime. Force `SEARCHI_FORCE_CPU=1` for every Windows VM
+inference test; explicit CPU builds are internal test tools only. Production
+acceleration requires separate hardware evidence. Verify size/SHA-256, cache reuse,
 interrupted downloads, errors, and retries. Keep inference local. Never substitute
 MiniLM, simulated vectors, or hosted inference. Ship the worker with the app.