[RealizmModz]

Problem

The TesseractOCR 5.5.2 update builds and passes Windows tests, but the production Linux parser remains unready. Docker still creates the 5.3-era native aliases, while the upgraded wrapper loads leptonica-1.85.0.dll and tesseract55.dll.

Changes

• Align Linux library aliases with TesseractOCR 5.5.2 in both API/worker package upgrades.
• Run actual OCR against the fixed synthetic readiness image during the parser image build; fail the image build if native loading, model initialization, or recognition fails. Emit pass/fail only. Production readiness continues to exercise the cgroup-contained process path.
• Refresh pinned .NET SDK/runtime images from chore(deps): bump dotnet/aspnet from 222759b to d563377 #700/chore(deps): bump dotnet/sdk from e70cdb7 to 0d99ebf #701, test SDK 18.11.0, and QRCoder 1.8.1 from Bump the non-major-updates group with 3 updates #725.

Validation

Exact-head CI must prove backend/OCR tests, all three production images, parser readiness, resource containment, real OCR, encrypted backup and isolated restore. No production deployment.

Supersedes dependency proposals #697, #700, #701 and #725 once this passes and merges.