Protect HTML literal contents during translation retries

Make structural retries sentinel-free

Remap local sentinels after structural retries

Preserve sentinels during structural retries

Restore ordinary attention before smoke parity check

Align continuous batching smoke generation settings

Load TranslateGemma text model for continuous batching

Document local preflight runtime requirements

Install exact Transformers runtime dependencies before translation

Align Hub dependency and add local runtime preflight

Import TranslateGemma auto classes from source-compatible module

Normalize GitHub slugs at clone boundary

Mark bootstrap submitter as authoritative Job path

Bootstrap runner snapshot inside Jobs

Use fresh Job config path to avoid stale mount entry

Fix immutable model revision diagnostics and runner pin

Use python3 in Jobs container command

Serialize Job status safely

Pass GitHub token securely to publishing Jobs

Use existing stevhliu fork for source PRs

Document staged Job gates

Add immutable Job launcher

Pin runner and Job volumes

Pin model and doc-builder revisions

Add source-preserving Japanese documentation runner

initial commit