fix(release): restore automatic updates by shipping the speech engine on demand

Auto-update could not work at all: the installer was 189 MB because it carried
the local speech engine and ffmpeg, and the download feed rejects uploads over
about 100 MiB, so update metadata could never be published.

The installer now leaves those components out and the app fetches them the first
time they are needed, verifying every part and the joined archive before
installing. The installer is 90.6 MiB, the update feed is published again, and
updates stay small because the engine is not re-sent on every release.

The fetch is visible and recoverable: the download runs with progress, a failed
install cleans up after itself, and Settings > STT shows the runtime status with
a manual download action for when the automatic one cannot run.

apps/desktop/src/main/update-feed.ts

@@ -16,6 +16,17 @@ export const UPDATE_FEED_URL =
 // GitLab Generic Registry legacy mirror. 2026-08 이전 설치본(0.2.1-alpha)은
 // 이 feed를 폴링하므로, 새 설치자가 Forgejo feed를 내장할 때까지 publisher가
 // 함께 게시한다. 마이그레이션 완료 후 제거 가능. 런타임은 참조하지 않는다.
+/**
+ * 로컬 AI 런타임(사이드카 엔진 / ffmpeg) 배포 위치.
+ * 진을 설치본에 넣으면 installer가 Cloudflare 업로드 한도(100MiB)를 넘어 업데이트
+ * 메타데이터(latest.yml)를 게시할 수 없다. 그래서 런타임은 별도 경로에서 필요할 때 받는다.
+ * 자동 업데이트 채널과 분리된 경로이며 서명이 필요 없다.
+ */
+export const RUNTIME_FEED_URL = UPDATE_FEED_URL.replace(/\/latest$/, '/runtime-latest')
+
+/** 런타임 인덱스 파일명 (feed 루트) */
+export const RUNTIME_INDEX_FILENAME = 'runtime.json'
+
 export const LEGACY_UPDATE_FEED_URL =
   'https://gitlab.twentyoz.kr:8443/api/v4/projects/1172/packages/generic/d3ro-voice/latest'