Favorites Harvester

Favorites Harvester

Use this skill when you need one entrypoint across platforms. It calls the per-platform skills (no RSSHub).

Docker-first:

• See docs/usage.md for docker compose run --rm runner ... examples (no host Python required).

Cookie auth (shared)

• Recommended: CookieCloud env vars (COOKIECLOUD_UUID, COOKIECLOUD_PASSWORD, optional COOKIECLOUD_SERVER_URL). Compatibility: also accepts COOKIECLOUDUUID / COOKIECLOUDPASSWORD.
• Or set platform-specific cookie env vars (BILIBILI_COOKIE, ZHIHU_COOKIES, XIAOHONGSHU_COOKIE)

Optional: export CookieCloud -> env file (no RSSHub)

docker compose run --rm runner python skills/favorites-harvester/scripts/cookiecloud_export_env.py \
  http://cookiecloud:8088 <uuid> <password> --env-file favorites.env

Quick Start

List favorites across platforms:

docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py list --platform all

List only XiaoHongShu saved boards (收藏专辑/收藏夹):

docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py list --platform xiaohongshu --xhs-mode boards

Fetch content by URL (auto-detect platform):

docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py content --url 'https://www.bilibili.com/video/BV...'
docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py content --url 'https://www.zhihu.com/question/.../answer/...'
docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py content --url 'https://www.xiaohongshu.com/explore/<noteId>'

List items in a container:

docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py items --platform bilibili --folder-id <mediaId> --limit 50
docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py items --platform zhihu --collection-id <id> --limit 50
docker compose run --rm runner python skills/favorites-harvester/scripts/favorites_harvester.py items --platform xiaohongshu --board-id <boardId> --limit 50

JSON mode:

• Add --json to list/items/content to emit machine-readable JSON.

How It Works

• Calls these atomic skills by invoking their scripts (prefers uv run if available; falls back to python):
  • skills/bilibili-favorites
  • skills/zhihu-favorites
  • skills/xiaohongshu-favorites
• Use the atomic skill directly when you need advanced flags (e.g. Bilibili subtitle language selection, XHS Playwright options).

Troubleshooting

• XHS failures: run with list ... --xhs-no-headless (visible browser).
• Bilibili transcript failures: the video may have no subtitle track; use skills/media-audio-download + skills/whisper-transcribe-docker (or other Whisper skills) after downloading audio.