视频瘦身 · Video Shrink
把一个或多个视频文件压缩成体积尽可能小、肉眼看不出明显损失的 MP4。默认策略是
libx265 CRF 28、保留分辨率与帧率、AAC 96k,输出 <原名>-compressed.mp4 放在原
文件旁边,原文件不动。质量档、目标体积、VMAF 门禁和替换模式都是可选开关。
Triggers
Activate when
- “把这个视频压小一点,发微信用。”
- “这条 4K 素材压到 1080p,尽量小。”
- “压缩到 50M 以内。”
- “压完直接替换原文件。”
- “Compress this video as small as possible without visible quality loss.”
- “Shrink these MP4s to fit in an email.”
Do not activate when
- 只是问 HEVC、H.264、ProRes 的区别或该选哪个 → 直接回答,不跑编码
- 要先找出电脑上有哪些视频 →
lov-list-videos - 要剪辑、加字幕、拼接、调色 →
lov-media-creator - 要对课程实录做增强与分段 →
lov-media-preprocessor - 要压缩图片 →
baoyu-compress-image
User Profile (cross-session)
Every generated Skill is connected to the shared user-profile/v1 contract in
skill.yaml. Read the shared user, brand, workspace, preferences, and this
Skill's skills.<skill_id> namespace at the start of every run. Keep the source
portable: resolved personal values belong in the shared profile, never here.
When the user directly states a durable preference or brand fact, persist it
through scripts/profile_store.py and report the saved profile path. Put
Skill-specific values under records.<field>; use brand.<field> or
user.<field> for shared values. Do not persist inferred secrets or credentials.
See references/user-profile.md for the complete contract.
本 Skill 读取的长期记录(都可选):
| 记录 | 含义 |
|---|---|
records.default_quality |
smallest / small(默认)/ balanced / high |
records.default_codec |
hevc(默认)/ h264 / av1 |
records.default_preset |
x264/x265 preset,默认按质量档选 |
records.audio_bitrate |
AAC 码率 kbit/s,默认按质量档选 |
records.max_edge |
长边像素上限,默认不缩放 |
records.min_vmaf |
画质门禁分数,设置后每次都会测 VMAF 并按需重编 |
records.output_dir |
固定输出目录,默认放原文件旁边 |
records.replace_original |
true 时默认替换原文件;只有用户明确说过才可写入 |
Skill Group Composition
Read references/skill-composition.md before deciding whether to invoke or
extend any adjacent capability. The record distinguishes optional upstream and
downstream handoffs from embedded Kit modules. Do not silently depend on a
sibling Skill that is not shipped with this source.
Workflow (MANDATORY)
You MUST follow these steps in order.
Step 0: Resolve skill root, dependencies, and runtime context
- Use
SKILL_DIRif the environment provides it. - Otherwise infer the installed skill directory from the current skill context.
- 必须确认
scripts/compress_video.py、scripts/profile_store.py、references/encoding-guide.md、references/skill-composition.md均存在。 ffmpeg -encoders里必须有libx265;缺libvmaf时不能用--vmaf/--min-vmaf,缺libsvtav1时不能用--codec av1。- If a required resource is missing, name its expected relative path and stop before producing a partial result.
When running scripts manually:
export SKILL_DIR="/path/to/lov-compress-video"
Resolve context.profile on every invocation. The precedence is current request,
project context, Skill-specific profile records, shared preferences, shared
brand/user profile, then safe defaults. A direct user statement about a durable
preference or brand fact should be saved with scripts/profile_store.py record
using --confirm, followed by a concise saved-path report.
Step 1: Understand the requested outcome
- 确认输入文件路径存在,多文件用
--output-dir或替换模式。 - 把用户话语映射到参数,不要先跑一遍再问:
- “尽量小 / 默认” → 不加参数(
small:CRF 28) - “画质要好一点” →
--quality balanced或high - “越小越好,画质无所谓” →
--quality smallest(CRF 32、slow、长边 1080、64k 音频) - “压到 XX M 以内” →
--target-size XXM(两遍码率模式) - “要发给 Windows 老电脑 / 网页播放” →
--codec h264 - “快点,别等太久” →
--fast(VideoToolbox,体积大约多两成,画质分更低) - “压完替换原文件” →
--replace;原文件进废纸篓,--permanent才直接删除
- “尽量小 / 默认” → 不加参数(
- 用户没有明确说替换时,一律保留原文件。替换是高影响操作,需要用户本轮明确
授权或 Profile 里由用户亲口设定的
replace_original。 - 用户在意画质或素材珍贵时加
--min-vmaf 90:脚本会测分并在不达标时降 CRF 重编,最多两次。VMAF 测量约等于再解码一遍两条视频,长片要先告知耗时。
Step 1.5: Analyze nearby Skills before implementation
- 先读
references/skill-composition.md。批量场景的上游是lov-list-videos(先列出候选再压),下游是lov-media-publisher(压完上传)。 - 不要在本 Skill 里做剪辑、字幕或增强,那些各有归属。
Step 2: Execute the workflow
- 先看计划再跑(大文件尤其如此):
python3 "$SKILL_DIR/scripts/compress_video.py" input.mov --dry-run
- 默认压缩,输出到原文件旁边:
python3 "$SKILL_DIR/scripts/compress_video.py" input.mov
- 常见变体:
# 画质门禁:VMAF 不到 90 就降 CRF 重编
python3 "$SKILL_DIR/scripts/compress_video.py" input.mov --min-vmaf 90
# 目标体积 50 MB(两遍编码)
python3 "$SKILL_DIR/scripts/compress_video.py" input.mov --target-size 50M
# 4K 压成 1080p、30fps、H.264,给兼容性优先的场景
python3 "$SKILL_DIR/scripts/compress_video.py" input.mov --max-edge 1080 --fps 30 --codec h264
# 批量:输出到一个目录,JSON 结果供后续统计
python3 "$SKILL_DIR/scripts/compress_video.py" *.MP4 --output-dir ~/compressed --json
# 替换模式:写 <原名>.mp4,原文件移到废纸篓
python3 "$SKILL_DIR/scripts/compress_video.py" input.mov --replace
进度显示在 stderr:百分比、速度倍数与 ETA。长片可以直接告诉用户预计时长 (1080p 在 Apple Silicon 上 libx265 medium 约 3 到 6 倍速,即 1 小时视频 10 到 20 分钟)。
用户说出长期偏好时写回 Profile,例如“以后都用 h264”:
python3 "$SKILL_DIR/scripts/profile_store.py" record \
--skill-id lov-compress-video \
--path records.default_codec \
--value '"h264"' \
--confirm
Step 3: Validate the deliverable
- 脚本自带门禁:输出时长与源相差超过 0.5 秒或 1% 会丢弃输出并报
duration-mismatch;输出不比源小会丢弃并报skipped-larger(--force保留)。 两种情况下替换模式都不会动原文件。 - 回读 stderr 摘要:源大小、结果大小、百分比、节省量、用时、VMAF(若测)。 报告给用户时给出这几个数字和输出路径;替换模式还要说明原文件去了废纸篓。
- 输出文件用 ffprobe 复核编码、分辨率、时长;
--json里已包含这些字段。 - 编码失败时保留 ffmpeg 最后 15 行错误,不重试相同命令;常见原因是源文件损坏、
4:2:2 ProRes 未被 VideoToolbox 支持(去掉
--fast)或磁盘空间不足。 - Validate
skill-card.yaml,cases/cases.json, andpricing-card.yamlas the standard trust bundle for this Skill.
Dependencies
- Python 3.9+(仅标准库)
- FFmpeg 与 FFprobe,含 libx265;
--codec h264需 libx264,--codec av1需 libsvtav1,--vmaf/--min-vmaf需 libvmaf;Homebrew 的 ffmpeg 默认全部包含 --fast与废纸篓式--replace依赖 macOS(VideoToolbox、Finder)- 无外部 Skill 依赖;交接见
references/skill-composition.md
