Video Stitcher

视频片段拼接和后期处理。输入视频片段列表，输出完整视频。支持转场效果、背景音乐、字幕叠加。底层使用 FFmpeg 或 Remotion。触发词：拼接视频、合并视频、视频剪辑、video stitch、concatenate videos、add transitions。

Security Scan

The skill's instructions match a video-stitching purpose, but there are mismatches (it expects FFmpeg/Remotion but declares no required binaries) and some operational vagueness that could lead to unexpected file/network access or resource use.

Purpose & Capability

The SKILL.md clearly describes video stitching using FFmpeg or Remotion which is coherent with the skill name/description. However the registry metadata lists no required binaries or runtime dependencies even though the instructions assume FFmpeg (and optionally Remotion/Node). The absence of declared runtime requirements is an inconsistency: a user installing this skill would reasonably need FFmpeg and/or a Node/Remotion environment.

Instruction Scope

The runtime instructions stay within the stated purpose: validating clip accessibility, normalizing formats/resolutions, building ffmpeg filter graphs, and running rendering. They reference creating files (e.g., list.txt), reading local paths and URLs, and executing ffmpeg commands. These are expected for a video-stitcher. Note: steps like "verify all clips are accessible" and accepting clips as URLs are somewhat vague and grant the agent discretion to fetch remote resources and access local files—this is operationally normal but should be acknowledged by an operator.

Install Mechanism

This is an instruction-only skill with no install spec and no code files — lowest install risk. No external downloads or install actions are specified in the metadata or SKILL.md.

Credentials

The skill declares no required environment variables or credentials, which fits its offline media-processing purpose. But it implicitly requires system binaries (FFmpeg) and possibly Node/Remotion; those runtime needs are not declared. That mismatch (no 'required binaries' listed while examples call ffmpeg and Remotion) reduces proportionality transparency and may surprise users or operators.

Persistence & Privilege

The skill does not request 'always' presence and does not modify other skills or system-wide configuration. Autonomous invocation is allowed (platform default), which is reasonable for a user-invocable tool of this type.

What to consider before installing

Before installing or enabling this skill, note the following: (1) it expects FFmpeg (and possibly a Node/Remotion environment) even though the registry metadata lists no required binaries—ensure those runtimes are installed and sandboxed if needed; (2) the skill's runtime steps accept local file paths and URLs and run ffmpeg commands, so avoid passing untrusted remote URLs or arbitrary local paths (these could cause unexpected network fetches or process/file access); (3) video rendering is CPU/disk intensive—run large renders in controlled environments and check resource limits; (4) review any ffmpeg command lines generated before execution to avoid unsafe command injection if user-supplied inputs are concatenated into shell commands; and (5) if you need stricter assurance, ask the publisher to update metadata to declare required binaries and provide a minimal test script or provenance (homepage/source) so you can verify expected behavior. Because of the metadata/instruction mismatch, proceed cautiously or run the skill in an isolated/sandboxed environment.

视频拼接

功能

1. 拼接：多个视频片段 → 完整视频
2. 转场：添加片段间过渡效果
3. 音频：背景音乐、音量调整
4. 字幕：叠加字幕/文字
5. 导出：多种格式和分辨率

工具选择

工具	适用场景
FFmpeg	简单拼接、快速处理、命令行
Remotion	复杂动效、React 组件、品牌化

默认使用 FFmpeg。复杂需求用 Remotion。

输入参数

参数	必填	说明
clips	✓	视频片段列表（路径或URL）
output	✓	输出路径
transition	-	转场类型
transition_duration	-	转场时长（默认0.5s）
bgm	-	背景音乐路径
bgm_volume	-	BGM音量（0-1，默认0.3）
subtitles	-	字幕文件（SRT/ASS）
resolution	-	输出分辨率
fps	-	帧率（默认30）
format	-	输出格式（mp4/mov/webm）

输入格式

clips:
  - path: "scene_01.mp4"
    trim_start: 0        # 可选：裁剪起点
    trim_end: 5          # 可选：裁剪终点
    
  - path: "scene_02.mp4"
    transition: "fade"   # 到下一个片段的转场
    
  - path: "scene_03.mp4"

output: "final_video.mp4"

transition: "fade"           # 全局默认转场
transition_duration: 0.5

bgm: "background_music.mp3"
bgm_volume: 0.2
fade_bgm: true              # BGM 渐入渐出

subtitles: "captions.srt"

resolution: "1080x1920"     # 竖屏
fps: 30
format: "mp4"

转场类型

FFmpeg 支持

转场	效果	命令
fade	淡入淡出	xfade=fade
dissolve	溶解	xfade=dissolve
wipeleft	左擦除	xfade=wipeleft
wiperight	右擦除	xfade=wiperight
slideup	上滑	xfade=slideup
slidedown	下滑	xfade=slidedown
none	无转场	硬切

Remotion 额外支持

• 自定义 React 动画
• 品牌化转场
• 复杂遮罩效果

FFmpeg 命令示例

简单拼接（无转场）

# 创建文件列表
cat > list.txt << EOF
file 'scene_01.mp4'
file 'scene_02.mp4'
file 'scene_03.mp4'
EOF

# 拼接
ffmpeg -f concat -safe 0 -i list.txt -c copy output.mp4

带转场拼接

ffmpeg -i scene_01.mp4 -i scene_02.mp4 \
  -filter_complex "[0:v][1:v]xfade=transition=fade:duration=0.5:offset=4.5[v]" \
  -map "[v]" output.mp4

添加背景音乐

ffmpeg -i video.mp4 -i bgm.mp3 \
  -filter_complex "[1:a]volume=0.3[bgm];[0:a][bgm]amix=inputs=2:duration=first[a]" \
  -map 0:v -map "[a]" output.mp4

添加字幕

ffmpeg -i video.mp4 -vf subtitles=captions.srt output.mp4

详见已有的 ffmpeg-video-editor skill。

工作流程

输入: clips[] + 配置
  ↓
验证所有片段可访问
  ↓
统一格式/分辨率（如需要）
  ↓
计算转场 offset
  ↓
生成 FFmpeg 命令 / Remotion 项目
  ↓
执行渲染
  ↓
输出: 完整视频

与上下游对接

上游输入：

• scene-video-generator 的视频片段
• digital-avatar 的口播视频
• Demo 视频片段

Pipeline 集成示例：

# 从 video-script-generator 输出
scenes:
  - id: 1
    video: "hook.mp4"      # scene-video-generator 输出
    duration: 3s
    
  - id: 2
    video: "talking.mp4"   # digital-avatar 输出
    duration: 5s
    
  - id: 3
    video: "demo.mp4"      # demo skill 输出
    duration: 10s
    
  - id: 4
    video: "cta.mp4"       # scene-video-generator 输出
    duration: 3s

# 自动拼接
output: "final_promo.mp4"
transition: "fade"
bgm: "upbeat_music.mp3"

输出格式建议

平台	分辨率	比例	格式
抖音/快手	1080x1920	9:16	mp4
小红书	1080x1440	3:4	mp4
YouTube Shorts	1080x1920	9:16	mp4
B站/YouTube	1920x1080	16:9	mp4
微信视频号	1080x1920	9:16	mp4

注意事项

1. 输入片段分辨率不一致时，会自动缩放（可能影响质量）
2. 转场会减少总时长（每个转场减少 transition_duration）
3. BGM 时长不够会循环，太长会裁剪
4. 大量片段时建议分批处理
5. 4K 视频渲染较慢，考虑先用 1080p 预览