← Back to list
video-assemble Checked
A skill from the GitHub repository zenstory-ai/video-recap-skills.
9.0Overall rating
Installation
🤖 Install via AI
Copy the prompt below and send it to your AI assistant (e.g. Claude Code) — it will follow the instructions and install automatically.
Install the "video-assemble" skill by following the instructions at https://skill123.me/install/video-assemble.
⌨️ Command line install
Run in your terminal — downloads and installs to ~/.claude/skills/.
curl -fsSL https://skill123.me/install/video-assemble.sh | bash
📦 Download ZIP
Download the zip and extract it into your skills directory (e.g. ~/.claude/skills/), then restart your session.
⬇ Download v1.0.0 · 96 KBAbout this skill
Overview
A skill from the GitHub repository zenstory-ai/video-recap-skills.
Source
- Repo: https://github.com/zenstory-ai/video-recap-skills
- Path: skills/video-assemble
Score breakdown
Trigger
7.5
Clear What/When as a pipeline final stage with bilingual trigger words (视频合成/混音/压字幕/mux/ducking); user-invocable=false makes it router-facing; no explicit When-Not in the description.
Structure
9.3
98-line SKILL.md over a well-factored 27-script (~5.8k lines) toolkit with clean contracts; references are JSON protocol templates rather than load-on-demand docs, so progressive disclosure is thinner.
Workflow
9.2
Strong execution design: explicit input/output contracts, named failure gates (no_safe_fit, timeline_audio_mismatch), ffmpeg subtitle-filter preflight, QC manifest (assembly_qc.json) as release gate, atomic draft writing, and a clear capability-boundary section.
Content
9.0
Precise, consistent, directly runnable CLI with full flag documentation; a few deep implementation rules live in SKILL.md prose that mirror script internals.
Engineering
9.0
Valid frontmatter (plus pipeline-oriented user-invocable flag); no shebangs but documented python3 invocation; ffmpeg/libass dependency preflighted; all referenced paths exist.
Security
10.0
Capability manifest matches declared scope: zero network calls, no credential access, destructive ops limited to its own temp dirs, vendored third-party templates commit-pinned with license and example credentials explicitly excluded.
Final-stage video assembly skill from zenstory-ai/video-recap-skills: lays TTS narration over source video with ducking, SRT/ASS subtitles and loudness normalization via ffmpeg, plus optional JianYing draft export. Capability manifest (6.0): reads only the explicitly passed video, args and work_dir artifacts; side effects confined to work_dir/output-dir and optional JianYing drafts dir (atomic tmp-dir writes, non-destructive sibling-dir numbering); network destinations: none — all subprocess calls are local ffmpeg/ffprobe with list-args, no shell; credentials: none (vendored JianYing protocol templates explicitly stripped of upstream example credentials, MIT-licensed and commit-pinned); persistence: none. Capabilities match declared scope.