Video Post Production

End-to-end short-video post-production from one raw talking-head video: transcribe speech, build timed subtitle phrases, highlight key words, place sound eff...

This skill appears to do what it claims: it runs local Python scripts plus ffmpeg to transcribe, generate ASS subtitles, and render a final video. Before installing/running: 1) ensure ffmpeg, ffprobe, and python3 are present; expect faster-whisper/Whisper to download model weights (large, network + disk usage) the first time you transcribe; 2) the reference docs include curl examples to download BGM/SFX from third‑party sites — only run those if you trust the source; 3) no credentials or secrets are requested by the skill, but you should still run it on a machine where large downloads and CPU/GPU usage are acceptable; 4) the subtitle generator uses simple text replacement for highlights (may over-replace in edge cases), so verify outputs on a sample video first. If you are processing sensitive videos, run the skill in an isolated environment to limit exposure.

Type: OpenClaw Skill Name: video-post-production Version: 1.0.0 The video-post-production skill bundle provides a legitimate and well-structured workflow for automating video editing tasks such as transcription, subtitle generation, and audio mixing. It utilizes standard tools like ffmpeg, ffprobe, and the faster-whisper library. The Python scripts (align_speech.py, generate_subtitles.py, render_video.py) use safe subprocess calls with list-based arguments, minimizing shell injection risks. The documentation (SKILL.md and references/) provides clear instructions and helpful resources for acquiring assets from reputable sites like Mixkit and Pixabay, with no evidence of malicious intent, data exfiltration, or harmful prompt injection.