Video Edit Skill
Edit long videos into concise, story-driven clips by analyzing the transcript, selecting key moments, and assembling them with ffmpeg.
Workflow
Phase 1: Discover & Transcribe
- Find the video — locate file, get duration/format with
ffprobe
- Extract audio — convert to 16kHz mono WAV for transcription:
ffmpeg -y -i INPUT -vn -acodec pcm_s16le -ar 16000 -ac 1 /tmp/audio.wav
- Transcribe with Whisper — use
whisper Python library (install via pip3 install openai-whisper if needed):import whisper, json
model = whisper.load_model("base")
result = model.transcribe("/tmp/audio.wav", language="en", word_timestamps=True)
# Save full result for reference
with open("/tmp/transcript.json", "w") as f:
json.dump(result, f)
# Print timestamped segments
for seg in result["segments"]:
print(f'[{seg["start"]:.1f}-{seg["end"]:.1f}] {seg["text"].strip()}')
Phase 2: Analyze & Select Moments
Read the full transcript and identify the story arc. Look for:
- Setup: Where the task/prompt is clearly stated
- Progress beats: AI working, intermediate results appearing
- Positive reactions: "wow", "that's great", "it worked", "cool", "upgraded", excitement
- Key insights: Moments where approach changes or something clever happens
- Resolution: Final confirmation that it worked, "thank you", wrap-up
Selection rules:
- Skip duplicate/repeated prompts (e.g., if there was a reset and the same prompt was given twice, use only the second clean attempt)
- Skip troubleshooting tangents, debugging, off-topic conversations, bathroom breaks, silence
- Skip filler ("uh", "hmm", long pauses) — but these are fine within a selected segment
- Keep segments long enough for context (15-75s each) — don't micro-cut
- Aim for 10-20% of original duration as a starting point
- Preserve the chronological order — never rearrange
Present the plan to the user before cutting — list each segment with timestamp range and what it captures.
Phase 3: Extract & Assemble
Extract each segment as a separate file with re-encoding for clean cuts:
ffmpeg -y -ss START -t DURATION -i INPUT \
-c:v libx264 -preset fast -crf 23 \
-c:a aac -b:a 128k \
/tmp/segments/segNN.mp4
Create concat file and merge:
for f in /tmp/segments/seg*.mp4; do
echo "file '$f'" >> /tmp/segments/filelist.txt
done
ffmpeg -y -f concat -safe 0 -i /tmp/segments/filelist.txt \
-c:v libx264 -preset fast -crf 23 \
-c:a aac -b:a 128k \
OUTPUT.mp4
Verify — check duration, file size, playback.
Phase 4: Save
Save to the same directory as the source with _edited suffix, or as specified by user.
Common Mistakes & Fixes
| Mistake |
Fix |
| Including the first attempt when there was a reset/retry |
Only include the clean second attempt — watch for repeated prompts |
| Cutting segments too short (<10s) |
Keep 15-75s per segment for natural flow |
| Including long silences or "uh/hmm" segments |
Skip segments that are mostly filler, but filler within a good segment is fine |
Using -c copy for segment extraction |
Always re-encode (-c:v libx264) — copy mode causes keyframe alignment issues and glitchy cuts |
Forgetting -safe 0 in concat |
Required when using absolute paths in the file list |
| Rearranging chronological order |
Never do this — the story must flow naturally in time |
| Over-compressing (too few segments) |
The edit should still tell the complete story — don't skip important transitions |
| Including debugging/troubleshooting tangents |
Skip unless the debugging itself is the story |
What Makes a Good Edit
- Tells a story: setup → progress → "wow it worked" → done
- Highlights positive outcomes: reactions of surprise, satisfaction, confirmation
- Shows the AI being smart: clever approaches, autonomous decisions, learning
- No dead time: every second earns its place
- Natural transitions: segments should feel like they flow, even with jumps
- User can follow along: enough context in each segment to understand what's happening
Dependencies
ffmpeg / ffprobe (install via brew install ffmpeg)whisper Python library (install via pip3 install openai-whisper)- Sufficient disk space for extracted audio + segments (rough guide: 2x source file size)
1---2name: video-edit3description: Edit long videos into short, story-driven clips using transcript analysis and ffmpeg. Use when user asks to "edit video", "cut video", "shorten video", "make highlight reel", "trim video", "create clip", or wants to extract the best moments from a recording. Works with screen recordings, demos, meetings, and any video with speech.4---56# Video Edit Skill78Edit long videos into concise, story-driven clips by analyzing the transcript, selecting key moments, and assembling them with ffmpeg.910## Workflow1112### Phase 1: Discover & Transcribe13141. **Find the video** — locate file, get duration/format with `ffprobe`152. **Extract audio** — convert to 16kHz mono WAV for transcription:16 ```bash17 ffmpeg -y -i INPUT -vn -acodec pcm_s16le -ar 16000 -ac 1 /tmp/audio.wav18 ```193. **Transcribe with Whisper** — use `whisper` Python library (install via `pip3 install openai-whisper` if needed):20 ```python21 import whisper, json22 model = whisper.load_model("base")23 result = model.transcribe("/tmp/audio.wav", language="en", word_timestamps=True)24 # Save full result for reference25 with open("/tmp/transcript.json", "w") as f:26 json.dump(result, f)27 # Print timestamped segments28 for seg in result["segments"]:29 print(f'[{seg["start"]:.1f}-{seg["end"]:.1f}] {seg["text"].strip()}')30 ```3132### Phase 2: Analyze & Select Moments3334Read the full transcript and identify the **story arc**. Look for:3536- **Setup**: Where the task/prompt is clearly stated37- **Progress beats**: AI working, intermediate results appearing38- **Positive reactions**: "wow", "that's great", "it worked", "cool", "upgraded", excitement39- **Key insights**: Moments where approach changes or something clever happens40- **Resolution**: Final confirmation that it worked, "thank you", wrap-up4142**Selection rules:**43- Skip duplicate/repeated prompts (e.g., if there was a reset and the same prompt was given twice, use only the second clean attempt)44- Skip troubleshooting tangents, debugging, off-topic conversations, bathroom breaks, silence45- Skip filler ("uh", "hmm", long pauses) — but these are fine *within* a selected segment46- Keep segments long enough for context (15-75s each) — don't micro-cut47- Aim for 10-20% of original duration as a starting point48- Preserve the chronological order — never rearrange4950**Present the plan to the user** before cutting — list each segment with timestamp range and what it captures.5152### Phase 3: Extract & Assemble53541. **Extract each segment** as a separate file with re-encoding for clean cuts:55 ```bash56 ffmpeg -y -ss START -t DURATION -i INPUT \57 -c:v libx264 -preset fast -crf 23 \58 -c:a aac -b:a 128k \59 /tmp/segments/segNN.mp460 ```61622. **Create concat file** and merge:63 ```bash64 for f in /tmp/segments/seg*.mp4; do65 echo "file '$f'" >> /tmp/segments/filelist.txt66 done67 ffmpeg -y -f concat -safe 0 -i /tmp/segments/filelist.txt \68 -c:v libx264 -preset fast -crf 23 \69 -c:a aac -b:a 128k \70 OUTPUT.mp471 ```72733. **Verify** — check duration, file size, playback.7475### Phase 4: Save7677Save to the same directory as the source with `_edited` suffix, or as specified by user.7879## Common Mistakes & Fixes8081| Mistake | Fix |82|---------|-----|83| Including the first attempt when there was a reset/retry | Only include the clean second attempt — watch for repeated prompts |84| Cutting segments too short (<10s) | Keep 15-75s per segment for natural flow |85| Including long silences or "uh/hmm" segments | Skip segments that are mostly filler, but filler *within* a good segment is fine |86| Using `-c copy` for segment extraction | Always re-encode (`-c:v libx264`) — copy mode causes keyframe alignment issues and glitchy cuts |87| Forgetting `-safe 0` in concat | Required when using absolute paths in the file list |88| Rearranging chronological order | Never do this — the story must flow naturally in time |89| Over-compressing (too few segments) | The edit should still tell the complete story — don't skip important transitions |90| Including debugging/troubleshooting tangents | Skip unless the debugging itself is the story |9192## What Makes a Good Edit9394- **Tells a story**: setup → progress → "wow it worked" → done95- **Highlights positive outcomes**: reactions of surprise, satisfaction, confirmation96- **Shows the AI being smart**: clever approaches, autonomous decisions, learning97- **No dead time**: every second earns its place98- **Natural transitions**: segments should feel like they flow, even with jumps99- **User can follow along**: enough context in each segment to understand what's happening100101## Dependencies102103- `ffmpeg` / `ffprobe` (install via `brew install ffmpeg`)104- `whisper` Python library (install via `pip3 install openai-whisper`)105- Sufficient disk space for extracted audio + segments (rough guide: 2x source file size)
m13v