Ffmpeg

Claude skills, scripts and configs for everyday use

Install

npx -y skills add sakydev/claude --skill ffmpeg

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use this skill whenever the user asks to process, convert, compress, trim, merge, extract, or otherwise manipulate video or audio files using ffmpeg. Covers format conversion, compression, clipping, filtering, and any other ffmpeg operation.

SKILL.md

2.9 KB, 747 tokens by cl100k_base, as published. Nobody here has run it

FFmpeg Skill

Hard rules

Never overwrite the original. Write to a new file. Derive the name if none given: video.mp4 becomes video_compressed.mp4. Never use -y on the same path as the input.
Never destroy quality. Minimum floors:

Codec Flag Floor Good default
H.264 (libx264) -crf 28 max 18-23
H.265 (libx265) -crf 32 max 24-28
VP9 (libvpx-vp9) -crf 40 max 31-36
AAC -b:a 128k min 192k
MP3 -q:a 4 max 2

If the target size requires going below these floors, say so. Offer the best size achievable at the floor instead.
File over 1 GB? Stop. Say: "This file is [size]. Processing will take a while and produce large intermediates. Continue?" Only proceed with confirmation.

Codec	Flag	Floor	Good default
H.264 (libx264)	`-crf`	28 max	18-23
H.265 (libx265)	`-crf`	32 max	24-28
VP9 (libvpx-vp9)	`-crf`	40 max	31-36
AAC	`-b:a`	128k min	192k
MP3	`-q:a`	4 max	2

Step 1 — Inspect first

Always run this before building any command:

ffprobe -v quiet -print_format json -show_format -show_streams "<input>"

Read the size, codec, resolution, bitrate, duration. Check the 1 GB rule here.

Step 2 — Pick the right approach

Compression: libx264 -crf 23 -preset slow. Copy audio unless asked otherwise: -c:a copy.

Format conversion: If only the container changes, use -c copy. No re-encode. No quality loss. Instant.

Trimming: -ss before -i for fast seek. -c copy when possible.

Audio extraction: -vn -c:a copy to extract without re-encoding.

Scaling: -vf scale=1280:-2 to preserve aspect ratio. Always pair with a re-encode.

Step 3 — Show the command before running

ffmpeg -i "input.mp4" -c:v libx264 -crf 23 -preset slow -c:a copy "input_compressed.mp4"
# -crf 23     : quality (lower = better)
# -preset slow : smaller file, same quality
# -c:a copy   : audio untouched

Show it. Explain non-obvious flags. Ask before running on large files or long operations.

Step 4 — Report after

Original size vs output size.
Any quality warnings from stderr.
Output file path.

Output larger than input? Flag it. Suggest -c copy if applicable.

Concerns

Situation	Say
Target size needs CRF above floor	"Hitting [size] needs CRF [N]. Below acceptable quality. Best I can do at the floor is [size]."
User says "compress as much as possible"	"How small do you need it? Below CRF [N] it degrades noticeably."
Input already matches target format	"Already [format]. I can remux without re-encoding. Lossless and instant."
User asks to overwrite original	"Not overwriting the original. Saving as [derived name] instead."
File over 1 GB	Stop. Confirm before proceeding.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most video audio skills give in 747 tokens

Counted across 622 of the 795 authors here whose files we hold, read 2026-08-07

read individual rule files for detailed explanationsin 21 of 622, across 10 files
render final videoin 13 of 622, across 6 files
Use WAV PCM 16kHz mono audio formatin 12 of 622, across 3 files
Use this skill when dealing with Remotion codein 11 of 622, across 4 files
save generated audio to a WAV filein 11 of 622, across 4 files
handle conversion errors gracefullyin 10 of 622, across 6 files
add captions to videos alwaysin 10 of 622, across 4 files
generate music from text descriptions using MusicGenin 9 of 622, across 2 files
do not skip pipeline layersin 9 of 622, across 3 files
do not make one tool do everythingin 9 of 622, across 3 files
use azure document intelligence for complex pdfsin 9 of 622, across 4 files
never ask the user to paste their full API keyin 9 of 622, across 3 files

Said here and by no other author read

derive output filename if none is given
inspect input with ffprobe before building command
do not drop quality below specified codec floors
stop and confirm when processing files over 1 gb
use stream copy for audio extraction
show full command before running it

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.