菜单
Audio/video studio production — multi-track editing, color grading, motion graphics, compositing, live streaming, screen recording, YouTube publishing.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Tibsfox/gsd-skill-creator --skill av-studio命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
Use the moment a request is underspecified and you are about to act on an ASSUMED goal. Enumerate the candidate goals the request could mean, project the action or artifact each one implies, and check whether they diverge: if two or more plausible goals lead to materially different actions or artifacts, ask exactly ONE targeted clarifying question; otherwise proceed on the most-likely goal and state the assumption in one line. Clarification is an evidence-producing action, not a delay — but one question, never an interrogation. Distinct from gsd-spec-phase (GSD-phase-bound, emits a heavy SPEC.md) and intent-router (fetch strategy, not goal disambiguation). Backed by agent goal-state inference (arxiv 2606.16813v1). Triggers on acting under an assumed goal when the request admits more than one materially different reading.
Use at write time to vector/embedding memory — Grove content-addressed insertion, chroma/pgvector upserts, memory-consolidation promoting session traces to MEMORY.md, or embedding externally-ingested content. Scores each candidate record against a fixed panel of sentinel queries and quarantines any record that would become the nearest neighbor of too many unrelated queries — a hub — whether from adversarial poisoning or accidental over-generality. This is the memory-record-side sibling of skill-injection-guardian (file-side) and the write-side complement of memory-use-warrant (read-side). Quarantine, never silently drop; a human reviews. Backed by the admission-time hubness gate (arxiv 2606.19692v1). Triggers on inserting into vector memory, consolidating memory, and embedding stranger content.
Run this appropriateness check the moment you are about to integrate a retrieved long-term memory — a Grove content-addressed hit, a chroma/pgvector neighbour, a memory-consolidation digest, or a MEMORY.md line — into a response, especially anything touching private origins, Fox Companies IP, credentials, or Center Camp / consent-governed content. It answers a question intent-router never asks: not WHAT to fetch or HOW, but WHETHER a correctly-retrieved item should reach output. Relevance is not appropriateness — a perfect similarity match can still be a boundary violation. Default is FAIL-CLOSED: if in-context authorization is uncertain, the memory may inform behaviour but must not be surfaced. Backed by RBI-Eval (arxiv 2606.06055v1). Triggers on surfacing recalled sensitive memory into a response.
正在显示 SKILL.md
| name | av-studio |
| description | Audio/video studio production — multi-track editing, color grading, motion graphics, compositing, live streaming, screen recording, YouTube publishing. |
| activationKeywords | ["video editing","color grading","color correction","motion graphics","compositing","green screen","chroma key","screen recording","screencast","OBS","streaming","live stream","YouTube upload","YouTube publish","render","timeline","keyframe","transition","fade","crossfade","title card","lower third","intro","outro","B-roll","cutaway","jump cut","montage","timelapse","slow motion","stabilization","aspect ratio","letterbox","pip","picture in picture"] |
| type | skill |
| category | media |
| status | stable |
| origin | tibsfox |
| modified | false |
| first_seen | "2026-04-11T00:00:00.000Z" |
| first_path | examples/skills/media/av-studio/SKILL.md |
| superseded_by | null |
Combines ffmpeg-media (codecs/formats), audio-engineering (mastering/mixing), and gource-visualizer (repo visualization) into a unified production toolkit for video content creation, live streaming, and multimedia publishing.
# 1. Screen recording (with audio)
ffmpeg -f x11grab -video_size 1920x1080 -framerate 30 -i :0.0 \
-f pulse -i default \
-c:v libx264 -preset ultrafast -crf 18 \
-c:a aac -b:a 192k \
raw_recording.mkv
# 2. Trim to content
ffmpeg -ss 00:00:05 -i raw_recording.mkv -to 00:45:00 -c copy trimmed.mkv
# 3. Add intro + outro
cat > concat.txt << EOF
file 'intro.mp4'
file 'trimmed.mkv'
file 'outro.mp4'
EOF
ffmpeg -f concat -safe 0 -i concat.txt -c:v libx264 -c:a aac assembled.mp4
# 4. Color grade + normalize audio
ffmpeg -i assembled.mp4 \
-vf "eq=brightness=0.04:contrast=1.1:saturation=1.15,unsharp=5:5:0.5" \
-af "loudnorm=I=-14:TP=-1:LRA=11" \
-c:v libx264 -preset medium -crf 20 -pix_fmt yuv420p \
-c:a aac -b:a 192k \
-movflags +faststart \
final.mp4
# 5. Generate thumbnail
ffmpeg -i final.mp4 -ss 00:00:30 -frames:v 1 \
-vf "scale=1280:720" thumbnail.jpg
# Record screen + webcam PiP
ffmpeg -f x11grab -video_size 1920x1080 -framerate 30 -i :0.0 \
-f v4l2 -video_size 320x240 -framerate 30 -i /dev/video0 \
-f pulse -i default \
-filter_complex "[0:v][1:v]overlay=W-w-20:H-h-20[out]" \
-map "[out]" -map 2:a \
-c:v libx264 -preset fast -crf 22 \
-c:a aac -b:a 128k \
presentation.mp4
# Record with mouse cursor highlight
ffmpeg -f x11grab -video_size 1920x1080 -framerate 30 \
-draw_mouse 1 -i :0.0 \
-f pulse -i default \
-c:v libx264 -preset fast -crf 20 \
-c:a aac -b:a 128k \
tutorial.mp4
# Add chapter markers (for YouTube)
ffmpeg -i tutorial.mp4 -c copy \
-metadata:c:0 title="Introduction" \
-metadata:c:0 timebase=1/1000 \
chaptered.mp4
# Apply a 3D LUT (cube format)
ffmpeg -i input.mp4 -vf graded.mp4
ffmpeg -i input.mp4 -vf graded.mp4
# Lift/Gamma/Gain (shadows/midtones/highlights)
ffmpeg -i input.mp4 -vf "\
colorbalance=rs=0.1:gs=-0.05:bs=0.15:\
rm=0:gm=0:bm=0:\
rh=-0.05:gh=0:bh=0.05" \
corrected.mp4
# White balance correction (warm up)
ffmpeg -i input.mp4 -vf "colortemperature=temperature=6500" warm.mp4
# Curves (S-curve for contrast)
ffmpeg -i input.mp4 -vf "curves=master='0/0 0.25/0.15 0.5/0.5 0.75/0.85 1/1'" contrast.mp4
# Fade-in title card (2 seconds)
ffmpeg -i input.mp4 -vf "\
drawtext=text='Episode Title':\
fontsize=72:fontcolor=white@0.0:\
x=(w-text_w)/2:y=(h-text_h)/2:\
alpha='if(lt(t,1),t,if(lt(t,3),1,if(lt(t,4),4-t,0)))'" \
titled.mp4
# Lower third (name + title)
ffmpeg -i input.mp4 -vf "\
drawbox=x=0:y=ih-120:w=500:h=120:color=black@0.7:t=fill,\
drawtext=text='Jane Doe':fontsize=36:fontcolor=white:x=20:y=ih-110,\
drawtext=text='Software Engineer':fontsize=24:fontcolor=0xcccccc:x=20:y=ih-65" \
lower_third.mp4
# Crossfade between two clips (1 second overlap)
ffmpeg -i clip1.mp4 -i clip2.mp4 \
-filter_complex "\
[0:v]trim=0:10,setpts=PTS-STARTPTS[v0];\
[1:v]trim=0:10,setpts=PTS-STARTPTS[v1];\
[v0][v1]xfade=transition=fade:duration=1:offset=9[vout];\
[0:a]atrim=0:10,asetpts=PTS-STARTPTS[a0];\
[1:a]atrim=0:10,asetpts=PTS-STARTPTS[a1];\
[a0][a1]acrossfade=d=1[aout]" \
-map "[vout]" -map "[aout]" crossfade.mp4
# Available xfade transitions:
# fade, wipeleft, wiperight, wipeup, wipedown, slideleft, slideright,
# slideup, slidedown, circlecrop, circleopen, circleclose, dissolve,
# pixelize, diagtl, diagtr, diagbl, diagbr, hlslice, hrslice,
# vuslice, vdslice, hblur, fadegrays, squeezeh, squeezev
# Webcam overlay (bottom right, 25% size)
ffmpeg -i screen.mp4 -i webcam.mp4 \
-filter_complex "\
[1:v]scale=iw/4:ih/4[pip];\
[0:v][pip]overlay=W-w-20:H-h-20" \
-c:v libx264 -crf 22 -c:a aac pip.mp4
# Animated PiP (slides in from right)
ffmpeg -i screen.mp4 -i webcam.mp4 \
-filter_complex "\
[1:v]scale=320:240[pip];\
[0:v][pip]overlay=x='if(lt(t,1),W,W-w-20)':y=H-h-20" \
animated_pip.mp4
# Two-pass stabilization with vidstab
# Pass 1: analyze
ffmpeg -i shaky.mp4 -vf vidstabdetect=shakiness=5:accuracy=15 -f null -
# Pass 2: apply
ffmpeg -i shaky.mp4 -vf vidstabtransform=smoothing=10:input=transforms.trf stabilized.mp4
# Timelapse from photos (one frame per image)
ffmpeg -framerate 30 -pattern_type glob -i 'photos/*.jpg' \
-c:v libx264 -crf 18 -pix_fmt yuv420p timelapse.mp4
# Speed ramp (normal → 4x → normal)
ffmpeg -i input.mp4 -filter_complex "\
[0:v]trim=0:10,setpts=PTS-STARTPTS[v1];\
[0:v]trim=10:30,setpts=0.25*(PTS-STARTPTS)[v2];\
[0:v]trim=30:40,setpts=PTS-STARTPTS[v3];\
[v1][v2][v3]concat=n=3:v=1:a=0[out]" \
-map "[out]" speed_ramp.mp4
# Smooth slow motion (frame interpolation with minterpolate)
ffmpeg -i input.mp4 -vf "minterpolate=fps=120:mi_mode=mci:mc_mode=aobmc:vsbmc=1" \
-r 30 slowmo.mp4
# Stream to YouTube via RTMP
ffmpeg -f x11grab -video_size 1920x1080 -framerate 30 -i :0.0 \
-f pulse -i default \
-c:v libx264 -preset veryfast -maxrate 4500k -bufsize 9000k \
-pix_fmt yuv420p -g 60 \
-c:a aac -b:a 128k -ar 44100 \
-f flv "rtmp://a.rtmp.youtube.com/live2/YOUR_STREAM_KEY"
# Stream to Twitch
ffmpeg -f x11grab -video_size 1920x1080 -framerate 30 -i :0.0 \
-f pulse -i default \
-c:v libx264 -preset veryfast -maxrate 6000k -bufsize 12000k \
-pix_fmt yuv420p -g 60 \
-c:a aac -b:a 160k -ar 44100 \
-f flv "rtmp://live.twitch.tv/app/YOUR_STREAM_KEY"
| Format | Ratio | Resolution | Use |
|---|---|---|---|
| Widescreen | 16:9 | 1920x1080 | YouTube, presentations |
| Vertical | 9:16 | 1080x1920 | TikTok, Reels, Shorts |
| Square | 1:1 | 1080x1080 | Instagram feed |
| Ultrawide | 21:9 | 2560x1080 | Cinematic |
| Classic | 4:3 | 1440x1080 | Legacy, some webcams |
| Cinema | 2.39:1 | 2048x858 | Film standard |
# Add letterbox (16:9 → 21:9 cinematic bars)
ffmpeg -i input.mp4 -vf "pad=iw:iw*9/21:(ow-iw)/2:(oh-ih)/2:black" cinematic.mp4
# Convert vertical to horizontal (blur background)
ffmpeg -i vertical.mp4 -filter_complex "\
[0:v]scale=1920:1080:force_original_aspect_ratio=increase,crop=1920:1080,boxblur=20[bg];\
[0:v]scale=-1:1080[fg];\
[bg][fg]overlay=(W-w)/2:0" \
horizontal_blurbg.mp4
-movflags +faststart for progressive download-pix_fmt yuv420p for compatibilitydrawtext filter| Asset | Role |
|---|---|
ffmpeg-media skill | Codec/format foundation |
audio-engineering skill | Audio mastering/mixing |
ffmpeg-processor agent | Format conversion |
audio-engineer agent | Audio mastering |
podcast-producer agent | Podcast pipeline |
music-analyzer agent | Audio analysis |
gource-visualizer skill | Repository visualization → video |