Agent Skills

remotion-motion-graphics

Create and edit professional motion graphics videos with Remotion (React-based video). Use this skill EVERY time the user wants to create a video, edit a video, animate something, build an intro/outro/logo animation, make a Reel/Short/promo/launch video, add text animations or captions to footage, composite images and B-roll into a video, or mentions Remotion, motion graphics, video rendering, or "make this video look better/less generic". Also trigger when editing an existing Remotion project o

Install

npx skills add https://github.com/haidrrrry/claude-remotion-skill --skill remotion-motion-graphics
SKILL.md

Remotion Motion Graphics

Remotion renders React components frame-by-frame into video. Code quality is not the bottleneck — motion design craft is. Untrained generation produces linear easing, opacity-only fades, simultaneous entrances, flat colors, and no texture. That is the "generic AI video" look. This skill exists to prevent it.

Non-negotiable rules (apply to EVERY composition)

  1. NEVER use linear interpolation. Every interpolate() gets an easing curve; every entrance prefers spring(). Always extrapolateLeft/Right: "clamp".
  2. Entrances animate 2–3 properties together (opacity + translateY + scale). A lone fade is forbidden.
  3. Stagger everything. Lists, words, rays, cards: 3–6 frame offsets. Nothing enters simultaneously.
  4. Exits exist and are faster than entrances (~10 frames vs ~20).
  5. Five-layer stack in every scene, bottom to top: background mesh → assets → graphics/type → color grade → grain + vignette. Never a flat solid background.
  6. Every still image gets Ken Burns (slow scale 1→1.08 + pan). Every video asset uses <OffthreadVideo>, never <Video>.
  7. Idle elements breathe: anything on screen >2s gets sin-wave micro-motion.
  8. All timing derives from fps via useVideoConfig(). No magic frame numbers.
  9. One theme object at the top of the project (colors, easings, spring presets, fonts). Never inline a hex color or easing in a component.
  10. Render, extract frames, LOOK at them, fix, re-render. Never deliver an unverified render. This loop is mandatory — see Verification below.

Workflow

Step 1 — Scope

Determine: duration, fps (30 default; 60 only for heavy fast motion), dimensions (1080×1920 Reels/Shorts, 1920×1080 landscape), what assets exist (images, footage, audio, logos), and whether this is a new composition or an edit to an existing project. If editing an existing project: read src/ fully first, find the theme (or create one), and refactor violations of the rules above before adding features.

Step 2 — Setup

New project:

npm install remotion @remotion/cli react react-dom
# optional: @remotion/transitions @remotion/motion-blur @remotion/google-fonts

Copy assets/theme.ts from this skill into src/theme.ts and adjust the palette to the user's brand. Structure: src/index.ts (registerRoot) → src/Root.tsx (Composition, duration/fps/size) → src/scenes/*.tsx → src/components/*.tsx. User assets go in public/, loaded via staticFile().

Step 3 — Build

Read references/motion-patterns.md for the reusable component implementations (BgMesh, Grade, Grain, Vignette, KenBurns, WordReveal, Stagger, Counter, Spark, transitions, parallax). Compose scenes from those patterns. For pacing, rhythm, typography, color palettes, and sound design rules, read references/design-rules.md.

Scene rhythm: HIT → hold (15–20 still frames) → build → HIT. Something must move in the first 15 frames. Never >90 frames without a new visual element.

Step 4 — Render

npx remotion render src/index.ts <CompId> out/video.mp4 --codec h264 --crf 17

Remotion needs a Chromium binary. If its auto-download fails (sandboxes, offline CI), find one and pass it explicitly:

which chromium chromium-browser google-chrome 2>/dev/null
ls /opt/pw-browsers 2>/dev/null   # Playwright installs live here
npx remotion render ... --browser-executable=<path>

If full Chrome errors with "Old Headless mode has been removed", use a headless_shell binary instead (Playwright ships one as chromium_headless_shell-*/chrome-linux/headless_shell).

Step 5 — VERIFY (mandatory, never skip)

Extract frames at key moments and visually inspect each one. The portable method — works on machines with no system ffmpeg, renders the exact frame:

for f in 15 45 90 150; do
  npx remotion still src/index.ts <CompId> out/check_$f.png --frame $f --overwrite
done

If system ffmpeg IS available, extracting from the finished mp4 also verifies the encode itself (-ss seeking is more portable than select= filters, whose quoting breaks in some shells and in Remotion's bundled ffmpeg):

ffmpeg -v error -ss 1.5 -i out/video.mp4 -frames:v 1 check_1.png

Look for, and fix, in order of frequency:

  • Spacing bugs: gap/margin in em resolves against the PARENT font-size (often 16px), not the text size — use px values in flex containers around big type.
  • Text overflowing or touching frame edges (keep critical content in middle 75% vertically for 9:16 — platform UI covers top/bottom).
  • Elements visible before their entrance or after their exit (missing clamp).
  • Color/contrast failures: hero color on >1 element per frame, dim text unreadable over the grade.
  • Layer order mistakes (grain/vignette must be on top, grade above content). Fix → re-render → re-extract → re-inspect. Only deliver after a clean pass. Then run the final checklist at the bottom of references/design-rules.md.

Editing the user's existing footage

To enhance an existing mp4 (captions, grade, intro/outro) rather than build from scratch: put the file in public/, render it as the asset layer with <OffthreadVideo src={staticFile("clip.mp4")} />, set composition duration from the clip length, and stack graphics/grade/grain above it. Get the clip's duration and fps with ffprobe before setting up the composition. For word-synced captions over speech, see the captions section of references/motion-patterns.md.

Reference files

  • references/motion-patterns.md — copy-paste component library: backgrounds, grade/grain/vignette, Ken Burns, text reveals, counters, transitions, parallax, audio sync, captions. Read before writing components.
  • references/design-rules.md — palettes, typography rules, scene architecture, sound design, pre-delivery checklist. Read before designing scenes and before final delivery.
  • assets/theme.ts — the theme template to copy into every project.

Common failure modes to actively avoid

  • Emoji as icons. Emoji render as full-color platform glyphs (green ✳️, blue 🌐) that ignore your palette and silently break the one-hero-color rule, and they sit on whatever background you gave them (orange mascot on orange tile = invisible). Draw glyphs with CSS/SVG in theme colors, or verify every emoji against the extracted frames.
  • No SFX assets is not a reason to ship silent. Synthesize a minimal kit as 16-bit WAVs from a Node script (noise-burst whoosh, pitch-drop pop, sine-thump kick/bass, detuned-sine pad) into public/sfx/ — see examples/scripts/ in the repo. Zero downloads, fully deterministic.
  • Generating one giant component instead of themed, reusable pieces.
  • durationInFrames mismatch between Composition and scene content (dead air).
  • Forgetting --overwrite on re-renders, then inspecting the stale file.
  • Fonts: never rely on system defaults for hero text; load a display font via @remotion/google-fonts or @font-face + staticFile.
  • Trying to "describe" the result to the user instead of rendering and verifying it.

Related skills

video-editgenmedia-labs715KEdit existing video on RunComfy — this skill is a smart router that matches the user's intent to the right edit model in the RunComfy catalog. Picks Wan 2.7 Edit-Video (general restyle / background swap / packaging swap, identity + motion preservation), Kling 2.6 Pro Motion Control (transfer precise motion from a reference video to a target character), or Lucy Edit Restyle (lightweight identity-stable restyle / outfit swap). Bundles each model's documented prompting patterns so the skill gets shai-video-generationgenmedia-labs714KGenerate AI videos on RunComfy via the `runcomfy` CLI — a smart router across the full video-model catalog: HappyHorse 1.0 (Arena #1, native in-pass audio), Wan-AI Wan 2-7 (open weights, audio-driven lip-sync), ByteDance Seedance v2 / 1-5 / 1-0 (multi-modal cinematic), Kling 3.0 / 2-6, Google Veo 3-1, MiniMax Hailuo 2-3, ByteDance Dreamina 3-0. Covers text-to-video (t2v), image-to-video (i2v), and Veo's video-extend endpoint. The skill picks the right model for the user's intent (Arena-#1 qualitai-musicgenmedia-labs714KGenerate AI music on RunComfy via the `runcomfy` CLI — a smart router across the music-model catalog. Routes to ElevenLabs AI Music Generation (premium 44.1 kHz stereo vocal tracks, 5 s–5 min, $0.0083/s) and ACE Step / ACE Step 1.5 (StepFun-AI open-weights, tag-driven composition, multilingual lyrics, $0.0002–0.0003/s, ~27× cheaper), plus ACE Step audio-inpaint (regenerate a time range inside an existing track) and ACE Step audio-outpaint (extend a track before or after). Picks the right model fimage-to-videogenmedia-labs713KAnimate any still image on RunComfy — this skill is a smart router that matches the user's intent to the right i2v model in the RunComfy catalog. Picks HappyHorse 1.0 I2V (Arena #1, native audio, identity preservation) for general animations, Wan 2.7 with `audio_url` for custom-voiceover lip-sync, or Seedance 2.0 Pro for multi-modal animation from image + reference video + reference audio. Bundles each model's documented prompting patterns so the caller gets sharper output without burning iterat

Search skills and MCP servers

Fuzzy search across 23,137 skills and servers