agentleFS
VUXXEVUXXE/cartoon-3d-short/SKILL.md

cartoon-3d-short

Skill4 starsChanged 9 days agoMITView on GitHub

Make a ~55-70 s vertical cartoon story short in the "2D characters in a colourful 3D world" style (flat bean-head cartoon cards with ink outlines, painted 3D town/world with black ink lines, saturated sky, depth-of-field backgrounds, gliding cameras, voice narration, synthesized SFX + march-style music), rendered with three.js inside Remotion. Proven on "The Bridge That Fell for a Song" (2026-10-31). Use when the user asks for another video "like the bridge video", "like the soldiers/bridge short", "in that cartoon 3D style", a history/science story short with cartoon characters in 3D, or says "use the cartoon short skill".

4 other projects use a version of this, 4 of them edited.

Adopted by

Most recently taken up by

  • Other versions4 projects use a version of this.

    Word for word · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

The file

Show its rules and text

We split this skill into the rules it gives an agent

18 rules. 9 are common, 2 are distinctive, and 7 are very specific to this project.

  • 9

    Common

    Most of the 4 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Common:

    Never copy their scripts, characters or shots.

    line 15
  2. Common:

    Never say the channel's name in a voice-over or description.

    line 16
  3. Very specific:

    Premade 3D Asset Library (Polyfork): Use Polyfork Catalog (references/polyfork/, 3,086 low-poly models across 47 themed kits) whenever …

    line 20
  4. Common:

    No captions or labels in the video.

    line 22
  5. Very specific:

    Write the facts with their source at the top of SCRIPT.md.

    line 34
  6. Common:

    Never describe on screen what the viewer already sees.

    line 44
  7. Very specific:

    Always use Mark via Chatterbox (the default in scripts/generate_voice.py):

    line 46
  8. Common:

    Every visual reaction lands 0-3 frames before its word.

    line 69
  9. Distinctive:

    Search 3,086+ low-poly models across 47 themed kits:

    line 72
  10. Very specific:

    Place in scene via placePolyforkProp(batch, model, x, y, z, height, yaw).

    line 80
  11. Common:

    Every pixel parameter scales with output height / 1920.

    line 92
  12. Very specific:

    For any topic: Do not hardcode sounds from previous shorts.

    line 94
  13. Very specific:

    Run: node scripts/bridge-stills.mjs out/<name>/x_ 100,400 0.5 (automatically runs with GPU rasterization and EGL/ANGLE hardware …

    line 99
  14. Distinctive:

    Always pass GPU hardware acceleration flags:

    line 103
  15. Very specific:

    Start every new short from templates/bridge/ (copy to src/videos/<name>/): the look lives in these values, the story lives in story.ts / …

    line 145
  16. Common:

    Every hold pushes in 0.15-0.35 %/f, with 0.6 %/f on punch words.

    line 180
  17. Common:

    No hand-held float. Impact shakes flip sign every frame and decay over 7 f.

    line 182
  18. Common:

    No blush or eye highlights.

    line 187

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Cartoon 3D story short (three.js + Remotion)

Proven example: The Bridge That Fell for a Song. It is an original 67 s short about the 1831 Broughton bridge and 74 soldiers marching in step.

  • Code: src/videos/bridge/ of your Remotion project (composition BridgeShort, 1080x1920, 30 fps).
  • Script, voice and sound: ~/bridge-short/.
  • Style research: references/.

The style reference is a channel studied privately (references/: STYLE_BIBLE.md, reports/*.md, frames/).

  • Those videos are private study material. Never upload, publish or send them.
  • Never copy their scripts, characters or shots. Story, characters and voice are always OUR OWN.
  • Never say the channel's name in a voice-over or description.

User rules that apply:

  • Always use Mark for voiceover: Synthesize narration using Chatterbox TTS cloning the Mark ("Mark - Natural Conversations") voice (templates/audio/voices/mark.wav), with optional expressive tags ([chuckle], [gasp], [sigh]).
  • Premade 3D Asset Library (Polyfork): Use Polyfork Catalog (references/polyfork/, 3,086 low-poly models across 47 themed kits) whenever assembling 3D scenery, buildings, nature, vehicles, props, or landmarks. Use python scripts/polyfork.py to search, inspect, download GLBs, or generate code snippets.
  • Always render with GPU when available: enforce hardware acceleration across WebGL (EGL/ANGLE), Remotion renders, Whisper alignment, and Chatterbox/Kokoro voice generation.
  • No captions or labels in the video.
  • English only.
  • Voice: one steady conversational tone (or controlled expressive tags).
  • Jokes are researched real facts.
  • Standard mix: music +3 dB, SFX -6 dB, voice unchanged.
  • Final export 2K vertical: 1440x2560, --scale 1.3333.
  • Model verification (Gemini Vision) verifies every look check.
  • Save after every step.
  • "Just do it" once the script is approved.

Pipeline

  1. Topic and facts. Pick one surprising true story with a physical cause and effect (it should be visual), plus a simple everyday analogy that explains it.
    • Write the facts with their source at the top of SCRIPT.md.
  2. Script (~3.7 words/s, 55-70 s). Beat formula from the bridge short:
    • Hook: the absurd fact in one line ("Seventy-four soldiers knocked down an iron bridge just by walking across it.").
    • Setting: year, place, pride.
    • Action: escalation over three or four beats, each a little bigger.
    • Snap: the moment it breaks.
    • Deadpan double line: "Luckily ... / Unluckily ...".
    • Explanation: one physics idea through an everyday analogy (the kid on a swing).
    • Consequence: the rule it created.
    • Callback twist: the ending.
    • Never describe on screen what the viewer already sees.
  3. Voice & Word Timings (voice.wav + words.json). Always use Mark via Chatterbox (the default in scripts/generate_voice.py):
    # Default: Generates Mark's voice using Chatterbox with GPU acceleration:
    python scripts/generate_voice.py --out-dir ~/bridge-short/voice --script SCRIPT.md
    
    # With expressive tags & custom emotion dial:
    python scripts/generate_voice.py --out-dir ~/bridge-short/voice --script SCRIPT.md --exaggeration 0.35
    • Mark Voice Profile (templates/audio/voices/mark.wav): Natural conversational pacing, warm timbre, articulate storytelling delivery.
    • Chatterbox Expression Controls:
      • Tags: [chuckle], [gasp], [sigh], [pause], [groan]. These vocal reactions are synthesized in Mark's voice and timestamped into words.json so camera kicks and visual reactions land on the exact beat.
      • --exaggeration: continuous control from 0.25 (deadpan BBC documentary style) to 0.35 (natural conversational) to 0.65+ (theatrical).
    • Output Format:
      • voice.wav: 48 kHz mono uncompressed WAV.
      • words.json: [{"w": "soldiers", "s": 0.42, "e": 0.88}, ...]. Everything in story.ts and mix.py is timed to words.json.
  4. Shot plan in story.ts: copy templates/bridge/ (the final bridge source, same as src/videos/bridge/ of your Remotion project) to src/videos/<name>/, register a new composition next to BridgeShort in src/Root.tsx, and rename it.
    • SHOTS: id, a/b frames, light day/golden, crowd mode, tau story time.
    • XFADE: crossfades at place/time changes.
    • FLASH_AT: a white flash after the hook.
    • CAMS: per-shot keys {f, dur?, cut?, s: f => {pos, look, fov, roll}}.
    • KICKS: impact flip-shakes.
    • DOF/dofAt: lens blur per beat.
    • FOREGROUND: near framing props.
    • Every visual reaction lands 0-3 frames before its word.
  5. World & Premade 3D Assets (Polyfork).
    • Premade Assets from Polyfork Catalog (references/polyfork/):
      • Search 3,086+ low-poly models across 47 themed kits:
        python scripts/polyfork.py search "house" --kit medieval-village-kit-ca0b43
        python scripts/polyfork.py kits
        python scripts/polyfork.py download a-frame-cabin-6cb550 --out-dir ./public/models
        python scripts/polyfork.py snippet a-frame-cabin-6cb550
      • Load via loadPolyforkKit(["a-frame-cabin-6cb550", ...]) from polyfork.ts.
      • Place in scene via placePolyforkProp(batch, model, x, y, z, height, yaw). Models automatically convert to two-tone cartoon shading (toLambert()) and receive the ink outline pass.
    • lights.ts: LIGHTS plus the LOOK sky/river/cloud palette per light.
    • shade.ts: two-tone shading and untone().
    • build.ts: town, houses and Polyfork / Kenney CC0 kit props, recoloured, with sight-line culling of anything blocking a camera.
    • paint.ts: HD Canvas2D painted textures with seams painted in as dark tints.
    • world.ts: sky shader, fog = horizon colour, river, per-frame posing, sun frozen per shot and camera-relative.
  6. Characters.
    • toon.ts: the Canvas2D Pen drawing API, 600x1000 design canvas, soles at y = 975, bean heads, stick limbs, plum variable-width outline, face library.
    • cast.ts: camera-facing billboards, crowds, effect cards, blinks, walk cycles. All billboards go on the NOINK layer.
    • cardshade.ts: cards lit by the scene light, no emissive.
  7. Post (Bridge.tsx + ink.ts + lensblur.ts + grade.ts).
    • Composer order: ink → lens pre → lens main → bloom → NEUTRAL tone map → grade (MSAA 4).
    • Every pixel parameter scales with output height / 1920.
  8. Topic-Adaptive Sound & SFX Generation (audio/: sfx.py, sfx_gen.py, mix.py, music.py, dsp.py).
    • For any topic: Do not hardcode sounds from previous shorts. Run python scripts/generate_sfx.py --project-dir . --render.
    • The generator scans SCRIPT.md and voice/words.json, identifies narrative keywords/action beats, builds layered ambience beds, and generates timestamped procedural sound cues from the 48+ DSP generators in sfx_gen.py (weather, vehicles, metal stress, cable snaps, impacts, splashes, foley scuffs, cartoon wobbles, and crowds).
    • audio/mix.py masters the final audio to ITU-R BS.1770-4 compliance: -15.0 LUFS, True Peak -1.2 dBTP, with automated speech ducking and section rides.
    • Output: audio/mix.wav (48 kHz stereo, exact length), stems (stem_sfx.wav, stem_music.wav, stem_voice.wav), and cues.md.
  9. Look checks (GPU-accelerated stills).
    • Run: node scripts/bridge-stills.mjs out/<name>/x_ 100,400 0.5 (automatically runs with GPU rasterization and EGL/ANGLE hardware acceleration).
    • Make side-by-sides against reference frames with python scripts/sbs_sheet.py out.jpg --ours ... --ref ....
    • Send the sheet to Gemini Vision: evaluate against target metrics, score 1-10 across 8 dimensions, provide a concise verdict and numbered frame fixes.
  10. Final render (GPU-accelerated).
    • Always pass GPU hardware acceleration flags:
      # On Linux / Arch (EGL backend):
      npx remotion render src/index.ts <Comp> out/<name>/video.mp4 --codec h264 --crf 16 --scale 1.3333 --gl egl \
        --chromium-options='{"args":["--enable-gpu","--use-gl=egl","--ignore-gpu-blocklist","--enable-gpu-rasterization"]}'
      
      # On Windows (ANGLE backend):
      npx remotion render src/index.ts <Comp> out/<name>/video.mp4 --codec h264 --crf 16 --scale 1.3333 --gl angle \
        --chromium-options='{"args":["--enable-gpu","--use-gl=angle","--ignore-gpu-blocklist"]}'
    • Mux with -map 0:v:0 -map 1:a:0 (Remotion writes a silent audio track).
    • QA: python scripts/short_qa.py final.mp4 out/<name>/qa shots.json checks every frame for black, white, frozen and pop frames, plus loudness and sheets.
    • Fix anything it finds, then run a final visual audit.
  11. Deliver. Provide the final mp4 path, summarize shot count and mix specs.

Premade 3D Assets: Polyfork Catalog (47 Kits, 3,086 Models)

The skill includes the full offline database of Polyfork Catalog at references/polyfork/catalog.json. All assets are low-poly (300–2,000 tris), Y-up, meter-scaled, centered at y=0, and download directly from https://polyfork.dev/cdn/<id>.glb.

Key Themed Kits:

  • Towns & Architecture: Medieval Village (medieval-village-kit-ca0b43), Little Tokyo (japanese-suburban-city-street-baa482), Medieval Japan (medieval-japan-cb01ad), Frontier Town (frontier-town-a0085b), NYC (nyc-city-kit-fbca88), Roadside Americana (roadside-americana-50d40c).
  • Nature & Landscape: Nature & Forest (nature-forest-kit-f29d6a), Campground & Lake (campground-lake-ad32b0), Prehistoric Valley (prehistoric-valley-8638dc), Coral Reef (coral-reef-a7128a), Cozy Farm (cozy-farm-cc979f).
  • Transportation & Sci-Fi: Countryside Railway (countryside-railway-9b4c57), Retro Cars (racing-cars-kit-660d53), Space Base (space-base-kit-cf65de), Spaceship Wars (spaceship-wars-kit-234642), Airport (airport-43b137).
  • Props, UI & Interiors: 3D UI kit (3d-ui-kit-5635f7), Dungeon & Crypt (dungeon-crypt-c27d93), Hospital (hospital-kit-2ebfcb), Hotel Casino (casino-3dd029), Pirate Cove (pirate-cove-2ebea4).

CLI Commands:

# 1. Search assets by keyword or filter
python scripts/polyfork.py search "tree" --limit 10
python scripts/polyfork.py search "car" --type vehicle --max-triangles 1500

# 2. Browse themed kits
python scripts/polyfork.py kits --search "train"
python scripts/polyfork.py kit countryside-railway-9b4c57

# 3. Download asset GLB to Remotion public/ folder
python scripts/polyfork.py download steam-locomotive-3b8c21 --out-dir ./public/models

# 4. Generate ready-to-paste Three.js snippet
python scripts/polyfork.py snippet steam-locomotive-3b8c21

Final look values (bridge after 4 verifier rounds; all in templates/bridge/)

Start every new short from templates/bridge/ (copy to src/videos/<name>/): the look lives in these values, the story lives in story.ts / build.ts layout / cast.ts roles.

  • Light presets (lights.ts, sun camera-relative and frozen per shot):
    • day: az 75 / el 48, lit [1.58,1.5,1.33], shade [.30,.29,.36], lift [.05,.06,.11] (cool tint ONLY in shade), sky top #3e7fd2 → horizon #b8d0ec, fog = horizon, 25→300 m, uSat 1.08, ~15 % blue haze past 60 m.
    • golden: az 6 / el 7 (backlit low sun), key toward #FFB36B ×1.5, violet-blue shade (facade ≈ #444380), sky #ffc47c → #8a6f8a, peach fog only past 50 m, sun disc #ffcf9e + bloom, contact blobs + real card shadows.
  • Albedos (paint.ts/build.ts): muted brick / cream / ochre / terracotta houses with plinth, cornice, sills, hoods, door step, chimney + pot, heights 7 / 8.4 / 9.8 m; warm grey/beige stone (saturation −35 % from "storybook"); ochre wood deck C* ~51-58 with near-black seams; iron dark and desaturated. Never lemon, never lavender lit faces.
  • Ink (ink.ts): #030303, 5.1 px near (6 m) → 3.3 px far (90 m) at 1080x1920, creases ×0.8 from 35°, depth jumps from 1 %, fog tint 0.3 only far; character cards on layer NOINK (they carry their own plum ink).
  • Lens (lensblur.ts + dofAt in story.ts): cInf px at 1080: hook/river CU 15-16, close-ups 12-14, MS 10-12, park 7, golden sway 5, wides/aerials 0. Focus on the eyes of the front row.
  • Post (Bridge.tsx): ink → lens pre → lens main → bloom (thr 1.0, int .5, 6 levels, radius .6) → NEUTRAL tone map → grade; MSAA 4. Film grain exists but is OFF (reference measures no grain; it quadruples the file size).
  • Characters (toon.ts/cast.ts/cardshade.ts): bean heads h/w .94, plum ink #241726, eyes .26 W apart, no blush, 14 expressions; adults 3.5 head-heights; walk bob 5 %, lean 1-7°, planted 0.8 m stride; blinks 0.5-2.3 s; tight contact shadow ~.5 under the feet + sun-direction tail by day; golden rim inner only (never an offset outer copy).
  • Camera (story.ts): cuts on big jumps, smootherstep glides 18-40 f, every hold creeps (no static hold ≥4 f), flip kicks decaying over 7 f on impacts, 4-4.6° rolls, foreground props (FOREGROUND) in ~1/3 of shots.

Look targets (measured from the reference; see references/reports/)

  • Ink:
    • World lines: about 3 px at 1080 wide, near-black #0b0a0c, nearly constant with distance. Silhouettes, depth jumps and normal creases only. No line on wall/floor joins or the horizon. Character cards are never outlined by the world ink.
    • Character ink: dark plum #201628-#2f1d28, thicker on the bottom/shadow side.
  • Colour:
    • Day median luma ~.57. Non-ink p5 ~.23 (no crushed blacks). Lit chroma C* p90 ~55. Day sky top ~#5e93d9, horizon 1.28x brighter.
    • Golden: peach horizon #ffd08a, violet-blue shadows (never brown), backlit sun.
  • Depth:
    • DOF on about 1/3 of screen time (CU, MS, inserts), cInf 10-16 px at 1080, focus on the eyes. Establishing and wide shots stay sharp.
    • Fog = horizon colour, 25→420 m day, 20→360 m golden.
    • Bloom only on HDR emissives: threshold 1.0.
  • Camera:
    • Hard cuts on big jumps. Everything else is a smootherstep glide of 18-40 f, peak ≤ 25 px/f at 720p.
    • Every hold pushes in 0.15-0.35 %/f, with 0.6 %/f on punch words.
    • Rolls of 4-5°.
    • No hand-held float. Impact shakes flip sign every frame and decay over 7 f.
  • Characters:
    • About 3.5 head-heights tall, legs 1.4 HH.
    • Stick limbs about 0.055 of head width, ink-dot hands.
    • Bean head (h/w ~0.93, narrower chin, pointier crown).
    • No blush or eye highlights. Faces at brightness ~.93.
    • Walk bob 3-7 %, lean 1-7°, planted feet.
    • Blinks every 0.5-2.3 s, lasting 4 f.

Look-match loop (Gemini workflow)

  1. Measure the reference first. references/reports/*.md contains exact metrics for motion, depth, textures, characters, lighting, outlines, and color grading. Shared brief: references/reports/CONTEXT.md.
  2. Context Window Advantage: With Gemini's 1M–2M token context, you do not need brittle multi-agent sandbox splits unless desired. Gemini can hold and edit story.ts, world.ts, toon.ts, paint.ts, ink.ts, and lights.ts directly in the project directory.
  3. Verify with Gemini Vision:
    • Render preview stills: node scripts/bridge-stills.mjs out/<name>/x_ 100,400 0.5
    • Generate side-by-side sheets: python scripts/sbs_sheet.py out.jpg --ours out/<name>/x_100.jpg out/<name>/x_400.jpg --ref references/frames/short01_0060.jpg references/frames/short01_0180.jpg
    • Gemini Vision inspects out.jpg and scores each area out of 10 (colour day / golden, lighting, ink, textures, characters, depth, composition) against the reference reports, providing exact frame fix recommendations.
  4. Contracts between components live in exported constants (NOINK layer from ink.ts, dofAt/FOREGROUND from story.ts, LIGHTS/LOOK from lights.ts).

GPU Acceleration & Platform Gotchas

  • Always Render with GPU:
    • Linux / Arch: WebGL uses hardware-accelerated --gl=egl via Mesa Intel / NVIDIA drivers (--enable-gpu --use-gl=egl --ignore-gpu-blocklist --enable-gpu-rasterization).
    • Windows: Uses --gl=angle (--use-gl=angle).
    • Voice & Alignment: generate_voice.py automatically checks torch.cuda.is_available() and executes Chatterbox and faster-whisper on GPU with float16.
  • Arch Linux & Linux Disk/Tmpfs: On Linux, /tmp is typically a tmpfs RAM disk. Every Remotion bundle copies public/ (~1.3 GB) and can quickly fill /tmp with ENOSPC. Always keep REMOTION_TMP pointed to a physical drive (e.g. ~/.cache/remotion-tmp or /var/tmp/remotion-tmp).
  • Browser Executable: bridge-stills.mjs automatically searches for Google Chrome / Chromium on Linux/Arch (~/.local/bin/google-chrome, /usr/bin/google-chrome-stable, /usr/bin/chromium), or respects PUPPETEER_EXECUTABLE_PATH.
  • Sight-lines: Cameras end up inside houses: the sight-line culler in build.ts/world.ts removes anything blocking any camera ray.
  • Ink & Post-Processing: Ink in the same merged effect pass as DOF gets overwritten. Ink must be its own pass before the blur.
  • Pixel-sized post parameters must scale with resolution: stills are at 0.5, the final at 1.333.
  • Remotion writes a silent audio track: mux with -map 0:v:0 -map 1:a:0.
  • 2K renders run ~4 fps (≈10 min for 67 s on GPU). Stills at 0.5 take 3-8 s each after bundling.
  • Film grain at 2.5 % quadruples file size and will turn to mush on YouTube; keep it off.

Readability traps the verifier caught

  • A swing drawn as one rope that disappears behind the kid reads as a HANGING. Always two ropes to a wide plank with the kid between them (swingKid in toon.ts), and a frame whose legs clear the arc (A-frame at each end).
  • Props or flotsam closer than ~2 m to a lens become huge cut-off blobs at the frame edge. When a camera moves, re-check every prop near its new position.
  • A foreground bush cut by the frame edge reads as a "green ball"; a foreground prop must be recognisable or gone.
  • Flags: red + yellow star/diamond reads as Vietnam's. Use a period-correct flag (St George's cross for England 1831).
  • Painted signs lit by a side sun go grey-blue and unreadable: give the sign face an unlit material.
  • Golden rim light done as an offset outer copy turns every hat into a picture frame: inner rim only, head/torso only.
  • Balcony/railing rails at chest height through characters: put the lens ~2.5 m up so rails cross legs, not torsos.
  • Flat holds read as freezes in QA (frozen span ≥12 f): every hold needs a creeping push-in.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.