mascot-rig
Skill0 starsChanged 7 days agoMITView on GitHub
SVG rig conventions for PocketForge mascots (part names, pivots, views, face states), pose-data JSON format, render commands for sheet/poses/scenes/animations, and the consistency checklist. Use before building, posing, rendering or animating a mascot.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- templates/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- .cursor/rules/pocketforge.mdcCursor rule
- anti-slop-writingSkill
- ebook-structureSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 73
Read sheet.png: the 3 views share head size, body height, color placement and features.
- Distinctive: line 74
Read every pose PNG next to the sheet: same proportions, no missing part, hands attached, face state valid.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Mascot rig
One rig file drives every still and every animation frame. Poses never redraw the character: they only rotate/translate parts and switch face states. That is what keeps the character identical everywhere.
Files
assets/character/
brief.md design brief + originality check
build-rig.mjs optional: code that writes rig.svg (recommended, keeps geometry shared across views)
rig.svg the rig
pose-data/sheet/ front.json, three-quarter.json, side.json, expr-*.json (6 expressions)
pose-data/*.json the pose library (pointing, thinking, celebrating, warning, explaining-with-board, holding-checklist)
scenes/*.svg chapter scenes (may embed the rig via <use> or a pasted posed copy)
sheet.png poses/*.png scenes/*.png anim/*.mp4 anim/*.gif qa/*.pngrig.svg conventions
viewBox="0 0 400 520", character centered on x=200, feet at y≈470. Root<svg id="rig" data-palette="#hex,#hex,#hex,#hex" data-stroke="6">.- One top-level
<g data-view="front|three-quarter|side">per view. Only one view is visible at a time (runtime setsdisplay). - Parts are
<g data-part="<name>" data-pivot="x y">. Same part names in every view. Required parts:leg-l,leg-r,torso,head,arm-l-upper>arm-l-lower>hand-l,arm-r-upper>arm-r-lower>hand-r,props.-l/-r= viewer's left/right. Lower arm is nested in upper arm, hand in lower arm, so rotations chain. - Pivots: shoulder for upper arm, elbow for lower arm, wrist for hand, hip for legs, neck for head, hip-center for torso.
- The main head shape carries
data-shape="head-base"and the main body shapedata-shape="torso-base".consistency.mjsmeasures these two in every pose. - Limb tube paths carry
data-tube="outer"/data-tube="inner"(see Style). - Face lives inside
head:<g data-face="eyes">,<g data-face="brows">,<g data-face="mouth">. Each holds state children<g data-state="...">, one visible at a time.- eyes:
open,half,closed - brows:
neutral,up,worried,determined - mouth:
neutral,smile,talk-a,talk-e,talk-o,worried
- eyes:
- Props are
<g data-prop="<name>">insideprops(e.g.board,checklist,coin), hidden unless a pose lists them. A prop that a hand holds goes inside that hand's group so it follows the arm. - Style: flat fills from the palette only,
stroke= palette ink, onestroke-width(thedata-strokevalue),stroke-linecap="round" stroke-linejoin="round". No gradients, filters, opacity tricks or text in the rig. - Limbs as outlined tubes: two stacked strokes on the same path (ink at width W+2·stroke, fill color at width W). Line weight stays identical at any rotation.
pose-data JSON
{
"view": "front",
"gaze": "right",
"parts": {
"arm-r-upper": { "rotate": -95 },
"arm-r-lower": { "rotate": -10 },
"head": { "rotate": 6, "translate": [0, -2] },
"torso": { "scaleY": 1.0 }
},
"face": { "eyes": "open", "brows": "up", "mouth": "smile" },
"props": ["board"]
}- Angles in degrees, SVG convention: positive = clockwise on screen. Arms hang down at 0. Right arm pointing right ≈ -90; left arm pointing left ≈ +90.
gaze= the side the character faces/points to. Layout puts the pose on the opposite side of the content so it faces it.- Omitted parts stay at rest. Never scale a part except
torso.scaleYwithin 0.98–1.03 (breathing).
Runtime and render commands
scripts/character/rig-runtime.js (browser) exposes PF.applyPose(svg, pose), PF.lerpPose(a, b, t), PF.ease.*, PF.spring(t, opts), PF.blinkAt(t, schedule).
| Output | Command |
| sheet.png (3 views + 6 expressions) | node scripts/character/render-sheet.mjs projects/<slug> |
| poses/.png + scenes/.png (2x, transparent) | node scripts/character/render-poses.mjs projects/<slug> |
| consistency report | node scripts/character/consistency.mjs projects/<slug> |
| animation MP4 + GIF + QA frames | node scripts/character/render-anim.mjs <page.html> <out-base> [--fps 30 --w 1920 --h 1080] |
Animation pages: plain HTML that defines window.DURATION (seconds) and window.seek(t). The renderer injects window.PF_RIG (rig.svg text) and window.PF_POSES (all pose JSON by name) and loads rig-runtime.js before page scripts run. Every visible property must be a pure function of t: no timers, no Date.now(), no unseeded random.
Consistency checklist (run before any layout or delivery)
consistency.mjspasses: only palette colors, one stroke width, head and torso size within 1% of rest in every pose.- Read
sheet.png: the 3 views share head size, body height, color placement and features. - Read every pose PNG next to the sheet: same proportions, no missing part, hands attached, face state valid.
- Animations: read
qa/<clip>-{start,mid,end}.png; the character matches the sheet in all three; loops end on frame 0's pose. - Originality: no resemblance to a known character, mascot or brand figure (see
brief.md). Any drift: fix the rig or the pose data, never patch a single PNG.
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
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.

