agentleFS
reescalderreescalder/claude-skills/infinite-canvas/SKILL.md

infinite-canvas

Skill0 starsChanged 4 months agoMITView on GitHub

Build nested, navigable canvas UIs with React Flow. Users can zoom into nodes to reveal child canvases - infinite depth.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Debounce position saves: - don't spam API on every drag

    line 253
  2. Distinctive:

    Lazy load: - only fetch nodes for current level

    line 255

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.

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

Infinite Canvas Skill

Build nested, navigable canvas UIs with React Flow. Users can zoom into nodes to reveal child canvases - infinite depth.

When to Use

  • Mind maps, knowledge graphs, visual databases
  • Nested folder/category systems
  • Any UI where items can contain other items
  • Visual organization tools

Architecture

React Flow (canvas) + Zustand (state) + Redis/DB (persistence)

Core Concepts

1. Data Model

interface CanvasNode {
  id: string;
  parentId: string | null;  // null = root level
  type: 'category' | 'container' | 'item';
  title: string;
  position: { x: number; y: number };
  childCount: number;  // cached for badges
  // Type-specific fields
  color?: string;
  icon?: string;
  content?: string;
  url?: string;
}

interface NavigationState {
  path: PathItem[];           // breadcrumb trail
  currentParentId: string | null;
  viewport: Viewport;         // saved per level
}

2. State Management (Zustand)

import { create } from 'zustand';

interface CanvasStore {
  nodes: CanvasNode[];
  path: PathItem[];
  currentParentId: string | null;
  viewport: Viewport;
  loading: boolean;
  
  // Navigation
  navigateInto: (nodeId: string) => void;
  navigateBack: () => void;
  loadLevel: (parentId: string | null) => Promise<void>;
  
  // CRUD
  createNode: (data: Partial<CanvasNode>) => Promise<void>;
  updateNode: (id: string, data: Partial<CanvasNode>) => Promise<void>;
  deleteNode: (id: string) => Promise<void>;
  updatePositions: (updates: PositionUpdate[]) => Promise<void>;
}

export const useCanvasStore = create<CanvasStore>((set, get) => ({
  nodes: [],
  path: [{ id: 'root', title: 'Home' }],
  currentParentId: null,
  viewport: { x: 0, y: 0, zoom: 0.75 },
  loading: false,

  navigateInto: async (nodeId) => {
    const node = get().nodes.find(n => n.id === nodeId);
    if (!node) return;
    
    set(state => ({
      path: [...state.path, { id: node.id, title: node.title }],
      currentParentId: node.id,
    }));
    
    await get().loadLevel(nodeId);
  },

  navigateBack: async () => {
    const { path } = get();
    if (path.length <= 1) return;
    
    const newPath = path.slice(0, -1);
    const newParentId = newPath.length > 1 ? newPath[newPath.length - 1].id : null;
    
    set({ path: newPath, currentParentId: newParentId });
    await get().loadLevel(newParentId);
  },

  loadLevel: async (parentId) => {
    set({ loading: true });
    const res = await fetch(`/api/nodes?parentId=${parentId || 'root'}`);
    const { nodes, viewport } = await res.json();
    set({ nodes, viewport, loading: false });
  },
}));

3. Canvas Component

import { ReactFlow, Background, Controls, MiniMap } from '@xyflow/react';
import { useCanvasStore } from '../lib/store';

export default function Canvas() {
  const { nodes, currentParentId, updatePositions } = useCanvasStore();
  
  const flowNodes = nodes.map(node => ({
    id: node.id,
    type: node.type,
    position: node.position,
    data: { node },
  }));

  const handleNodeDragStop = (event, node) => {
    // Debounce save positions to API
    updatePositions([{ id: node.id, position: node.position }]);
  };

  return (
    <ReactFlow
      nodes={flowNodes}
      edges={[]}
      nodeTypes={nodeTypes}
      onNodeDragStop={handleNodeDragStop}
      minZoom={0.1}
      maxZoom={2}
    >
      <Background />
      <Controls />
      <MiniMap />
    </ReactFlow>
  );
}

4. Custom Node Components

function CategoryNode({ data, selected }) {
  const { navigateInto } = useCanvasStore();
  const node = data.node;

  const handleDoubleClick = () => {
    if (node.childCount > 0) {
      navigateInto(node.id);
    }
  };

  return (
    <div
      onDoubleClick={handleDoubleClick}
      className={`p-4 rounded-xl border-2 cursor-pointer ${
        selected ? 'ring-2 ring-white/30' : ''
      }`}
      style={{ borderColor: node.color }}
    >
      <span className="text-2xl">{node.icon}</span>
      <h3 className="font-bold">{node.title}</h3>
      {node.childCount > 0 && (
        <span className="text-xs opacity-60">{node.childCount} items</span>
      )}
    </div>
  );
}

5. Breadcrumb Navigation

function Breadcrumb() {
  const { path, loadLevel } = useCanvasStore();

  return (
    <nav className="flex items-center gap-2">
      {path.map((item, index) => (
        <Fragment key={item.id}>
          {index > 0 && <ChevronRight className="w-4 h-4" />}
          <button
            onClick={() => {
              const newPath = path.slice(0, index + 1);
              const parentId = index === 0 ? null : item.id;
              loadLevel(parentId);
            }}
            className={index === path.length - 1 ? 'font-bold' : 'opacity-60'}
          >
            {item.title}
          </button>
        </Fragment>
      ))}
    </nav>
  );
}

6. API Routes (Next.js)

// /api/nodes/route.ts
export async function GET(request: Request) {
  const { searchParams } = new URL(request.url);
  const parentId = searchParams.get('parentId');
  
  const nodes = await redis.getChildren(parentId === 'root' ? null : parentId);
  const viewport = await redis.getViewport(parentId);
  
  return Response.json({ nodes, viewport });
}

export async function POST(request: Request) {
  const body = await request.json();
  const node = await redis.createNode(body);
  return Response.json({ node });
}

Keyboard Shortcuts

useEffect(() => {
  const handleKeyDown = (e: KeyboardEvent) => {
    if (e.key === 'Escape' && path.length > 1) {
      navigateBack();
    }
  };
  window.addEventListener('keydown', handleKeyDown);
  return () => window.removeEventListener('keydown', handleKeyDown);
}, [path.length, navigateBack]);

Dependencies

{
  "@xyflow/react": "^12.0.0",
  "zustand": "^4.5.0",
  "framer-motion": "^11.0.0"
}

Best Practices

  1. Save viewport per level - restore when navigating back
  2. Cache childCount - show badges without loading children
  3. Debounce position saves - don't spam API on every drag
  4. Use Framer Motion - smooth transitions between levels
  5. Lazy load - only fetch nodes for current level

Files

  • examples/store.ts - Full Zustand store
  • examples/Canvas.tsx - Canvas component
  • examples/nodes/ - Custom node components
  • templates/ - Starter template

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.