Agent skill

reactflow-expert

Builds DAG visualizations using ReactFlow v12 with custom nodes, ELKjs auto-layout, Zustand state management, and live state updates via WebSocket. Use when implementing workflow visualization dashboards, creating custom agent node components, integrating ELK layout algorithms, or wiring execution state into React components. Activate on "ReactFlow", "workflow visualization", "DAG visualization", "ELKjs", "custom nodes", "node-based editor", "graph visualization". NOT for writing Mermaid diagrams (use mermaid-graph-writer), general React development, or static diagram rendering.

Stars 81
Forks 12

Install this agent skill to your Project

npx add-skill https://github.com/curiositech/some_claude_skills/tree/main/.claude/skills/reactflow-expert

Metadata

Additional technical details for this skill

tags
reactflow workflow-visualization dag-visualization
pairs with
[
    {
        "skill": "websocket-streaming",
        "reason": "Live DAG state updates flow through WebSocket connections to ReactFlow visualizations"
    },
    {
        "skill": "human-gate-designer",
        "reason": "Human approval gates are rendered as interactive ReactFlow custom nodes"
    },
    {
        "skill": "task-decomposer",
        "reason": "Decomposed task DAGs are visualized as ReactFlow node graphs with ELKjs layout"
    }
]

SKILL.md

ReactFlow Expert

Builds DAG visualizations using ReactFlow v12 with custom agent nodes, ELKjs auto-layout, Zustand state management, and live execution state updates.


When to Use

Use for:

  • Building workflow/DAG visualization dashboards
  • Creating custom ReactFlow node components for agent state
  • Integrating ELKjs auto-layout for automatic graph positioning
  • Wiring WebSocket execution events into ReactFlow state
  • Implementing zoom, pan, selection, and node interaction

NOT for:

  • Static Mermaid diagrams (use mermaid-graph-writer)
  • General React component development
  • Non-graph visualizations (charts, tables)

Architecture

mermaid
flowchart TD
  subgraph "State Layer"
    Z[Zustand Store] --> N[nodes + edges]
    Z --> U[updateNodeData]
    Z --> A[applyNodeChanges / applyEdgeChanges]
  end
  
  subgraph "Layout Layer"
    E[ELKjs] --> P[Compute positions]
    P --> Z
  end
  
  subgraph "Data Layer"
    WS[WebSocket] --> Z
    API[REST API] --> Z
  end
  
  subgraph "Render Layer"
    Z --> RF[ReactFlow component]
    RF --> CN[Custom AgentNode]
    RF --> CE[Custom edges]
    RF --> PA[Panel controls]
  end

Core Patterns (ReactFlow v12)

Zustand Store (Recommended over useNodesState for complex editors)

typescript
import { create } from 'zustand';
import { applyNodeChanges, applyEdgeChanges, type Node, type Edge } from '@xyflow/react';

interface DAGStore {
  nodes: Node[];
  edges: Edge[];
  onNodesChange: (changes: any) => void;
  onEdgesChange: (changes: any) => void;
  setNodes: (nodes: Node[]) => void;
  setEdges: (edges: Edge[]) => void;
  updateNodeData: (nodeId: string, data: Record<string, any>) => void;
}

const useDAGStore = create<DAGStore>((set, get) => ({
  nodes: [],
  edges: [],
  onNodesChange: (changes) => set({ nodes: applyNodeChanges(changes, get().nodes) }),
  onEdgesChange: (changes) => set({ edges: applyEdgeChanges(changes, get().edges) }),
  setNodes: (nodes) => set({ nodes }),
  setEdges: (edges) => set({ edges }),
  // CRITICAL: create NEW object to trigger ReactFlow re-render
  updateNodeData: (nodeId, data) => set({
    nodes: get().nodes.map((n) =>
      n.id === nodeId ? { ...n, data: { ...n.data, ...data } } : n
    ),
  }),
}));

Custom Agent Node

typescript
import { Handle, Position, type NodeProps } from '@xyflow/react';

const STATUS_COLORS = {
  pending: '#9CA3AF', scheduled: '#60A5FA', running: '#3B82F6',
  completed: '#10B981', failed: '#EF4444', retrying: '#F59E0B',
  paused: '#8B5CF6', skipped: '#D1D5DB', mutated: '#EAB308',
};

function AgentNode({ data }: NodeProps) {
  return (
    <div className={`agent-node status-${data.status}`}
         style={{ borderColor: STATUS_COLORS[data.status] }}>
      <Handle type="target" position={Position.Top} />
      <div className="node-header">
        <span className={`status-dot ${data.status}`} />
        <span>{data.role}</span>
      </div>
      {data.skills && (
        <div className="node-skills">
          {data.skills.map((s: string) => <span key={s} className="badge">{s}</span>)}
        </div>
      )}
      {data.status === 'completed' && data.output?.summary && (
        <div className="node-output">{data.output.summary.slice(0, 60)}...</div>
      )}
      {data.metrics?.cost_usd > 0 && (
        <div className="node-meta">${data.metrics.cost_usd.toFixed(3)}</div>
      )}
      <Handle type="source" position={Position.Bottom} />
    </div>
  );
}

// MUST define outside component (or useMemo) to avoid re-registration
const nodeTypes = { agentNode: AgentNode };

ELKjs Auto-Layout Hook

typescript
import ELK from 'elkjs/lib/elk.bundled.js';
import { useCallback } from 'react';
import { useReactFlow } from '@xyflow/react';

const elk = new ELK();

export function useAutoLayout() {
  const { fitView } = useReactFlow();

  return useCallback(async (nodes: Node[], edges: Edge[], direction = 'DOWN') => {
    const isHorizontal = direction === 'RIGHT';
    const layouted = await elk.layout({
      id: 'root',
      layoutOptions: {
        'elk.algorithm': 'layered',
        'elk.direction': direction,
        'elk.spacing.nodeNode': '80',
        'elk.layered.spacing.nodeNodeBetweenLayers': '100',
        'elk.edgeRouting': 'ORTHOGONAL',
      },
      children: nodes.map((n) => ({
        ...n,
        targetPosition: isHorizontal ? 'left' : 'top',
        sourcePosition: isHorizontal ? 'right' : 'bottom',
        width: n.measured?.width ?? 220,
        height: n.measured?.height ?? 120,
      })),
      edges,
    });
    const result = layouted.children!.map((elkN) => ({
      ...nodes.find((n) => n.id === elkN.id)!,
      position: { x: elkN.x!, y: elkN.y! },
    }));
    window.requestAnimationFrame(() => fitView());
    return result;
  }, [fitView]);
}

Dashboard Assembly

typescript
import { ReactFlow, ReactFlowProvider, Panel } from '@xyflow/react';
import '@xyflow/react/dist/style.css';

function DAGDashboard({ dagId }: { dagId: string }) {
  const { nodes, edges, onNodesChange, onEdgesChange } = useDAGStore();
  const layout = useAutoLayout();

  // WebSocket → Zustand (see websocket-streaming skill)
  useDAGStream(dagId);

  return (
    <ReactFlow nodes={nodes} edges={edges} nodeTypes={nodeTypes}
      onNodesChange={onNodesChange} onEdgesChange={onEdgesChange} fitView>
      <Panel position="top-right">
        <button onClick={() => layout(nodes, edges, 'DOWN')}>↓ Vertical</button>
        <button onClick={() => layout(nodes, edges, 'RIGHT')}>→ Horizontal</button>
      </Panel>
    </ReactFlow>
  );
}

export default function DAGPage({ dagId }: { dagId: string }) {
  return <ReactFlowProvider><DAGDashboard dagId={dagId} /></ReactFlowProvider>;
}

v12 Gotchas

Pitfall Fix
nodeTypes defined inside component → infinite re-render Define OUTSIDE component or wrap in useMemo
State update doesn't trigger re-render Must create NEW node object: { ...node, data: { ...node.data, ...update } }
xPos/yPos in custom node → undefined Use positionAbsoluteX/positionAbsoluteY (v12 rename)
nodeInternals → undefined Use nodeLookup (v12 rename)
ELK layout ignores node size Pass node.measured?.width and height explicitly
fitView fires before DOM paint Wrap in requestAnimationFrame(() => fitView())
Interactive elements drag the node Add className="nodrag" to inputs, buttons, selects

Anti-Patterns

Canvas Rendering for Debugging

Wrong: Using canvas-based libraries (GoJS) where you can't inspect nodes in dev tools. Right: ReactFlow renders SVG + HTML. Every node is inspectable in React DevTools and the DOM.

Re-running Layout on Every State Update

Wrong: Calling ELK layout every time a node's status changes (expensive, causes visual jitter). Right: Only re-layout when topology changes (add/remove node/edge). Status color changes are just data updates — no layout needed.

Monolithic Node Component

Wrong: One giant node component handling all node types. Right: Register separate node types: agentNode, humanGateNode, pluripotentNode. Each is a focused React component.

Expand your agent's capabilities with these related and highly-rated skills.

curiositech/some_claude_skills

3d-cv-labeling-2026

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

81 12
Explore
curiositech/some_claude_skills

project-management-guru-adhd

Expert project manager for ADHD engineers managing multiple concurrent projects. Specializes in hyperfocus management, context-switching minimization, and parakeet-style gentle reminders. Activate on 'ADHD project management', 'context switching', 'hyperfocus', 'task prioritization', 'multiple projects', 'productivity for ADHD', 'task chunking', 'deadline management'. NOT for neurotypical project management, rigid waterfall processes, or general productivity advice without ADHD context.

81 12
Explore
curiositech/some_claude_skills

large-scale-map-visualization

Master of high-performance web map implementations handling 5,000-100,000+ geographic data points. Specializes in Leaflet.js optimization, Supercluster algorithms, viewport-based loading, canvas rendering, and progressive disclosure UX patterns.

81 12
Explore
curiositech/some_claude_skills

adhd-design-expert

Designs digital experiences for ADHD brains using neuroscience research and UX principles. Expert in reducing cognitive load, time blindness solutions, dopamine-driven engagement, and compassionate design patterns. Activate on 'ADHD design', 'cognitive load', 'accessibility', 'neurodivergent UX', 'time blindness', 'dopamine-driven', 'executive function'. NOT for general accessibility (WCAG only), neurotypical UX design, or simple UI styling without ADHD context.

81 12
Explore
curiositech/some_claude_skills

liaison

Translate multi-agent ecosystem activity into human-readable status briefings, decision requests, and progress summaries. Use for 'status update', 'brief me', 'what happened', 'summarize progress'. NOT for project planning (use project-management-guru-adhd), code review, or technical documentation.

81 12
Explore
curiositech/some_claude_skills

windows-95-web-designer

Modern web applications with authentic Windows 95 aesthetic. Gradient title bars, Start menu paradigm, taskbar patterns, 3D beveled chrome. Extrapolates Win95 to AI chatbots, mobile UIs, responsive layouts. Activate on 'windows 95', 'win95', 'start menu', 'taskbar', 'retro desktop', '95 aesthetic', 'clippy'. NOT for Windows 3.1 (use windows-3-1-web-designer), vaporwave/synthwave, macOS, flat design.

81 12
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results