{"claude_md":"# Canvas Neural Network Visualizer\n\nAnimated neural network visualization with signal propagation, node activations, and weighted connections. Feed-forward architecture with sigmoid activation.\n\n## Quick Reference\n\n# Canvas Neural Network Visualizer\n\n> **Difficulty:** Intermediate | **Runtime:** HTML5 Canvas | **Output:** Animated visualization\n\nTeach an AI agent to visualize a feed-forward neural network with animated signal propagation, node activations, and weight connections.\n\n## What You'll Build\n\nA self-contained HTML file showing an animated neural network:\n- Configurable layer sizes (input → hidden → output)\n- Weighted connections drawn as lines (thickness = weight magnitude)\n- Signal pulses traveling along connections during forward pass\n- Node activation glow effect (brighter = higher activation)\n- Continuous forward pass animation with random inputs\n\n## Teaching Points\n\n- Neural network architecture visualization\n- Canvas arc drawing for nodes\n- Line thickness mapping for weight visualization\n- Particle/pulse animation along paths\n- Color interpolation for activation display\n\n## Agent Docs (read when relevant)\n\n| File | When to read |\n|------|-------------|\n| `agent_docs/neural-net.html` | neural net |\n| `agent_docs/README.md` | README |\n","agent_docs":{"agent_docs/neural-net.html":"\n\n\n\n\nNeural Network Visualizer\n\n\n\nFeed-Forward Neural Network — Forward Pass
\n\n\n\n","agent_docs/README.md":"# Neural Network Visualizer\n\nAnimated feed-forward neural network visualization with signal propagation.\n\n## Features\n- Configurable layer architecture (4→6→6→3)\n- Animated signal pulses traveling along weighted connections\n- Node activation glow based on sigmoid output\n- Color-coded positive (green) and negative (red) weights\n- Continuous forward pass with random inputs\n\n## Run\nOpen `neural-net.html` in any browser.\n"}}