Agentic Coding with Claude Code
Deep dive into context engineering for AI agents. Features the shoreman script for unified logging, vite-console-forward-plugin for browser-to-terminal debugging, and error handling training to teach agents robust service management.
Key Highlights
Context Engineering Focus
Shoreman Script Integration
- Unified logging system that runs both backend and frontend services
- Creates single log file that AI agents can read and understand
- Manages multiple services with consolidated output
- Hides unnecessary information to prevent agent confusion
- Auto-reloads and auto-compiles with single command entry point
- Error handling training - teaches agents to handle “service already running” messages gracefully
- Prevents agents from killing processes unnecessarily for more reliable service management
- Essential for maintaining long-running development servers
Vite-Console-Forward-Plugin
- Forwards browser console logs directly to terminal
- Enables real-time debugging without switching between browser and terminal
- Creates unified development environment for AI agents
- Eliminates the need to paste logs back into claude code
Resources
- Shoreman Script: github.com/mitsuhiko/minibb/blob/main/scripts/shoreman.sh
- Vite Console Forward Plugin: github.com/mitsuhiko/vite-console-forward-plugin
This video demonstrates advanced context engineering techniques that create AI-readable environments, significantly improving agent reliability and reducing debugging complexity.