shinpr/ai-coding-project-boilerplate

Implements a specialized multi-agent workflow where parent sessions delegate tasks to Claude Code sub-agents, each operating with isolated context to prevent exhaustion during large projects. Built-in quality checks (lint, type, test, build) run locally before commits, while a Skills system provides reusable knowledge modules for coding standards, TypeScript patterns, testing strategies, and project-specific context. Includes a CLI-based update mechanism to keep agent definitions, commands, and AI rules synchronized across team projects while preserving custom source code.

185 stars.