<aside> 🚀

Context-Portable Project Template

Stop fighting boilerplate. Stop repeating context to your AI. Start building.

</aside>

An input-driven, AI-ready engineering starter that generates exactly the tech stack you need—and equips it with a cross-tool context synchronization system so your repo’s “truth” follows you across Claude, Copilot, Cursor, Gemini, and beyond.

• 🔗 Repository: View on GitHub

🛑 The Problem

1. Boilerplate bloat

   Most starters assume too much (language/framework/build tooling), so you spend day 1 deleting decisions you didn’t make.

2. AI context silos

   Teams use multiple AI tools. Switching tools or onboarding someone new forces you to re-explain architecture, commands, and safety rules—causing token waste and context drift.

💡 The Solution

This template stays generic until you answer a few setup questions.

When you bootstrap a new repo, it adds only what you ask for:

• ✅ Only the requested tech stack files
• ✅ Only the requested build-tool support
• ✅ Only the requested AI tool integrations

More importantly: it introduces a Context-Portable AI Memory System so your rules and workflows stay consistent across tools.

✨ Key Features

🛠️ 1) Input-driven scaffolding

Generate a tailored repo on day one.

• Tech stacks: typescript-node, react-vite, python-fastapi, java-spring-boot
• Build tools: npm, pnpm, poetry, gradle, maven

🧠 2) Universal AI tool support

Your AI knows your project regardless of the tool.

• Claude Code (CLAUDE.md + safety hooks)
• Cursor (.cursor/rules/repository-context.mdc)
• GitHub Copilot (.github/copilot-instructions.md)
• Gemini / Codex / JetBrains Junie (and more)

🔄 3) Zero context drift (sync engine)

Write rules once in AGENTS.md, generate tool-specific instructions automatically.

🔒 4) Private local memory layer

Keep personal workflow memory in ai/local/ (git-ignored) and merge locally when needed.