Stack Customization

VibeReady ships with a default tech stack (Prisma, Clerk, Vitest, React Query, shadcn/ui, and more), but every technology is designed to be swappable. The framework separates universal principles from library-specific implementation, so your architectural decisions survive any technology swap.

---

How It Works

Each technology’s patterns live in isolated, clearly marked sections across the codebase:

• Rule files use a Principles section (universal, technology-agnostic) and an Implementation section (swappable, library-specific)
• Skill references contain technology-specific patterns in dedicated reference directories
• Stack manifest maps each technology to all files containing its patterns, so the AI knows exactly what to update

This separation means you can replace any library without losing the architectural decisions that make your codebase consistent.

---

Swapping a Technology

To swap a technology (e.g., Prisma to Drizzle, Clerk to NextAuth):

1. Invoke the stack-swap skill with what you want to swap
2. The AI reads the stack manifest to find all affected files
3. It keeps the Principles sections and rewrites the Implementation sections
4. It updates skill references, the manifest, and regenerates per-tool rules

Example prompt:

Use the stack-swap skill to replace Prisma with Drizzle ORM.

The skill follows a structured multi-step process documented in the stack manifest. See Prompt Library and Skills for the full Agent Skills reference.

---

What Stays, What Changes

Stays (Principles)	Changes (Implementation)
“All queries scoped by organizationId”	Prisma where clause becomes Drizzle where clause
”Auth check on every protected route”	auth() from Clerk becomes getServerSession() from NextAuth
”Structured logging with context”	Pino child loggers become Winston child loggers

Your architectural decisions survive the swap. Only the library-specific code changes.

---

The Principles/Implementation Split in Practice

This split is not just a documentation convention — it is the mechanism that makes stack swaps safe. When the stack-swap skill runs:

1. Principles are preserved — Multi-tenancy rules, TDD requirements, security patterns, and design rationale remain untouched.
2. Implementation is rewritten — Code examples, import paths, API calls, and configuration are updated to the new library.
3. Rules are regenerated — The setup command regenerates scoped rules for all configured AI tools so they immediately use the new patterns.

For more on maintaining this split when editing rule files, see Keeping Docs Synced.

---

Related Documentation

• AI Development Framework — Scoped rules and the 14 rule areas
• Keeping Docs Synced — Maintaining the principles/implementation split
• Prompt Library and Skills — The stack-swap skill and invocation guide