Add MIT License to the project
TanStack Agent Skills
Comprehensive best practices for building applications with the TanStack ecosystem. These skills provide AI coding agents with structured guidelines for TanStack Query, TanStack Router, and TanStack Start.
Skills follow the Agent Skills format.
Available Skills
tanstack-query
Best practices for data fetching, caching, mutations, and server state management with TanStack Query (React Query).
Use when:
- Fetching or caching server data with
useQueryoruseMutation - Setting up React Query for a new project
- Reviewing data fetching patterns for performance issues
32+ rules across 8 categories:
- Query Keys (CRITICAL) — Proper key structure and organization
- Caching (CRITICAL) — staleTime, gcTime, and invalidation patterns
- Mutations (HIGH) — Optimistic updates, error handling
- Error Handling (HIGH) — Error boundaries, retry logic
- Prefetching (MEDIUM) — Intent-based and route prefetching
- Infinite Queries (MEDIUM) — Pagination patterns
- SSR Integration (MEDIUM) — Hydration and dehydration
- Performance (LOW) — Select transforms, memoization
tanstack-router
Best practices for type-safe routing, data loading, search params, and navigation with TanStack Router.
Use when:
- Setting up routing for a React app
- Working with search params, loaders, or code splitting
- Configuring type-safe navigation
36 rules across 10 categories:
- Type Safety (CRITICAL) — Router registration, type narrowing, context typing
- Route Organization (CRITICAL) — File-based routing, route trees, pathless layouts
- Router Config (HIGH) — Global router defaults
- Data Loading (HIGH) — Loaders, Query integration, deferred data
- Search Params (HIGH) — Validation, type inheritance, middleware, defaults
- Error Handling (HIGH) — Not-found routes, error boundaries
- Navigation (MEDIUM) — Link component, active states, programmatic navigation
- Code Splitting (MEDIUM) — Lazy routes, auto splitting, critical path
- Preloading (MEDIUM) — Intent-based, stale time, manual preloading
- Route Context (LOW) — Dependency injection, beforeLoad
tanstack-start
Best practices for full-stack React applications with TanStack Start, including server functions, middleware, SSR, and authentication.
Use when:
- Building full-stack apps with server functions
- Implementing authentication or middleware
- Configuring SSR, streaming, or deployment
32 rules across 10 categories:
- Server Functions (CRITICAL) — createServerFn patterns
- Security (CRITICAL) — Input validation, CSRF protection
- Middleware (HIGH) — Request/function middleware
- Authentication (HIGH) — Sessions, route protection
- API Routes (MEDIUM) — External endpoint patterns
- SSR (MEDIUM) — Hydration safety, streaming, selective SSR
- Error Handling (MEDIUM) — Server errors, redirects, not-found
- Environment (MEDIUM) — Configuration management
- File Organization (LOW) — Code separation patterns
- Deployment (LOW) — Environment config
tanstack-integration
Best practices for integrating TanStack Query with TanStack Router and TanStack Start together.
Use when:
- Combining Query + Router in the same app
- Setting up SSR with data prefetching
- Wiring loaders to query cache
13 rules across 4 categories:
- Setup (CRITICAL) — QueryClient context, provider wrapping
- Data Flow (HIGH) — Loader + Query patterns
- Caching (MEDIUM) — Single source of truth
- SSR (LOW) — Dehydration/hydration
Installation
npx skills add DeckardGer/tanstack-agent-skills
Or manually add to your project's .cursor/ or Claude Code configuration.
Usage
Skills are automatically available once installed. The agent will use them when relevant tasks are detected.
Skill Structure
Each skill follows the Agent Skills standard:
skills/
├── tanstack-query/
│ ├── SKILL.md # Main skill instructions
│ └── rules/ # Individual rule files
│ ├── qk-array-structure.md
│ ├── cache-stale-time.md
│ └── ...
├── tanstack-router/
│ ├── SKILL.md
│ └── rules/
├── tanstack-start/
│ ├── SKILL.md
│ └── rules/
└── tanstack-integration/
├── SKILL.md
└── rules/
Rule File Structure
Each rule file contains:
- Priority — CRITICAL, HIGH, MEDIUM, or LOW
- Explanation — Why this pattern matters
- Bad Example — Anti-pattern to avoid with explanation
- Good Example — Recommended implementation
- Context — When to apply or skip the rule
Contributing
Contributions welcome! Please ensure:
- Rules are practical and battle-tested
- Examples are clear and runnable
- Priority levels are appropriate
- Context helps agents decide applicability
Agent Compatibility
Works with any AI coding agent that supports AGENTS.md, including Claude Code, Codex, Opencode, and others.
Resources
License
MIT — see LICENSE for details.