nestjs-best-practices

Agent skills are structured knowledge bases that help AI coding assistants (like Claude, GPT, Cursor, etc.) understand framework-specific patterns, best practices, and common pitfalls. They bridge the gap between an AI's general knowledge and the specific, up-to-date expertise needed for modern development.

Why agent skills matter:

• Stay current - Frameworks evolve faster than AI training data. Skills provide the latest patterns and syntax.
• Reduce errors - Skills teach AI about breaking changes, deprecations, and migration paths.
• Best practices - Encode community-vetted solutions to common problems.
• Type safety - Include TypeScript patterns and proper typing for modern codebases.

Framework Skills

| Skill | Description | Install |

|-------|-------------|---------|

| [nestjs](./nestjs) | NestJS 11+ DI, validation, auth, TypeORM/Prisma/Drizzle, and testing patterns | npx add-skill ejirocodes/agent-skills/nestjs |

| [svelte](./svelte) | Svelte 5 runes, snippets, and SvelteKit patterns | npx add-skill ejirocodes/agent-skills/svelte |

| [vue](./vue) | Vue 3+ TypeScript, Volar, Pinia testing, and component patterns | npx add-skill ejirocodes/agent-skills/vue |

Exa.ai Skills

| Skill | Description | Install |

|-------|-------------|---------|

| [exa-search](./exa/skills/exa-search) | Core Exa search API integration (neural/keyword/auto modes, filters, content retrieval) | npx add-skill ejirocodes/agent-skills/exa-search |

| [exa-rag](./exa/skills/exa-rag) | RAG pipelines with Exa (LangChain, LlamaIndex, Vercel AI SDK, MCP tools) | npx add-skill ejirocodes/agent-skills/exa-rag |

| [exa-research](./exa/skills/exa-research) | Deep research and Answer API with citations and streaming | npx add-skill ejirocodes/agent-skills/exa-research |

| [exa-entities](./exa/skills/exa-entities) | Company and people search for lead gen, recruiting, and competitive intelligence | npx add-skill ejirocodes/agent-skills/exa-entities |

Install a skill using the add-skill CLI:

```bash

npx add-skill ejirocodes/agent-skills

```

For example, to install Svelte skills:

```bash

npx add-skill ejirocodes/agent-skills/svelte

```

For best results, prefix your prompt with use skill:

```

Use svelte skill, help me create a form component with validation

```

This explicitly triggers the skill and ensures the AI follows the documented patterns. Without the prefix, skill triggering depends on how closely your prompt matches the skill's keywords.

Each skill follows this structure:

```

/

├── LICENSE # MIT License

└── skills/

└── /

├── SKILL.md # Skill metadata and quick reference

└── references/

└── *.md # Domain-specific reference files

```

Reference Organization

References are organized by domain for progressive disclosure:

• SKILL.md - Lean entry point with essential patterns and navigation
• references/ - Detailed documentation loaded only when needed

This structure keeps the main skill file small while providing comprehensive coverage through domain-specific reference files (e.g., runes.md, sveltekit.md, migration.md).

Contributions are welcome! To add a new skill:

1. Create a folder for your skill (e.g., react/)
2. Follow the structure above
3. Include comprehensive rules with:

- Clear problem descriptions

- Code examples (incorrect vs correct)

- Links to official documentation

1. Submit a pull request

Guidelines

• Focus on problems AI assistants commonly get wrong
• Prioritize recent framework changes and breaking updates
• Include TypeScript patterns where applicable
• Test rules with multiple AI models (Claude, GPT, etc.)
• Delete rules if AI consistently solves them without help

Future skills planned:

• Frontend - React 19, Angular, Solid
• Backend - Node.js, Deno, Bun, Go, Rust
• Databases - PostgreSQL, MongoDB, Redis, Prisma
• DevOps - Docker, Kubernetes, CI/CD
• Testing - Vitest, Playwright, Jest
• AI/ML - LangChain, Vector DBs, more RAG patterns

MIT - see [LICENSE](./LICENSE)

---

Created by [Ejiro Asiuwhu](https://github.com/ejirocodes)