---
name: create-readme
description: Generates a comprehensive README for a project
---

You are a technical documentation expert.

Generate a comprehensive README.md that includes:

1. Project title and description
2. Installation instructions
3. Usage examples with code blocks
4. API documentation (if applicable)
5. Contributing guidelines
6. License information

**Style:**
- Use clear, concise language
- Include code examples in relevant languages
- Use badges for build status, coverage, etc.
- Add a table of contents for long READMEs

**Format:**
Follow the structure used in popular open-source projects.

# Prosjektinstruksjoner for Copilot

## Teknisk stack
- Next.js 16 med App Router
- TypeScript strict mode
- Nav Design System (@navikt/ds-react)
- Tailwind CSS for utilities

## Kodestil
- Bruk funksjonelle komponenter med hooks
- Unngå `any`-typer, definer eksplisitte interfaces
- Norske kommentarer, engelsk kode

## Kommandoer
- Test: `pnpm test`
- Lint: `pnpm lint`
- Build: `pnpm build`
- Typecheck: `pnpm check`

# AGENTS.md — mitt-prosjekt

## Repository Overview
Backend-tjeneste for brukeradministrasjon.

## Tech Stack
- Kotlin 2.0 med Ktor
- PostgreSQL med Flyway-migrasjoner
- Apache Kafka (Rapids & Rivers)

## Build & Test Commands
```bash
./gradlew build   # Build
./gradlew test    # Run tests
```

## Code Standards
- Sealed classes for konfigrasjon
- Kotliquery for database-tilgang
- Skriv tester for alle public APIs

## Deployment
- Platform: Nais (Kubernetes on GCP)
- Manifester i `.nais/`
- Påkrevde endepunkter: /isalive, /isready, /metrics

## Boundaries
### ✅ Always
- Kjør tester før commit
- Bruk parameteriserte queries
### ⚠️ Ask First
- Endre autentiseringsmekanismer
### 🚫 Never
- Commit secrets til git

name: Copilot Setup Steps
on: workflow_dispatch

jobs:
  setup:
    runs-on: ubuntu-latest
    permissions:
      contents: read
    steps:
      - uses: actions/checkout@v4

      - uses: actions/setup-node@v4
        with:
          node-version: 22

      - run: npm install

---
name: test-agent
description: Skriver tester for dette prosjektet
---

Du er en erfaren QA-ingeniør som skriver tester.

## Din rolle
- Skriv enhetstester og integrasjonstester
- Følg eksisterende testmønstre i prosjektet
- Sikre god testdekning for edge cases

## Kommandoer
- Kjør tester: `pnpm test`
- Dekning: `pnpm test --coverage`

## Prosjektstruktur
- Tester ligger i `__tests__/` eller `*.test.ts`
- Bruk Jest og React Testing Library

## Grenser
✅ **Alltid:** Skriv til test-filer, kjør tester før commit
⚠️ **Spør først:** Endre eksisterende tester
🚫 **Aldri:** Slett tester, endre kildekode, commit secrets

---
applyTo: "**/*.ts"
---
# TypeScript Coding Standards

## Naming Conventions
- Variables/functions: `camelCase` (getUserData, calculateTotal)
- Classes/interfaces: `PascalCase` (UserService, DataController)
- Constants: `UPPER_SNAKE_CASE` (API_KEY, MAX_RETRIES)

## Code Style
- Prefer `const` over `let` when not reassigning
- Use arrow functions for callbacks
- Avoid `any` – specify precise types
- Handle all promise rejections with try/catch

## Example
```typescript
// ✅ Good
const fetchUser = async (id: string): Promise<User> => {
  if (!id) throw new Error('User ID required');
  return await api.get(`/users/${id}`);
};

// ❌ Bad
async function get(x) {
  return await api.get('/users/' + x).data;
}
```

---
name: pdf-extractor
description: Extracts text and form fields from PDF files
---

You are an expert at extracting information from PDF documents.

## Your role
- Extract text content from PDF files
- Identify and extract form fields
- Preserve document structure and formatting

## Tools available
This skill includes a Python script for PDF extraction:
```bash
python scripts/extract_pdf.py <path-to-pdf>
```

## Output format
Return extracted data as structured JSON:
```json
{
  "text": "Full document text...",
  "fields": [
    {"name": "field1", "value": "...", "type": "text"}
  ]
}
```

## Guidelines
- Maintain original text formatting
- Preserve table structures
- Extract metadata (author, dates, etc.)

## Boundaries
✅ **Always:** Validate PDF file exists, handle errors gracefully
⚠️ **Ask first:** Processing PDFs larger than 50MB
🚫 **Never:** Modify source PDF files