# AI Prompts Directory

This directory contains reusable prompt templates for common development tasks with AI coding assistants.

## Available Prompts

| Prompt                           | Purpose                         | Use When                           |
| -------------------------------- | ------------------------------- | ---------------------------------- |
| [code-review.md](code-review.md) | Code review against standards   | Reviewing new code or PRs          |
| [refactoring.md](refactoring.md) | Refactor code to meet standards | Improving existing code            |
| [new-feature.md](new-feature.md) | Implement new features          | Adding functionality               |
| [debugging.md](debugging.md)     | Debug issues                    | Fixing bugs or unexpected behavior |

## How to Use

### 1. Read the Relevant Prompt File

Each prompt file contains:

- General template
- Language-specific variations
- Examples
- LazyVim workflows
- Tips and best practices

### 2. Customize for Your Needs

Copy the template and fill in:

- Your specific code/issue
- Language and context
- Requirements and constraints

### 3. Use with Your AI Assistant

- **Claude.ai**: Paste prompt in chat
- **Claude Code**: Reference in command
- **GitHub Copilot**: Use in comments
- **Cursor**: Use in chat
- **Continue.dev**: Use slash commands

## Quick Examples

### Code Review

```
Please review this code following .ai/context/AGENTS_GO.md:

[paste Go code]

Check:
- Interface/concrete pattern
- Error handling
- Test coverage
```

### Refactoring

```
Refactor this Python code to add type hints per .ai/context/AGENTS_PYTHON.md:

[paste Python code]

Requirements:
- Pass mypy strict mode
- Use Pydantic models
- Update tests
```

### New Feature

```
Implement a caching layer following .ai/context/AGENTS_GO.md:

Requirements:
- Accept Cache interface
- Functional options for config
- TTL support
- Table-driven tests
```

### Debugging

```
Debug this TypeScript error:

[code + error]

Check against .ai/context/AGENTS_TYPESCRIPT_VUE.md:
- Are types correct?
- Is Composition API proper?
```

## Prompt Engineering Tips

### Be Specific

❌ "Fix this code"
✅ "Refactor this to follow the 'accept interfaces, return concrete' pattern from .ai/context/AGENTS_GO.md"

### Reference Standards

Always point to the relevant AGENTS file:

- `.ai/context/AGENTS_GO.md`
- `.ai/context/AGENTS_PYTHON.md`
- `.ai/context/AGENTS_TYPESCRIPT_VUE.md`
- `.ai/context/AGENTS_POWERSHELL.md`
- `.ai/context/AGENTS_BASH.md`

### Include Context

Provide:

- Language/framework versions
- Environment (dev/prod)
- Error messages (complete)
- What you've tried

### Request LazyVim Help

Ask for editor workflow from `.ai/context/EDITORS.md`:

- Keybindings to use
- Available snippets
- LSP actions
- Which-key commands

### Ask for Tests

Always request tests following patterns:

- Table-driven (Go)
- Parametrized (Python)
- Component tests (Vue)
- Pester (PowerShell)

## Multi-Turn Conversations

### Start Broad

```
I need to implement user authentication. What approach should I take following .ai/context/AGENTS_*.md patterns?
```

### Get Specific

```
Great, now implement the JWT token service in Go following those patterns.
```

### Iterate

```
Add tests for the edge cases you mentioned, following the table-driven pattern.
```

### Refine

```
Show me the LazyVim workflow for debugging this with DAP.
```

## Combining Prompts

You can combine multiple prompt types:

```
1. Review this code against standards (code-review.md)
2. If issues found, refactor to fix (refactoring.md)
3. Add the caching feature (new-feature.md)
4. Help debug the timeout issue (debugging.md)
```

## Language-Specific Workflows

### Go Development

1. **New Feature**: Use new-feature.md template
   - Request interfaces + concrete types
   - Functional options
   - Table-driven tests

2. **Review**: Use code-review.md
   - Check interface pattern
   - Verify error handling
   - Confirm context usage

3. **Debug**: Use debugging.md
   - Check for race conditions
   - Verify defer usage
   - Test with `-race` flag

### Python Development

1. **New Feature**: Use new-feature.md
   - Request type hints
   - Pydantic models
   - Async patterns

2. **Review**: Use code-review.md
   - Run mypy check
   - Verify uv usage
   - Check docstrings

3. **Refactor**: Use refactoring.md
   - Add missing type hints
   - Convert to async
   - Add Pydantic validation

### Vue/TypeScript Development

1. **New Feature**: Use new-feature.md
   - Composition API
   - Typed props/emits
   - Composables

2. **Review**: Use code-review.md
   - Check TypeScript strict
   - Verify reactivity
   - Component structure

3. **Debug**: Use debugging.md
   - Reactivity issues
   - Type errors
   - Props/emits

### PowerShell Development

1. **New Feature**: Use new-feature.md
   - Multi-subscription
   - Graph API
   - Error handling

2. **Review**: Use code-review.md
   - Check Graph usage
   - Verify error handling
   - Approved verbs

3. **Refactor**: Use refactoring.md
   - Replace loops with Graph
   - Add error handling
   - Update tests

## Tool-Specific Usage

### Claude.ai

```
1. Upload relevant AGENTS_*.md file to Project
2. Use prompt templates
3. Reference uploaded files
```

### Cursor

```
1. Keep AGENTS files open in workspace
2. Use prompts in Cmd+K or chat
3. Reference .ai/context/ files
```

### Continue.dev

```
1. Use slash commands with prompts
2. /standards - Show all standards
3. Reference in @files
```

### GitHub Copilot

```
1. Use prompts in comments
2. Reference patterns in code
3. Keep AGENTS files open
```

## Creating New Prompts

To add a new prompt template:

1. Copy existing template structure
2. Add language-specific variations
3. Include examples
4. Reference AGENTS\_\*.md patterns
5. Add LazyVim workflow
6. Update this README

## Feedback

Improve these prompts:

1. Use them with AI assistants
2. Note what works/doesn't
3. Submit updates
4. Share variations

## Additional Resources

- [General Instructions](../instructions.md)
- [LazyVim Setup](../context/EDITORS.md)
- [Go Patterns](../context/AGENTS_GO.md)
- [Python Patterns](../context/AGENTS_PYTHON.md)
- [TypeScript/Vue Patterns](../context/AGENTS_TYPESCRIPT_VUE.md)
- [PowerShell Patterns](../context/AGENTS_POWERSHELL.md)
- [Bash Patterns](../context/AGENTS_BASH.md)

## Version History

- 2024-01-24: Initial version