Custom Commands

Creating Commands

Method 1: Markdown Files

Create ~/.config/abov3/command/test.md:

---
description: Run tests with coverage
agent: build
model: anthropic/claude-3-sonnet
---

Run the full test suite with coverage reporting.
Show any failing tests and suggest fixes.

$ARGUMENTS

Use in TUI: /test unit tests for auth module

Method 2: JSON Configuration

Add to ~/.config/abov3/abov3.json:

{
  "command": {
    "review": {
      "template": "Review this code for best practices:\n$ARGUMENTS",
      "description": "Code review assistant",
      "agent": "plan",
      "model": "anthropic/claude-3-opus"
    },
    "optimize": {
      "template": "Optimize this code for performance:\n$ARGUMENTS",
      "description": "Performance optimizer"
    }
  }
}

Command Configuration Options

Template Features

Dynamic Variables

---
template: |
  Debug this error:
  $ARGUMENTS
---

Shell Command Integration

---
template: |
  Analyze the current git status:
  `git status`

  Recent commits:
  `git log --oneline -5`

  $ARGUMENTS
---

File Content Inclusion

---
template: |
  Review this configuration:
  @package.json

  Check for:
  $ARGUMENTS
---

Example Commands

---
description: Deploy application to production
agent: build
model: anthropic/claude-3-opus
---

Deploy the application following these steps:
1. Run tests: `npm test`
2. Build: `npm run build`
3. Check deployment config: @deploy.config.js

$ARGUMENTS

Ensure all tests pass before deployment.

---
description: Security audit for current project
agent: plan
temperature: 0.3
---

Perform a security audit on:
`find . -name "*.js" -o -name "*.ts" | head -20`

Check for:
- Exposed credentials
- SQL injection vulnerabilities
- XSS risks
- Insecure dependencies: `npm audit`

$ARGUMENTS

---
description: Suggest refactoring improvements
agent: plan
model: anthropic/claude-3-sonnet
---

Analyze this code for refactoring opportunities:

$ARGUMENTS

Consider:
- Design patterns
- Code duplication
- Performance improvements
- Readability enhancements
- Modern JavaScript/TypeScript features

---
description: Generate TODO list from code comments
subtask: true
---

Find all TODO comments in the codebase:
`grep -r "TODO\|FIXME\|HACK" --include="*.js" --include="*.ts"`

Organize them by:
1. Priority (FIXME > TODO > HACK)
2. File location
3. Estimated effort

$ARGUMENTS

Advanced Patterns

Conditional Logic

---
template: |
  Check environment: `echo $NODE_ENV`

  If production:
    - Verify backups
    - Check monitoring

  If development:
    - Enable debug mode
    - Skip optimizations

  $ARGUMENTS
---

Multi-File Analysis

---
template: |
  Analyze the API structure:

  Routes: @src/routes/index.js
  Controllers: @src/controllers/index.js
  Models: @src/models/index.js

  Suggest improvements for: $ARGUMENTS
---

Interactive Workflows

---
description: Create new feature
agent: build
---

Create a new feature: $ARGUMENTS

Steps:
1. Create feature branch
2. Generate boilerplate: `ls src/templates/`
3. Write tests first (TDD)
4. Implement the feature
5. Update documentation

Ask me questions if you need clarification.

Best Practices

Command Discovery

In TUI

• Type / to see all available commands
• Use /help to see command descriptions
• Tab completion works for command names

Command Locations

Commands are loaded in this priority order:

1. .abov3/command/ - Project-specific
2. ~/.config/abov3/command/ - User global
3. abov3.json - Configuration file
4. Built-in commands