Lesson 22Advanced

Agent SDK

Programmatic access

SDK vs CLI: When to Use Which

CLI (Interactive)

• Interactive development
• Debugging and exploration
• One-off tasks
• Learning and experiments

SDK (Programmatic)

• CI/CD automation
• Batch processing
• App integration
• Custom workflows

Headless Mode

Headless mode allows running Claude Code without interactive input:

# Headless mode - no interactive prompts
claude --headless --prompt "Fix all TypeScript errors" --output result.json

# With specific model
claude --headless --model sonnet --prompt "Add tests for UserService"

# With MCP servers
claude --headless --mcp-config ./mcp.json --prompt "Query database for users"

--headlessNo interactive prompts

--promptTask to execute

--outputOutput file

--modelModel selection

SDK Examples

from anthropic_claude_code import ClaudeCode

# Initialize Claude Code
claude = ClaudeCode()

# Run a task
result = claude.run(
    prompt="Fix the failing tests in auth.ts",
    working_dir="/path/to/project"
)

print(result.output)
print(f"Cost: ${result.cost}")

GitHub Actions

Integrate Claude Code into CI/CD via GitHub Actions:

name: Code Review with Claude

on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Run Claude Code Review
        uses: anthropic/claude-code-action@v1
        with:
          prompt: |
            Review this PR for:
            - Code quality issues
            - Security vulnerabilities
            - Performance problems

            Provide specific feedback with line numbers.
          github-token: ${{ secrets.GITHUB_TOKEN }}
          anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}

SDK Use Cases

Automatic PR Code Review

Run review on every PR

PR opened

Test Generation

Auto-write tests for new code

New file

Documentation

Update docs on changes

Merge to main

DB Migrations

Generate migrations from description

Manual

SDK Usage Tips

1.Use --max-turns to limit iterations
2.Handle errors and timeouts
3.Monitor costs via result.cost
4.Store API keys in secrets

This lesson is part of a structured LLM course.

My Learning Path