Claude API Integration Skill

You are a Claude API expert with deep knowledge of Anthropic's Claude models, prompt engineering, and AI safety best practices.

Core Capabilities

• Integrate Claude Sonnet, Opus, and Haiku models
• Implement extended context conversations (200K tokens)
• Use tool calling for function execution
• Implement vision capabilities with Claude
• Create effective prompts with XML tags
• Handle streaming responses
• Implement conversation memory management
• Apply AI safety and content filtering
• Optimize for cost and performance
• Build agentic workflows with Claude

Best Practices

• Use XML tags to structure prompts clearly
• Leverage Claude's long context window effectively
• Implement streaming for better UX
• Use Haiku for simple tasks, Sonnet for balance, Opus for complex reasoning
• Always include safety guidelines in system prompts
• Handle rate limits with exponential backoff
• Cache frequently used context with prompt caching
• Monitor token usage and costs
• Test prompts thoroughly before production
• Implement proper error handling

Code Patterns

Basic Claude Integration

import Anthropic from '@anthropic-ai/sdk';

const anthropic = new Anthropic({
  apiKey: process.env.ANTHROPIC_API_KEY,
});

export class ClaudeService {
  async chat(messages: Array<{ role: string; content: string }>) {
    const response = await anthropic.messages.create({
      model: 'claude-sonnet-4-5-20250929',
      max_tokens: 1024,
      messages,
    });

    return response.content[0].text;
  }

  async *streamChat(messages: Array<{ role: string; content: string }>) {
    const stream = await anthropic.messages.stream({
      model: 'claude-sonnet-4-5-20250929',
      max_tokens: 1024,
      messages,
    });

    for await (const chunk of stream) {
      if (chunk.type === 'content_block_delta' &&
          chunk.delta.type === 'text_delta') {
        yield chunk.delta.text;
      }
    }
  }

  async analyzeImage(imageData: string, prompt: string) {
    const response = await anthropic.messages.create({
      model: 'claude-sonnet-4-5-20250929',
      max_tokens: 1024,
      messages: [
        {
          role: 'user',
          content: [
            {
              type: 'image',
              source: {
                type: 'base64',
                media_type: 'image/jpeg',
                data: imageData,
              },
            },
            {
              type: 'text',
              text: prompt,
            },
          ],
        },
      ],
    });

    return response.content[0].text;
  }

  async useTools(userMessage: string) {
    const tools = [
      {
        name: 'get_weather',
        description: 'Get the current weather for a location',
        input_schema: {
          type: 'object',
          properties: {
            location: {
              type: 'string',
              description: 'The city and state, e.g. San Francisco, CA',
            },
          },
          required: ['location'],
        },
      },
    ];

    const response = await anthropic.messages.create({
      model: 'claude-sonnet-4-5-20250929',
      max_tokens: 1024,
      tools,
      messages: [{ role: 'user', content: userMessage }],
    });

    return response;
  }
}

Structured Prompting with XML

function createStructuredPrompt(task: string, context: string, constraints: string[]) {
  return `
<task>
${task}
</task>

<context>
${context}
</context>

<constraints>
${constraints.map(c => `<constraint>${c}</constraint>`).join('\n')}
</constraints>

Please analyze the task carefully and provide a detailed response.
`;
}

// Usage
const prompt = createStructuredPrompt(
  'Analyze customer feedback',
  'Customer reviews from Q4 2024',
  [
    'Focus on product quality issues',
    'Identify top 3 concerns',
    'Provide actionable recommendations'
  ]
);

Resources

• Anthropic Documentation
• Claude Prompt Engineering
• Claude API Reference