claude-api

Guide Anthropic SDK selection and request architecture for Claude-powered applications.

1|Updated Apr 13, 2026
One-click install
npx skills add https://github.com/yahav123147/paid-ads-cro-skills --skill claude-api-yahav123147
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-api
Source: https://github.com/yahav123147/paid-ads-cro-skills/tree/main/skills/claude-api
Command: npx skills add https://github.com/yahav123147/paid-ads-cro-skills --skill claude-api-yahav123147

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you build reliable LLM-powered applications by choosing the right Anthropic/Claude API surface, handling SDK/API differences, and following best practices for models, thinking, tools, streaming, and production-ready request patterns.

Core Features & Use Cases

  • API selection & architecture guidance: Choose between single-call Claude API, tool workflows, and Agent SDK based on what your app needs (e.g., file access, web/terminal tools, multi-step orchestration).
  • Model defaults & correctness rules: Uses Claude Opus 4.6 by default (claude-opus-4-6) and enforces exact model ID strings.
  • Implementation-ready patterns: Covers language-specific setup (Python/TypeScript/Java/Go/Ruby/C#/PHP/cURL), plus practical sections on streaming, structured outputs, prompt caching, compaction, batching, files, and tool use loops.

Quick Start

Tell the skill which language you’re using (Python, TypeScript, etc.) and what you’re building (single-call extraction, tool-using workflow, or an agent with built-in web/file/terminal access).

Frequently Asked Questions about claude-api

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I implement tool use loops with the Claude API?

Implement tool use loops with the Claude API by configuring the Anthropic SDK to handle stop reasons and tool definitions iteratively, ensuring correct SDK-specific API routing for multi-step agentic workflows.

How does prompt caching work for structured outputs in the Claude API?

Prompt caching for structured outputs works by building deterministic prompts with exact model ID strings like claude-opus-4-6, allowing the Claude API to cache and reuse static prompt prefixes for faster LLM responses.

What's the best way to set up streaming for my Claude API integration?

The best way to set up streaming for your Claude API integration is selecting the appropriate Anthropic SDK surface for your language and runtime, enabling continuous response generation for single-call and agentic applications.

Does the Claude API support batching and file processing for structured outputs?

The Claude API supports batching, file processing, and structured outputs by leveraging specific SDK request architectures, allowing applications to handle bulk data extraction and format enforcement safely.

Can I use the Claude API with Python, TypeScript, and other programming languages?

You can use the Claude API with Python, TypeScript, Java, Go, Ruby, C#, PHP, and cURL by applying language-specific setup patterns to configure single-call, tool-enabled, or agentic workflows safely.

Why does my Claude API request fail when using incorrect model IDs?

Claude API requests fail when using incorrect model IDs because the integration enforces exact model ID strings, defaulting to claude-opus-4-6, to ensure deterministic prompt-building and proper SDK behavior.