claude-api

Implement and debug Anthropic SDK applications with correct Claude model parameters.

1|Updated May 17, 2026
One-click install
npx skills add https://github.com/mykcs/myk-skills --skill claude-api-mykcs
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-api
Source: https://github.com/mykcs/myk-skills/tree/main/claude-api
Command: npx skills add https://github.com/mykcs/myk-skills --skill claude-api-mykcs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides Claude API guidance for building, debugging, and migrating Anthropic/Claude SDK applications, with strict rules to ensure the code uses the correct Claude surface and model parameters.

Core Features & Use Cases

  • SDK-correct Claude integration: Ensures you call Claude via the official Anthropic SDK (or raw HTTP only when explicitly requested) and prevents mixing OpenAI-style code into Claude implementations.
  • Model- and parameter-aware troubleshooting: Enforces safe defaults (e.g., Opus 4.7 as claude-opus-4-7) and highlights breaking changes like adaptive thinking requirements and removed sampling/budget parameters.
  • Language-aware implementation support: Detects the project language from repo files and routes you to the appropriate language-specific Claude API reference.
  • Tooling, streaming, and advanced API patterns: Covers structured outputs, tool use loops, streaming/SSE handling guidance, prompt caching, compaction, and managed-agents concepts.

Quick Start

Ask: "Integrate the Claude API into my project, using the official Anthropic SDK in TypeScript, and stream the response while following your model and thinking defaults."

Frequently Asked Questions about claude-api

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

FAQPage Schema
How do I integrate the Claude API using the official Anthropic SDK?

Streaming Claude API responses requires handling SSE (Server-Sent Events) through the official Anthropic SDK. This skill provides streaming and SSE handling guidance to correctly process response chunks while adhering to Claude's adaptive thinking defaults and parameter constraints.

What are the breaking changes when migrating to newer Claude API models?

Claude API model migration introduces breaking changes like adaptive thinking requirements and removed sampling or budget parameters. This skill enforces migration-aware parameter handling, ensuring you use exact model IDs and apply safe defaults during new builds or feature additions.

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

Implementing tool use with the Claude API requires structuring tool definitions and managing iterative tool use loops via the Anthropic SDK. This skill covers advanced API patterns including tool use, structured outputs, and prompt caching for robust application behavior.

Does the Claude API support structured outputs and prompt caching?

Yes, the Claude API supports structured outputs and prompt caching through the official Anthropic SDK. This skill provides guidance on these advanced API patterns alongside compaction and managed-agents concepts for building efficient, safe Claude applications.

Why is my Claude API code failing after switching from OpenAI?

Claude API code fails when OpenAI-style parameters and code structures are mixed into Anthropic SDK implementations. This skill enforces strict Claude-vs-OpenAI provider separation, highlighting removed sampling parameters and adaptive thinking requirements to fix migration errors.