claude-api

Select Claude API, tool use, or Agent SDK integration surfaces.

Updated Apr 6, 2026
One-click install
npx skills add https://github.com/daanteijema-beep/ai-gids-platform --skill claude-api-daanteijema-beep
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-api
Source: https://github.com/daanteijema-beep/ai-gids-platform/tree/main/.agents/skills/claude-api
Command: npx skills add https://github.com/daanteijema-beep/ai-gids-platform --skill claude-api-daanteijema-beep

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you choose the correct Anthropic/Claude integration path and implement it correctly (Claude API vs tool use vs Agent SDK).

Core Features & Use Cases

  • Surface selection: Pick between single-call, tool workflows, and Agent SDK based on whether Claude needs tools, files/web/terminal access, or multi-step orchestration.
  • Guardrailed defaults: Enforce model choice (defaulting to Claude Opus 4.6), adaptive thinking, and streaming guidance for long inputs/outputs.
  • Language-aware guidance: Detect the project language and route you to the appropriate language-specific Claude API/tooling documentation, including Go/Java/C#/PHP/Ruby.
  • Operational best practices: Cover streaming, structured outputs, tool loops, prompt caching, compaction, and model/capability lookup guidance.

Quick Start

Tell the AI: "I’m building a TypeScript app that uses the Claude API with tools—show me the correct setup, streaming pattern, and a tool-use loop for this use case."

Frequently Asked Questions about claude-api

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

FAQPage Schema
When should I use the Claude API versus the Agent SDK for my application?

Use the Claude API for single-call tasks, and switch to the Agent SDK when your workflow requires multi-step tool orchestration, or file, web, and terminal access. Surface selection depends on whether Claude needs tools to complete the task.

How do I implement a tool-use loop with the Anthropic SDK?

To implement a tool-use loop with the Anthropic SDK, configure your integration to handle iterative tool calls and responses until the task completes. The Skill provides operational best practices for structuring these multi-step tool workflows correctly.

Does the Anthropic SDK support streaming and structured outputs simultaneously?

Yes, the Anthropic SDK supports streaming and structured outputs simultaneously. The Skill provides guidance on applying both features together, along with prompt caching and compaction, to handle long inputs and outputs efficiently.

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

Yes, the Claude API supports multiple programming languages. The Skill detects your project language and routes you to the appropriate language-specific documentation, including Go, Java, C#, PHP, and Ruby.

What is the default model and adaptive thinking behavior when building Claude apps?

The default model enforced is Claude Opus 4.6, paired with adaptive thinking enabled by default. These guardrailed defaults ensure deterministic behavior across single-call tasks and complex agentic workflows.

How do I handle model selection and capability discovery in my Claude API integration?

Handle model selection and capability discovery by applying the Skill's enforced deterministic defaults and model lookup guidance. This ensures your integration queries the correct model capabilities before executing tool workflows.