What problem does it solve?
This Skill helps you choose the correct Anthropic/Claude API surface and wire it up correctly so you can build LLM-powered applications without trial-and-error.
Core Features & Use Cases
- Guided surface selection: Pick Claude API for single calls or workflows, use tool use for orchestration, and choose Agent SDK when you need built-in web/file/terminal capabilities.
- Language-aware implementation guidance: Detect your project language and follow the matching SDK instructions (Python, TypeScript, Java, Go, Ruby, C#, PHP) or use cURL/raw HTTP when needed.
- Production-ready setup defaults: Apply consistent model/streaming/thinking defaults (notably
claude-opus-4-6 and adaptive thinking) and steer toward safer patterns like streaming and structured outputs.
- Tooling and agentic patterns: Cover tool use, manual agentic loops, tool runners (where available), code execution, web fetch/search concepts, batching, files API, and compaction for long conversations.
- Reliability and cost controls: Use prompt caching, token counting guidance, batch processing for cost savings, and error-handling references.
Quick Start
Ask: Build a Claude API (Messages API) integration in my language, using claude-opus-4-6, streaming, and tool use only if I need multi-step actions.