skills.rest
MrS-E
MrS-ECOMMUNITY
Explore all skills by this author →
GitHub

claude-api

Build, debug, and optimize applications using the Claude API and Anthropic SDKs.

Updated Sep 5, 2025
Search Tags:#streaming#tool use#llm integration#claude api#prompt caching#anthropic sdk#managed agents
One-click install
npx skills add https://github.com/MrS-E/config --skill claude-api-mrs-e
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: claude-api
Source: https://github.com/MrS-E/config/tree/main/junie/versions/1588.20/skills/claude-api
Command: npx skills add https://github.com/MrS-E/config --skill claude-api-mrs-e

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Writing correct Claude API and Anthropic SDK code requires up-to-date knowledge of model IDs, SDK bindings, beta headers, and features like prompt caching, tool use, streaming, and Managed Agents, which are easy to get wrong from memory alone. ## Core Features & Use Cases - Multi-language SDK guidance: Provides verified code patterns for Python, TypeScript, Java, Go, Ruby, C#, PHP, and raw cURL, with automatic language detection from project files. - Feature implementation: Covers tool use, streaming, structured outputs, prompt caching, compaction, adaptive thinking, batch processing, and the Files API with correct current model IDs. - Managed Agents support: Guides setup of server-managed stateful agents with sessions, environments, and workspaces, including an onboarding interview flow. - Use Case: A developer adding tool use to a TypeScript app gets correct @anthropic-ai/sdk code with the right model string, streaming helpers, and cache breakpoints instead of hallucinated API shapes. ## Quick Start Ask the assistant to add a Claude API call with tool use and prompt caching to your project file.

Frequently Asked Questions about claude-api

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

FAQPage Schema
How do I add tool use to a Claude API application?▼

Define tools with the SDK's tool runner (decorators, Zod schemas, or annotated classes) for automatic loop handling, or write a manual loop checking stop_reason for tool_use and returning tool_result blocks. Language-specific examples exist for Python, TypeScript, Java, Go, Ruby, and PHP.

Which Claude model should I use with the Anthropic SDK?▼

Default to claude-opus-4-6 with adaptive thinking (thinking: {type: "adaptive"}). Use the exact model ID strings without date suffixes, and only switch to Sonnet or Haiku when the user explicitly requests them.

Does the Anthropic SDK support streaming responses?▼

Yes, all official SDKs support streaming, and it is recommended for long outputs or high max_tokens to avoid HTTP timeouts. Use the SDK's get_final_message() or finalMessage() helper to accumulate the complete response.

Why is my prompt caching not hitting the cache?▼

Cache misses usually come from silent invalidators like timestamps in the system prompt, unsorted JSON, or changing tool lists, since caching requires an exact byte-stable prefix. Verify hits via usage.cache_read_input_tokens and keep volatile content after the last cache_control breakpoint.

Can I use Managed Agents on Amazon Bedrock or Google Vertex AI?▼

No, Managed Agents is first-party only and unavailable on Bedrock, Vertex AI, or Microsoft Foundry. On third-party providers, use the Claude API with tool use for all agent workloads.

When should I use raw HTTP instead of the Anthropic SDK?▼

Use raw HTTP only when the user explicitly asks for cURL/REST, the project is shell-based, or no official SDK exists for the language. Never mix raw HTTP with SDK calls in a language that has an official SDK.

Related Skills

Explore More Software Engineering Skills→
rhixecompanyrhixecompany

claude-api

Generate and debug Claude API and Anthropic SDK application code.

Community
Advanced
kitaekattkitaekatt
3

cache-report

Parse cache_report.py output to display prompt-cache hit rates and token consumption.

Community
Intermediate
AlexanderVilla47AlexanderVilla47

n8n-api-http-robusta

Configure n8n workflows with versioned, idempotent HTTP calls and jittered retries.

Community
Advanced
MeganHarrisonMeganHarrison

claude-opus-4-5-migration

Migrate prompts and code from older Claude models to Opus 4.5.

Community
Intermediate