anthropic-expert

Guide Claude Code CLI features including plugins, hooks, and MCP integration.

Updated Nov 25, 2025
One-click install
npx skills add https://github.com/gounthar/bazel-riscv --skill anthropic-expert-gounthar
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: anthropic-expert
Source: https://github.com/gounthar/bazel-riscv/tree/main/.claude/skills/anthropic
Command: npx skills add https://github.com/gounthar/bazel-riscv --skill anthropic-expert-gounthar

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Integrating with Anthropic's Claude API, mastering prompt engineering, and leveraging advanced features like function calling and vision can be complex and require deep technical understanding. This Skill provides comprehensive, up-to-date guidance, saving developers time and effort in building powerful, intelligent AI applications.

Core Features & Use Cases

  • Claude API Mastery: Gain expert knowledge of the Messages API, optimal model selection (Opus, Sonnet, Haiku), and efficient streaming responses for your applications.
  • Advanced Prompt Engineering: Get guidance on crafting effective system prompts, using XML tags for structured output, and few-shot examples to improve AI performance.
  • Function Calling & Vision: Learn to implement tool use for external actions and analyze images with Claude's vision capabilities, expanding your AI's interaction possibilities.
  • Use Case: If you need to integrate Claude's vision capabilities into your application to analyze images, this Skill can show you how to structure the API call to send an image and ask a question about its content, quickly and accurately.

Quick Start

Explain the key differences between Claude 3.5 Opus, Sonnet, and Haiku models, and when to use each.

Frequently Asked Questions about anthropic-expert

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

FAQPage Schema
How do I use Claude's vision capabilities to analyze images in my application?

Claude's vision API accepts images as base64-encoded data or URLs in your Messages API calls. Structure your request with image content blocks, specify the model (Opus, Sonnet, or Haiku), and ask questions about the image content. The API returns analyzed results directly in the response.

What are the key differences between Claude 3.5 Opus, Sonnet, and Haiku models?

Opus excels at complex reasoning and tasks requiring deep analysis. Sonnet balances intelligence and speed for most applications. Haiku prioritizes cost-efficiency and latency for simpler tasks. Choose based on task complexity, budget, and latency requirements.

How do I implement function calling with the Claude API?

Define tools in your API request with name, description, and input schema. Claude processes context and returns tool_use blocks indicating which functions to call. Parse these responses, execute the functions externally, and send results back via the API for continued reasoning.

Can I use XML tags to structure Claude's outputs for better formatting?

Yes. Include XML tags in your system prompt to define output structure—Claude learns these patterns and follows them consistently. This technique improves parsing reliability and helps standardize responses across different prompts and use cases.

What's the best way to optimize prompt engineering for better Claude API performance?

Combine clear system prompts defining Claude's role, use few-shot examples showing desired behavior, add XML tags for structured output, and iterate on results. Start with Sonnet for prototyping, then validate with your target model before production deployment.

Does streaming work with Claude's Messages API for real-time responses?

Yes. Enable streaming in API requests to receive tokens incrementally, reducing perceived latency and enabling real-time output display. Streaming works with all Claude models and integrates seamlessly into web applications and command-line tools.