Skills Architecture Design

The architecture design principles behind Agent Skills. This guide provides complete code examples and terminal operation demonstrations.

Quick Start

Launch Claude Code in the terminal:

claude

You: Help me implement the architecture design principles of Agent Skills

Claude: Sure, let me first look at the project structure...

Claude wants to:
  • Read file: src/main.py
Allow? [y/n/always/never]

You: y

Claude: Let me implement this for you...

Press Shift+Tab to switch permission modes:

# Normal Mode → Auto-Accept Mode → Plan Mode

# Auto-Accept: Auto-execute (bottom shows ⏵⏵ accept edits on)
You: Quick implementation
Claude: [Auto-executes without confirmation]

# Plan Mode: Read-only (bottom shows ⏸ plan mode on)
You: Create plan first
Claude: [Only creates plan, doesn't execute]

Basic Concepts

You: Explain the core concepts of Skills architecture design

Claude: The core concepts of Agent Skills architecture design include...

from anthropic import Anthropic

client = Anthropic()
response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Agent Skills architecture design principles"}]
)
print(response.content[0].text)

Implementation Steps

1. Environment Setup

# Install dependencies
pip install anthropic

# Configure API Key
export ANTHROPIC_API_KEY="your-api-key"

2. Core Implementation

You: Help me implement the core functionality of Skills architecture design

Claude: Sure, let me create the implementation file...

Claude wants to:
  • Write file: src/skills_architecture.py
Allow? [y/n/always/never]

You: y

Claude: File created.

# Skills Architecture Design core implementation
from anthropic import Anthropic

client = Anthropic()

def process(query: str) -> str:
    """Process user request"""
    response = client.messages.create(
        model="claude-sonnet-4-6",
        max_tokens=2048,
        messages=[{"role": "user", "content": query}]
    )
    return response.content[0].text

if __name__ == "__main__":
    result = process("Agent Skills architecture design principles")
    print(result)

3. Test Verification

You: Run tests to verify the implementation

Claude: Sure, let me run the tests...

Claude wants to:
  • Run command: python -m pytest tests/
Allow? [y/n/always/never]

You: y

Claude: All tests passed ✅

Advanced Usage

# Using streaming output
with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=2048,
    messages=[{"role": "user", "content": "Agent Skills architecture design principles"}]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

# Using tool calling
tools = [{
    "name": "process",
    "description": "Agent Skills architecture design principles",
    "input_schema": {
        "type": "object",
        "properties": {
            "query": {"type": "string"}
        },
        "required": ["query"]
    }
}]

response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    tools=tools,
    messages=[{"role": "user", "content": "Agent Skills architecture design principles"}]
)

Performance Optimization

# Use Prompt Caching to reduce costs
response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system=[{
        "type": "text",
        "text": "You are an expert in Agent Skills architecture design...",
        "cache_control": {"type": "ephemeral"}
    }],
    messages=[...]
)

# View costs
/cost

# Switch to faster model
/model haiku

# Use fast mode
/fast

Error Handling

from anthropic import APIError, APITimeoutError

try:
    response = client.messages.create(
        model="claude-sonnet-4-6",
        max_tokens=1024,
        messages=[{"role": "user", "content": "request"}]
    )
except APITimeoutError:
    print("Request timed out, please retry")
except APIError as e:
    print(f"API error: {e.status_code} - {e.message}")

Best Practices

1. Clear task descriptions - Tell Claude exactly what to do
2. Provide context - Use CLAUDE.md or directly provide relevant code
3. Verify step by step - Verify results after each step
4. Use Plan Mode - Create plans first for complex tasks
5. Monitor costs - Use /cost to view API usage
6. Choose models wisely - Use Haiku for simple tasks, Opus for complex tasks

Permission Mode Quick Reference

Mode	Shortcut	Bottom Display	Behavior
Normal	Default	No indicator	Each operation requires confirmation
Auto-Accept	Shift+Tab ×1	⏵⏵ accept edits on	Auto-execute edits
Plan	Shift+Tab ×2	⏸ plan mode on	Only create plans

Common Commands Quick Reference

/help          # View help
/clear         # Clear history
/model opus    # Switch to Opus 4.6
/model sonnet  # Switch to Sonnet 4.6
/model haiku   # Switch to Haiku 4.5
/cost          # View cost
/fast          # Fast mode
/quit          # Exit

Related Resources

• Claude Code Quick Start
• Basic Usage
• Cheat Sheet