What is the recommended way to handle API keys?

Always use environment variables such as ANTHROPIC_API_KEY or AWS_ACCESS_KEY_ID instead of hardcoding credentials directly in your source code.

Does this skill support AWS Bedrock or Google Vertex AI?

Yes, it provides configuration patterns for multiple cloud providers including Amazon Bedrock, Google Vertex AI, and Azure OpenAI.

How do I initialize a basic Claude agent?

Use the Agent constructor from the @anthropic-ai/claude-agent-sdk, specifying the model, system prompt, and allowed tools in the configuration object.

Can I restrict what tools an agent can use?

Yes, you can use the 'allowedTools' array to whitelist specific functions or 'disallowedTools' to blacklist them, ensuring the agent operates within a secure boundary.

How does the SDK handle project-specific context?

By setting 'settingSources' to ['project'], the agent automatically loads context and instructions from the .claude/CLAUDE.md file in your project root.

Claude Agent SDK Creation

Name: Claude Agent SDK Creation
Author: TheBushidoCollective

byTheBushidoCollective

•

Data Science & ML

Streamlines the initialization and configuration of autonomous AI agents using the Claude Agent SDK.

This skill provides developers with standardized patterns and best practices for building sophisticated AI agents using the Claude Agent SDK in TypeScript. It covers essential setup tasks including model selection, multi-provider authentication (Anthropic, AWS Bedrock, Google Vertex AI), and complex permission management. By implementing structured project memory via CLAUDE.md and defining clear subagent hierarchies, this skill ensures that agents are secure, context-aware, and aligned with professional software development standards.

Key Features

01Integrated project and user-level memory management

02Standardized TypeScript initialization for Claude agents

0397 GitHub stars

04Best practices for secure API key handling and environment setup

05Granular tool permission management with strict and allow-list modes

06Multi-cloud provider configuration (AWS, Google, Azure, Anthropic)

Use Cases

01Building multi-agent systems with hierarchical subagent definitions

02Developing specialized code review agents with restricted file system access

03Creating context-aware CLI tools that leverage local project documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thebushidocollective/han agent-creation

For use in Claude.ai and ChatGPT

Download Skill