How do I use Claude with AWS Bedrock or Google Vertex AI?

Set the 'provider' field in the SDK configuration (e.g., provider: 'bedrock') and ensure the required environment variables for that cloud provider are exported.

Can I control which tools an agent can access?

Yes, you can use the 'allowedTools' or 'disallowedTools' arrays in the configuration to explicitly manage permissions for tools like bash, grep, or file editing.

Where should I store agent-specific memory?

Standard practice is to use a .claude/ directory at the project root, storing project context in CLAUDE.md and subagent definitions in the agents/ subdirectory.

How do I initialize a basic Claude agent?

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

What is the best way to manage API keys?

Never hardcode keys. Use environment variables like ANTHROPIC_API_KEY and access them via process.env to maintain security and portability.

Claude Agent SDK Agent Creation

Name: Claude Agent SDK Agent Creation
Author: TheBushidoCollective

byTheBushidoCollective

•

API Development

Simplifies the initialization and configuration of Claude AI agents using the official TypeScript SDK.

About

This skill provides a comprehensive framework and standardized patterns for building specialized AI agents with the Claude Agent SDK. It guides developers through the entire setup process, including selecting the latest models, configuring complex system prompts, and managing tool permissions to ensure secure operation. By establishing a clear directory structure for project memory and subagent definitions, it enables the creation of sophisticated, context-aware agents that integrate seamlessly with Anthropic, AWS Bedrock, and Google Vertex AI.

Key Features

Multi-provider authentication (Anthropic, AWS, Google Cloud, Azure)
74 GitHub stars
Standardized TypeScript SDK initialization patterns
Project-level memory management via .claude directory structure
Fine-grained tool permission and restriction controls
Best practice implementation for model selection and security

Use Cases

Configuring secure tool-calling environments with restricted bash or filesystem access
Integrating AI agents into existing TypeScript projects with standardized memory sources
Bootstrapping specialized subagents for complex software development tasks

About

Key Features

Multi-provider authentication (Anthropic, AWS, Google Cloud, Azure)
74 GitHub stars
Standardized TypeScript SDK initialization patterns
Project-level memory management via .claude directory structure
Fine-grained tool permission and restriction controls
Best practice implementation for model selection and security

Use Cases

Configuring secure tool-calling environments with restricted bash or filesystem access
Integrating AI agents into existing TypeScript projects with standardized memory sources
Bootstrapping specialized subagents for complex software development tasks