Verify Connection¶

Confirm that your agent can reach the AWS MCP Server and retrieve information from AWS services.

Connectivity Test¶

Ask your agent the following prompt:

Search AWS documentation for Amazon S3 bucket naming rules

Your agent should return content from AWS documentation describing S3 bucket naming conventions (lowercase letters, numbers, hyphens, 3–63 characters, etc.).

Info

If your agent returns S3 naming rules from AWS documentation, your MCP Server connection is working correctly. You are ready to start the workshop!

Expected Result¶

A successful response means:

Your agent connected to the MCP Server endpoint
AWS credentials were accepted and authenticated
The documentation search tool executed and returned results
Your agent formatted and displayed the documentation content

You should see information about S3 bucket naming rules including character restrictions, length limits, and formatting requirements.

Troubleshooting¶

If your agent cannot retrieve documentation or returns an error, work through this checklist:

Troubleshooting Checklist

1. Credential Expiry¶

Workshop Studio credentials are temporary. If more than a few hours have passed since you set them up:

Return to the Workshop Studio dashboard
Click Get AWS CLI credentials
Re-export the new values in your terminal
Restart your agent from the same terminal

2. Network or Firewall¶

Verify your machine can reach the MCP endpoint:

curl -s -o /dev/null -w "%{http_code}" https://mcp.us-east-1.amazonaws.com/

If you do not get a response (or get a timeout), check: - Wi-Fi / network connectivity - VPN or corporate proxy settings that may block outbound HTTPS - Firewall rules allowing traffic to mcp.us-east-1.amazonaws.com on port 443

3. MCP Endpoint URL¶

Confirm the endpoint URL is exactly:

https://mcp.us-east-1.amazonaws.com/

Common mistakes: - Missing trailing slash - Wrong region (must be us-east-1) - HTTP instead of HTTPS - Extra path segments

4. Plugin or Configuration Syntax¶

::::tabs :::tab{label="Claude Code"} Verify the plugin is installed:

/plugin list

You should see aws-core@agent-toolkit-for-aws in the output. If not, reinstall:

/plugin marketplace add aws/agent-toolkit-for-aws
/plugin install aws-core@agent-toolkit-for-aws

KiroCursorCodex

Verify your MCP config file exists and has correct JSON:

cat ~/.kiro/settings/mcp.json

Expected content:

{
  "mcpServers": {
    "aws": {
      "url": "https://mcp.us-east-1.amazonaws.com/",
      "transport": "sse"
    }
  }
}

Restart Kiro after any config changes.

Verify the MCP config exists in your project:

cat .cursor/mcp.json

Expected content:

{
  "mcpServers": {
    "aws": {
      "url": "https://mcp.us-east-1.amazonaws.com/",
      "transport": "sse"
    }
  }
}

Restart Cursor after any config changes.

Verify the plugin is installed:

codex plugin list

You should see aws/agent-toolkit-for-aws in the output. If not, reinstall:

codex plugin marketplace add aws/agent-toolkit-for-aws

5. Environment Variables¶

Confirm credentials are set in your current shell:

echo $AWS_ACCESS_KEY_ID
echo $AWS_SECRET_ACCESS_KEY
echo $AWS_SESSION_TOKEN
echo $AWS_DEFAULT_REGION

All four should return values. If any are empty, re-export them from the Workshop Studio credentials page.

:::

Ready to Go¶

Once your agent successfully returns AWS documentation content, your setup is complete. You can now proceed to any module:

Module 1: Knowledge Tools — explore AWS documentation search
Module 2: Agent Skills — build a serverless API
Module 3: API Tools — configure S3 and security groups
Module 4: Guardrails & Security — set agent boundaries
Module 5: Troubleshooting — diagnose a broken stack
Module 6: Capstone — build an image processing pipeline

Info

Modules 1–5 are independent — complete them in any order. Module 6 (Capstone) benefits from familiarity with at least one earlier module.