Prompt Image Development Agentcore Agentcore Agentcore Agentcore Agentcore

Today, we are very happy to announce the agentcol agentcol model (agentcol for Amazon Bedrock Agentcore (MCP). of beedrock. You can use the AgentCore Prototyping server, to generate ai solutions, or measure your business agent infrastructure.

Agentic Agene such as Kiro, Claude Code, Gitub Copilot, and the indicator, and complex MCP servers change the way the engineers create agents Agents AI. What often takes important time and effort, for example learning the Bedrock Agentcore services, including Configuration tools, and sending Configuration Now will now be terminated with your chat authority.

In this post launches a new Agentcore MCP server and travels in installing measures to start.

The AgentCore MCP server skills

The AgentCore MCP server brings a new Agentic development experience, to provide specialized tools that automatically work for a comprehensive agent curve, remove the full learning curve, and reduce the demand for development that can reduce the new cycles. Coping with the challenges of agentcore MCP server development:

1. It converts agentcore Runtime By providing your jewup codes in a minimum processor of the required import library, updating, installing apps with BedrockAgentCoreApp()Converting Extoints on shoulder, and change the specific agent calls by loading download – while maintaining a logic agic and the Strands agents.
2. Automated the evolution of environmental development By managing the full set process of your code: installing the required dependency (Bedrock-agentCore SDK, configuring AWS agentCore regionals, and Building .Bedrock_agentcore configuration.
3. Facilitate the combination of tools with a beedrock agentcore gateway In the communication of agent-to-to -ot-to -ook area in the cloud.
4. Enables simple persuasion of agent and testing By providing environmental language instructions for your Code assistance to request services provided during the agentcore performance and ensure the complete transaction, including Calls at AgentCore Gateway's tools when applicable.

The method specified

When using an agentcore MCP server and your favorite client, we encourage you to consider legal formatting designed to provide comprehensive Age Agent support:

1. Layer 1: tall egedic or client – Use Kiro, Claude Code, CASE, VS extensions, or other interface interface. With very simple activities, Agentic-operating identities are equipped with relevant tools to view documents and perform specific tasks in the Bedrock agentcore. However, in this layer alone, enhancements can keep the highest performance in the Agentcore engineer.
2. Laser 2: AWS service – Enter the AWS DECESTION SERVICE MCP of complete AWS text, including context about the Bedrock AgentCore.
3. Layer 3: frame documents – Put strips, Langgraph, or other MCP's books draft servers or use the Special Policy LLMS.txt.
4. Background 4: SDK documents – Enter the MCP or use the Agent Framework SDK framework and the Bedrock AgentCore SDK integrated configuration of SDK agents and Bedrock AlentCore API.
5. BACKGRET 5: Directory files – Special guidelines for complex and repeated work performance. Each EIER has a unique way to use direction files (for example, see directing in the Kiro text).

Each layer creates the past, we have more than your Code assistant can handle everything from AWs working on the hard changes of agent and defagement.

Insertion

To get started with Amazon Bedrock Agentcore agentcore MCP Server You can use one to install in GitTub Repostory.

Each Ide meets a different MCP using the MCP.json file. Update MCP documents for your ipi, such as Kiro, cursor, IQ CLLI, CLAUDE code to find MCP.json location.

Use the following to your MCP.json:

{
  "mcpServers": {
    "awslabs.amazon-bedrock-agentcore-mcp-server": {
      "command": "uvx",
      "args": ["awslabs.amazon-bedrock-agentcore-mcp-server@latest"],
      "env": {
        "FASTMCP_LOG_LEVEL": "ERROR"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

For example, here is a Kiro-looking Kiro, by the server of agentcore MCP and two tools, to search_agentcore_docs and downloading

Using an agentcore MCP server server

While showing various cases of various cases of Kiro, the agentcore MCP server has been tested in Claude Code, Amazon Q CLLI, Q Q Q Q Q Q q. First, let's look at the general diagnostic development of the agentcore services (remember that this is only one example with the tools available, and to examine other such offenses by teaching your agent in Agentic):

Agent Development LifeCycle Follows these steps:

1. The user takes a set of toolbar or MCP servers and
   1. Creates a goal of the lambda gate of agentcore; or
   2. Recycle MCP server as-in-Agentcore Runtime
2. The user is preparing for the actual agent code using the framework you selected as the strands atens or Langgraph. The user can:
   1. Start from scratch (server can download documents from the strands or the Langgraph Docs)
   2. Start in a fully or part-validated agent code
3. The user asks an agency to change the code to the Competencore Runtime format with the purpose of sending a provider later. This is causing the agent to:
   1. Write the correct requirements of the requirements.txt
   2. Import the required libraries including beedrock_agentcore
   3. Decorating a large handle (or creating one) to access the agent's agent's agic agent or the installation manager
4. The user can ask an agent to send the running time to agentcore. The agent may view the documents and may use agentcore Cilder to enter the agent code for Ruretime
5. The user can test the agent by requesting an agent to do so. Agentcore CRI command is required in this regard and killed by the Customer
6. The user requests to modify the code to use the agentCore Gateway MCP server within the agentCore Runtime Agent server.
   1. The agent converts true code to add a MCP client to drive the entrance to
   2. Agent also uses new V2 version of the performance agent
   3. Agent and test this integration with a new TEMPT

Here is the Demo of the MCP server working with cussor Ide. We see the agent make the following steps:

1. Change weather_agent.py complying with running agentcore Runtime
2. Use AgentCore Cili to enter an agent
3. Check the agent used with a successful acceleration

Here is another example of using Langgraph agent in AgentCore Runtime with a cursor Identi-working in the same steps as seen above.

Clean

If you would like to download the MCP server, follow your ipple documents, such as Kiro, cursor, q cla, and CLOUTA code code.

Store

In this case, we showed how to use the agentcore MCP Server with your favorite development agentic to speed up your development flow.

We encourage you to update the gittib and study and use the following services in your development:

We encourage you to try the agentcore MCP Server and give any answer on the problems in our GitTub area.

About the authors