AgentSkillsCN

create-bedrock-agent

使用Terraform或Python Boto3创建并配置Amazon Bedrock Agent。当用户提出创建、部署或配置Bedrock Agent的需求，或在创建情境中提及“Bedrock Agent”时，可调用此技能。

SKILL.md
--- frontmatter
name: create-bedrock-agent
description: Create and configure Amazon Bedrock Agents using Terraform or Python Boto3. Use when the user asks to create, deploy, or setup a Bedrock Agent, or mentions "Bedrock Agent" in a creation context.

Create Amazon Bedrock Agent

Prerequisites

Before creating an agent, ensure the following IAM permissions are ready:

  1. •Agent Role: A service role allowing bedrock.amazonaws.com to assume it.
  2. •Model Access: The account must have access to the foundation model (e.g., Anthropic Claude).

Workflow

  1. •

    Determine Implementation Method:

    • •If the user uses Infrastructure as Code (IaC) -> Use Terraform patterns.
    • •If the user wants a script or programmatic creation -> Use Python (Boto3) patterns.
  2. •

    Select Components:

    • •Agent: The core orchestrator.
    • •Action Group: (Optional) For executing tasks via Lambda.
    • •Knowledge Base: (Optional) For RAG capabilities.

Terraform Implementation

See terraform.md for resource definitions.

Python (Boto3) Implementation

See python.md for SDK usage examples.

Validation

After creation, always advise the user to:

  1. •Prepare the agent (create a draft version).
  2. •Test in the AWS Console or via invoke_agent API using the DRAFT alias.