Bedrock Agent

This template demonstrates how to handle Amazon Bedrock Agent events using function-based actions.

Architecture

The template sets up:

Amazon Bedrock Agent: Orchestrates user requests and identifies tool usage.
AWS Lambda function: Resolves function calls using BedrockAgentFunctionResolver.
Amazon DynamoDB table: Stores and retrieves items used by the agent.

Deploy the stack using:

make deploy STACK=agent

Function-based Actions: Uses the @app.tool() decorator to expose functions to Bedrock Agents.
Automatic Parameter Mapping: Maps Bedrock Agent parameters directly to Python function arguments.
Type Safety: Uses Pydantic models for data validation and camelCase alias conversion.
Observability: Integrated with AWS Lambda Powertools for logging, tracing, and metrics.

Use the @app.tool() decorator to define tools that the agent can call:

@app.tool(name="getItem", description="Gets item details by ID")
def get_item(item_id: str) -> dict:
    return handler.get_item(item_id)

Field	Type	Description
`id`	UUID string	Unique item identifier (auto-generated)
`name`	string	Human-readable item name
`description`	string	Human-readable item description

Variable	Description
`TABLE_NAME`	DynamoDB table name
`SERVICE_NAME`	Powertools service name
`METRICS_NAMESPACE`	Powertools metrics namespace