New Workflow Submission: sfn-agentcore-bedrock-cdk

## New Workflow Submission: Step Functions with Bedrock AgentCore Multi-Agent Orchestration

**PR:** #413

### Pattern Overview

This workflow uses AWS Step Functions to orchestrate multiple Amazon Bedrock AgentCore agents in parallel via a Map state, then aggregates their responses into a unified summary using the Bedrock Converse API.

### example-workflow.json Details

| Field | Value |
|-------|-------|
| **Title** | Step Functions with Bedrock AgentCore Multi-Agent Orchestration |
| **Language** | TypeScript |
| **Framework** | CDK |
| **Type** | Standard |
| **Level** | 400 |
| **Simplicity** | 3 - Application |
| **Services** | bedrock, lambda, stepfunctions |

### Description (introBox.text)

This workflow deploys a Step Functions state machine that invokes multiple Bedrock AgentCore agents in parallel using a Map state, then aggregates their responses into a single coherent summary via Amazon Bedrock Converse.

The workflow: (1) A trigger Lambda starts the state machine with a prompt and list of agent runtime ARNs, (2) the Map state fans out to invoke each AgentCore agent in parallel via a bundled Lambda intermediary, (3) an aggregate Lambda uses Bedrock Converse to synthesize all agent responses into a unified summary.

Key architectural decisions: The @aws-sdk/client-bedrock-agentcore SDK is not available in the Lambda runtime and must be bundled using NodejsFunction. InvokeAgentRuntime returns a streaming response, requiring a Lambda intermediary rather than direct Step Functions SDK integration.

### Diagram

The state machine diagram is included at `resources/statemachine.png` showing:
- Start → InvokeAgents (Map State) → CallAgent (Lambda) → AggregateResults (Lambda) → End

### Resources

1. [Amazon Bedrock AgentCore Documentation](https://docs.aws.amazon.com/bedrock/latest/userguide/agentcore.html)
2. [AWS Step Functions Map State](https://docs.aws.amazon.com/step-functions/latest/dg/amazon-states-language-map-state.html)
3. [Amazon Bedrock Converse API](https://docs.aws.amazon.com/bedrock/latest/userguide/conversation-inference-call.html)
4. [The AWS Step Functions Workshop](https://catalog.workshops.aws/stepfunctions/en-US)

### Author

**Nithin Chandran R** — Technical Account Manager at AWS
[LinkedIn](https://www.linkedin.com/in/nithin-chandran-r/)

### Checklist

- [x] Used the [model template](https://github.com/aws-samples/step-functions-workflows-collection/tree/main/_workflow_model)
- [x] `example-workflow.json` includes all required fields
- [x] Description is 300-500 words
- [x] Diagram exported as PNG from Workflow Studio
- [x] README.md sections completed in full
- [x] Framework: CDK
- [x] Author bio with LinkedIn reference

Review PR #413 together with the workflow model template, example-workflow.json, README.md, and resources/statemachine.png. Check that the submission follows the listed metadata, description, diagram, resource links, and author-bio requirements, and that the workflow’s Step Functions, Lambda, and Bedrock components are represented consistently. Done means the checklist is satisfied and the workflow is ready for inclusion.

Written by the indexing model from the issue text.