AWS re:Invent 2025現地の様子Workshop(Kiro CLI編)

> STEP 4: SHOW THE CORRECTED CONFIGURATION

The CORRECT configuration should be:

Configuration:

S3Bucket: !Ref CodeArtifactBucket

S3ObjectKey: 'sampleapp.zip'

The fix is to change S3Bucket: CodeArtifactBucket to S3Bucket: !Ref CodeArtifactBucket

The !Ref intrinsic function returns the actual physical name of the S3 bucket resource created by CloudFormation, which includes the stack name and a unique identifier.

③Would you like me to proceed with updating the CloudFormation stack with this fix?

▸ Credits: 0.14 • Time: 4s

[serverless-agent] > yes

> STEP 5: UPDATE CLOUDFORMATION STACK

I'll modify the following file: /tmp/aws-api-mcp/workdir/serverlesspipeline-template.yaml (using tool: write)

Purpose: Fix S3Bucket reference in template

341, 341: Configuration:

- 342 : S3Bucket: CodeArtifactBucket

+ 342: S3Bucket: !Ref CodeArtifactBucket

343, 343: S3ObjectKey: 'sampleapp.zip'

344, 344: PollForSourceChanges: false

Updating: /tmp/aws-api-mcp/workdir/serverlesspipeline-template.yaml

- Completed in 0.0s

Running tool call_aws with the param (from mcp server: awslabs.aws-api-mcp-server)

⋮ {

⋮ "cli_command": "aws cloudformation describe-stacks --stack-name serverlesspipeline --query 'Stacks[0].StackStatus'"

⋮ }

- Completed in 0.129s

>④ Excellent! The stack update completed successfully (UPDATE_COMPLETE). Now let me verify the pipeline Source stage status:

Running tool call_aws with the param (from mcp server: awslabs.aws-api-mcp-server)

⋮ {

⋮ "cli_command": "aws codepipeline get-pipeline-state --name SampleApplicationPipeline --query 'stageStates[0].[stageName,latestExecution.status]' --output table"

⋮ }

- Completed in 0.132s

> The pipeline still shows the old failed execution. The pipeline needs to be triggered again to use the updated configuration. Let me start a new execution:

Running tool call_aws with the param (from mcp server: awslabs.aws-api-mcp-server)

⋮ {

⋮ "cli_command": "aws codepipeline start-pipeline-execution --name SampleApplicationPipeline"

⋮ }

- Completed in 0.188s

Running tool call_aws with the param (from mcp server: awslabs.aws-api-mcp-server)

⋮ {

⋮ "cli_command": "aws codepipeline get-pipeline-state --name SampleApplicationPipeline --query 'stageStates[0].[stageName,latestExecution.status]' --output table"

⋮ }

- Completed in 0.161s

> ⑤Perfect! The fix is successful. The Source stage now shows "Succeeded" status.