Including automatically resolved stack outputs in LambdaInvokeAction for CDK Pipelines

[PeterBaker0]

When creating a CdkPipeline and adding stages, the ShellScriptAction stage has the helpful optional field use_outputs which allows a mapping of key words to StackOutput which are automatically resolved at the time of the ShellScriptAction running and are available as environment variables as we expect in the rest of the CDK Lambda workflow.

Upon trying to add a LambdaInvokeAction to the pipeline from the aws_codepipeline_actions package, I see that there is the option for user_parameters, inputs and outputs (artifacts) but no direct equivalent. It seems that user_parameters are statically defined at the time of synthesis for the pipeline stack and when read inside the Lambda function are not resolved to their values.

It seems that there is some guidance here about how to manually pull the zip files from the relevant artefact s3 bucket, however this really breaks the workflow that CDK usually provides.

Is there an equivalent, or something closer, to the kindness of just being able to throw in StackOutputs from the pipeline.stack_output function directly into environment variables for a LambdaInvokeAction?

If not, I think this could be a valuable feature.