Try again

Signed-off-by: Martin Smithson <msmiths@uk.ibm.com>

Author: msmiths

docs/docs/section-2/step-04.md

@@ -6,8 +6,8 @@ In the previous step, you created **nested workflows** that combined sequential,
 build sophisticated multi-level orchestration.
 
 That design was intentionally concrete. Step 03 introduced separate feedback agents and a straightforward
-[`@ParallelAgent`](../../../section-2/step-03/src/main/java/com/carmanagement/agentic/workflows/FeedbackWorkflow.java) workflow
-so you could clearly see how multiple specialized agents collaborate inside a larger workflow.
+[`@ParallelAgent`](https://github.com/msmiths/langchain4j-workshop/blob/main/section-2/step-03/src/main/java/com/carmanagement/agentic/workflows/FeedbackWorkflow.java)
+workflow so you could clearly see how multiple specialized agents collaborate inside a larger workflow.
 
 In this step, the business problem evolves. We are still building on the same orchestration ideas, but the addition of a
 third feedback dimension creates a good opportunity to refactor the architecture. Rather than adding yet another nearly