Python: fix: concurrent_agents sample incorrectly treats output as list[Message]

[Benke Qu (benke520)]

Problem

The concurrent_agents.py sample treats get_outputs() results as list[Message] directly:

for output in outputs:
    messages: list[Message] | Any = output
    for i, msg in enumerate(messages, start=1):

However, at runtime the default ConcurrentBuilder aggregator yields AgentResponse objects, not list[Message]. This causes incorrect iteration behavior.

Verification

outputs = events.get_outputs()
for output in outputs:
    print(type(output).__name__)  # prints "AgentResponse", not "list"

Fix

Check isinstance(output, AgentResponse) and iterate output.messages instead of treating the output as a raw list.

Testing

Ran the sample locally with agent-framework v1.8.0 and confirmed the output is correctly printed with the fix.

[Copilot]

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

Updates the concurrent agents orchestration sample to treat each orchestration result as a structured AgentResponse instead of a loosely typed message list.

Changes:

• Import AgentResponse and remove the previous Any-based casting.
• Add a runtime type check for AgentResponse before printing results.
• Iterate over output.messages rather than assuming output is list[Message].

[Eduard van Valkenburg (eavanvalkenburg)]

Benke Qu (@benke520) there are merge conflicts and open comments, please have a look!

[Benke Qu (benke520)]

Benke Qu (@benke520) there are merge conflicts and open comments, please have a look!

Thanks Eduard van Valkenburg (@eavanvalkenburg) — both addressed! Merge conflicts resolved (rebased onto latest main) and replied to the Copilot review comment. Also added a warning print for unexpected output types to avoid silent skipping.

[Benke Qu (benke520)]

Benke Qu (@benke520) there are merge conflicts and open comments, please have a look!

Thanks Eduard van Valkenburg (@eavanvalkenburg) — both addressed! Merge conflicts resolved (rebased onto latest main) and replied to the Copilot review comment. Also added a warning print for unexpected output types to avoid silent skipping.