fix(a2a): document the metadata-attach except-pass in a2a_executor (closes #1787)

GitHub Code Quality bot flagged the empty `except (AttributeError,
TypeError): pass` at workspace/a2a_executor.py:424 as a nit on PR #1783.
The suppression IS intentional — `new_agent_text_message()` returns
a plain string in MagicMock paths in tests where assignment to
`.metadata` raises despite hasattr being true.

This:
  - Adds a why-comment citing the test-mock motivation, commit
    dcbcf19 (the original guard), and issue #1787 so the next
    code-quality pass doesn't re-flag it.
  - Adds `logger.debug("metadata attach skipped (non-Message ...")`
    for observability — debug-level so production logs stay quiet
    but ops can flip the level if metadata loss is ever suspected.

Behavior unchanged. 43 existing a2a_executor tests still pass.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

workspace/a2a_executor.py

@@ -422,7 +422,14 @@ class LangGraphA2AExecutor(AgentExecutor):
                     try:
                         msg.metadata = {"tool_trace": tool_trace}
                     except (AttributeError, TypeError):
-                        pass
+                        # `new_agent_text_message()` returns a plain string in
+                        # MagicMock paths in tests, where assignment to
+                        # .metadata raises despite hasattr being true (the
+                        # mock has the attribute as a property). Suppression
+                        # is intentional — production Message objects always
+                        # accept the assignment. See #1787 + commit dcbcf19
+                        # for the original test-mock motivation.
+                        logger.debug("metadata attach skipped (non-Message return from new_agent_text_message)")
                 await event_queue.enqueue_event(msg)
                 _result = final_text