Fix sub_results iteration in the skills examples

sub_results is a dict keyed by agent name, so iterating it yields strings.
32_skills_multi_agent.py parallel died on the print — PARALLEL is the strategy
that actually populates it:

  AttributeError: 'str' object has no attribute 'status'

The other two call sites have the same bug and survived only because their
sub_results was empty. Iterate .items() and print the agent name with its
output.

Verified against a 3.32.0-rc18 server: all 8 modes across both examples now
complete, and parallel prints all three sub-agent outputs.

Author: ling-senpeng13

examples/agents/30_skills_dg_review.py

@@ -70,11 +70,12 @@ def run_standalone():
         print(f"Status: {result.status}")
         print(f"Tokens: {result.token_usage}")
 
-        # Sub-agent results are individually visible
+        # Sub-agent results are individually visible: a dict keyed by agent
+        # name — see AgentResult.sub_results.
         if result.sub_results:
-            print(f"\nSub-agent executions:")
-            for sub in result.sub_results:
-                print(f"  - {sub.agent_name}: {sub.status} ({sub.token_usage})")
+            print(f"\nSub-agent outputs:")
+            for agent_name, agent_output in result.sub_results.items():
+                print(f"  - {agent_name}: {str(agent_output)[:120]}")
 
 
 # ── Example 2: Compose with regular agent in pipeline ──────────────

examples/agents/32_skills_multi_agent.py

@@ -151,10 +151,11 @@ def example_pipeline():
         print(f"Execution ID: {result.execution_id}")
         print(f"Status:      {result.status}")
         print(f"Tokens:      {result.token_usage}")
+        # sub_results is a dict keyed by agent name — see AgentResult.sub_results.
         if result.sub_results:
-            print("\nSub-agent executions:")
-            for sub in result.sub_results:
-                print(f"  - {getattr(sub, "execution_id", "?")}: {sub.status}")
+            print("\nSub-agent outputs:")
+            for agent_name, agent_output in result.sub_results.items():
+                print(f"  - {agent_name}: {str(agent_output)[:120]}")
         result.print_result()
 
 
@@ -213,10 +214,11 @@ def example_parallel():
         print(f"Execution ID: {result.execution_id}")
         print(f"Status:      {result.status}")
         print(f"Tokens:      {result.token_usage}")
+        # sub_results is a dict keyed by agent name — see AgentResult.sub_results.
         if result.sub_results:
-            print("\nParallel sub-agent executions:")
-            for sub in result.sub_results:
-                print(f"  - {getattr(sub, "execution_id", "?")}: {sub.status}")
+            print("\nParallel sub-agent outputs:")
+            for agent_name, agent_output in result.sub_results.items():
+                print(f"  - {agent_name}: {str(agent_output)[:120]}")
         result.print_result()