test(harness): exercise ToolResponseContent.is_error in parallel-tools fixture

declan-scale · claude · declan-scale · commit a04bf5eef7d9 · 2026-06-22T16:00:49.000-04:00
The field landed (AGX1-371), so the failing tool in the parallel-tools
fixture now sets is_error=True instead of relying on the "Error: ..."
string convention. The deriver threads it onto CloseSpan.is_error and the
cross-channel span-signal equivalence covers the error status.

Co-Authored-By: Claude Opus 4.8 (1M context) &lt;noreply@anthropic.com&gt;
diff --git a/tests/lib/core/harness/conformance/test_conformance.py b/tests/lib/core/harness/conformance/test_conformance.py
@@ -174,11 +174,10 @@
     # an error. It guards against the two channels agreeing with each other while
     # both mishandling interleaved/parallel spans or a failing tool.
     #
-    # The tool error is represented the way the harness encodes it today — an
-    # "Error: ..." string in ToolResponseContent.content (see
-    # claude_agents/hooks/hooks.py post_tool_use_failure_hook). Once the deferred
-    # ToolResponseContent.is_error field lands (AGX1-371), extend this fixture to
-    # assert the error status propagates onto the closed tool span.
+    # The failing tool sets ToolResponseContent.is_error=True (AGX1-371), which
+    # the span deriver threads onto the closed tool span's CloseSpan.is_error.
+    # Both channels feed the same deriver, so the recorded span signals — error
+    # status included — must match.
     Fixture(
         name="parallel-tools-with-error",
         events=[
@@ -217,6 +216,7 @@
                     tool_call_id="p-ls",
                     name="Bash",
                     content="Error: ls: /nope: No such file or directory",
+                    is_error=True,
                 ),
             ),
             # p-read succeeds and closes second.
