PR #1111 fix only suppresses WouldBlock - may not address the actual hang root cause

[knoal]

Problem

PR #1111 (Fix query() stream hang when CLI subprocess dies mid-run) is overstated. The fix only suppresses WouldBlock on the error enqueue; if the consumer is parked on _message_send.receive() and the buffer is not full, the fix does not help.

Impact

• The fix may not address the actual hang root cause
• A real hang could still occur in the unhandled case
• The test does not exercise the fixed code path - mock_receive raises ProcessError directly, but the original hang was caused by await self._message_send.send(...) blocking on a full/closed buffer

What was audited

MCE audit Tcas1111 (2026-07-19) caught 3 HIGH attacks via F-C:

1. The test does not actually exercise the fixed code path
2. P1 (p=0.85 "stream unhangs when CLI dies") is overstated - the fix only suppresses WouldBlock on the error enqueue
3. The tests use pytest.raises(Exception, match=r"Command failed") - this matches any Exception

Recommended fix

1. The author should:
   a. Address the actual hang scenario - consumer parked on receive with full buffer
   b. Add a test that exercises the real code path
   c. Use more specific exception matching in tests

Cross-references

• MCE audit: Tcas1111 (2026-07-19), 0/4 HIT, Brier 0.55
• F-C analysis: /home/alex/.hermes/mce/fc-critiques/Tcas1111-with-diff.json
• PR: Fix query() stream hang when CLI subprocess dies mid-run #1111