From 1a365b976b64b48553eae262a7d414c93a69d977 Mon Sep 17 00:00:00 2001
From: Claude <noreply@anthropic.com>
Date: Tue, 1 Sep 2026 11:43:47 +0000
Subject: [PATCH] fix(test): stop asserting openai free-pool admission after
 #1587

Merged #1587 ("separate free-pool admission from global discovery")
intentionally excluded OPENAI_API_KEY from FREE_POOL_CREDENTIAL_NAMES,
but left two of its own tests unmigrated:
test_build_catalog_applies_account_cap and
test_build_catalog_respects_limit both built discovery reports with
openai rows and asserted they were admitted to the free-pool catalog.
Every full-suite/coverage-evidence run on protected main -- and every
PR rebasing onto it -- inherited these two failures regardless of its
own diff.

Reproduced genuine RED against unmodified main (960b0845): both tests
fail, one with PolicyError ("no free model route is available"), the
other with a KeyError on the now-absent "openai" account count.

Swapped the openai rows in both tests for bytez, which is also
is_free-eligible but, unlike openai, remains in
FREE_POOL_CREDENTIAL_NAMES. This preserves each test's original
intent -- three distinct provider accounts each capped at 2, and a
single provider's rows truncated to the configured limit -- without
depending on the now-removed OpenAI free-pool admission.

Verified: full suite 2273 passed, 1 skipped, 21 subtests; 100%
coverage and docstrings maintained. No production code changed.
---
 CHANGELOG.md                                        | 11 +++++++++++
 tests/test_contextual_orchestrator_review_policy.py |  6 +++---
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 8e8633515b..b842e16927 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,17 @@ this file. The format follows Keep a Changelog, and versioned releases follow
 Semantic Versioning where the repository publishes a release.
 
 ## [Unreleased]
+- Fix two `tests/test_contextual_orchestrator_review_policy.py` tests left broken by merged
+  `#1587` ("separate free-pool admission from global discovery"), which intentionally excluded
+  `OPENAI_API_KEY` from `FREE_POOL_CREDENTIAL_NAMES` but did not update
+  `test_build_catalog_applies_account_cap` and `test_build_catalog_respects_limit`, both of which
+  still built discovery reports using `openai` rows and asserted they were admitted to the free
+  pool. Every full-suite/coverage-evidence run on protected `main` (and every PR rebasing onto it)
+  inherited these two failures regardless of its own diff. Swapped the `openai` rows in both tests
+  for `bytez` (also `is_free`-eligible but, unlike `openai`, still in `FREE_POOL_CREDENTIAL_NAMES`),
+  preserving each test's original intent — three distinct provider accounts each capped at 2, and a
+  single provider's rows truncated to the configured limit — without depending on the now-removed
+  OpenAI free-pool admission. No production code changed.
 - **Fix a live crash: `noema-review` failed with an unhandled `HTTPError` instead
   of failing closed.** Live incident on `ContextualWisdomLab/naruon#1486`:
   `scripts/ci/noema_review_gate.py::call_llm`'s `opener.open(request)` call sat
diff --git a/tests/test_contextual_orchestrator_review_policy.py b/tests/test_contextual_orchestrator_review_policy.py
index b10fc4a0b9..089f73a574 100644
--- a/tests/test_contextual_orchestrator_review_policy.py
+++ b/tests/test_contextual_orchestrator_review_policy.py
@@ -354,7 +354,7 @@ def test_build_catalog_applies_account_cap() -> None:
             for i in range(6)
         ]
         + [
-                {"provider": "openai", "model": f"o{i}", "agent_id": f"oa_{i}", "is_free": True, **FREE_PRICE}
+                {"provider": "bytez", "model": f"b{i}", "agent_id": f"bz_{i}", "is_free": True, **FREE_PRICE}
             for i in range(3)
         ]
     }
@@ -367,14 +367,14 @@ def test_build_catalog_applies_account_cap() -> None:
         account_counts[account] = account_counts.get(account, 0) + 1
     assert account_counts["nvidia_nim"] == 2
     assert account_counts["nvidia_nim_sub"] == 2
-    assert account_counts["openai"] == 2
+    assert account_counts["bytez"] == 2
 
 
 def test_build_catalog_respects_limit() -> None:
     """The catalog never exceeds the configured agent limit."""
     report = {
         "models": [
-                {"provider": "openai", "model": f"m{i}", "agent_id": f"oa_{i}", "is_free": True, **FREE_PRICE}
+                {"provider": "bytez", "model": f"m{i}", "agent_id": f"bz_{i}", "is_free": True, **FREE_PRICE}
             for i in range(20)
         ]
     }