Cairn
NousResearch/hermes-agent

OpenAI API parameter compatibility bugs and fixes

Cluster 60355 · agent. Data as of

Nothing open on GitHub. Fix merged in #65029 on 15 Jul 2026. Check it covers your case.

The short answer

0open issues
0open PRs
3items grouped
Known?
Yes. 3 related items grouped here, 1 issue already closed.
Fix event
A fix was merged to main in PR #65029. Check it covers your case. No release data is tracked.
Tracker
The maintainers' tracker marks this unresolved. That is separate from the fix event.
Still open
Nothing open on GitHub.

These issues and fixes concern incorrect API parameters being sent to OpenAI and compatible models, specifically around max_tokens vs max_completion_tokens selection, unsupported reasoning/include parameters, and model capability detection logic that breaks on custom or non-standard endpoints.

Open issues0

No open issues.

Open PRs0

No open pull requests.

Fixes on record

Merged to main. A stamp counts the issues a fix is linked to.

  • #65029 fix(auxiliary): route direct-create aux callers through call_llm

    merged to main

    files: hermes_cli/config.py, hermes_cli/goals.py, hermes_cli/kanban_decompose.py, hermes_cli/kanban_specify.py, hermes_cli/profile_describer.py, tests/hermes_cli/test_goals.py, tests/hermes_cli/test_kanban_cli.py, tests/hermes_cli/test_kanban_decompose.py, tests/hermes_cli/test_kanban_specify.py, tests/hermes_cli/test_profile_describer.py, tests/plugins/test_kanban_dashboard_plugin.py

    Show the diff
    • hermes_cli/config.py
      @@ -1700,6 +1700,18 @@ def _ensure_hermes_home_managed(home: Path):
             "extra_body": {},
             "reasoning_effort": "",  # per-task thinking level: none|minimal|low|medium|high|xhigh|max|ultra (empty = provider default)
         },
+        # Goal judge — evaluates whether a /goal run's latest response
+        # satisfies the goal/contract, and drafts goal contracts. Short
+        # structured-JSON calls; a fast cheap model is fine.
+        "goal_judge": {
+            "provider": "auto",
+            "model": "",
+            "base_url": "",
+            "api_key": "",
+            "timeout": 60,
+            "extra_body": {},
+            "reasoning_effort": "",  # per-task thinking level: none|minimal|low|medium|high|xhigh|max|ultra (empty = provider default)
+        },
         # Curator — skill-usage review fork. Timeout is generous because the
         # review pass can take several minutes on reasoning models (umbrella
         # building over hundreds of candidate skills). "auto" = use main chat
    • hermes_cli/goals.py
      @@ -878,20 +878,11 @@ def judge_goal(
         return "continue", "empty response (nothing to evaluate)", False, None
 
     try:
-        from agent.auxiliary_client import get_auxiliary_extra_body, get_text_auxiliary_client
+        from agent.auxiliary_client import call_llm
     except Exception as exc:
         logger.debug("goal judge: auxiliary client import failed: %s", exc)
         return "continue", "auxiliary client unavailable", False, None
 
-    try:
-        client, model = get_text_auxiliary_client("goal_judge")
-    except Exception as exc:
-        logger.debug("goal judge: get_text_auxiliary_client failed: %s", exc)
-        return "continue", "auxiliary client unavailable", False, None
-
-    if client is None or not model:
-        return "continue", "no auxiliary client configured", False, None
-
     # Build the prompt. Priority: contract > subgoals > plain. When both a
     # contract and subgoals exist, the subgoals are appended into the
     # contract block as extra criteria so the judge sees a single source of
@@ -935,8 +926,11 @@ def judge_goal(
         )
 
     try:
-        resp = client.chat.completions.create(
-            model=model,
+        # Route through call_llm so auxiliary.goal_judge.* config
+        # (provider/model/base_url, extra_body, reasoning_effort, retries)
+        # all apply — the direct-create path dropped extra_body (#35566).
+        resp = call_llm(
+            task="goal_judge",
             messages=[
                 {"role": "system", "content": JUDGE_SYSTEM_PROMPT},
                 {"role": "user", "content": prompt},
@@ -944,7 +938,6 @@ def judge_goal(
             temperature=0,
             max_tokens=_goal_judge_max_tokens(),
             timeout=timeout,
-            extra_body=get_auxiliary_extra_body() or None,
         )
     except Exception as exc:
         logger.info("goal judge: API call failed (%s) — falling through to continue", exc)
@@ -999,23 +992,15 @@ def draft_contract(objective: str, *, timeout: float = DEFAULT_JUDGE_TIMEOUT) ->
         return None
 
     try:
-        from agent.auxiliary_client import get_auxiliary_extra_body, get_text_auxiliary_client
+        from agent.auxiliary_client import call_llm
     except Exception as exc:
         logger.debug("goal draft: auxiliary client import failed: %s", exc)
         return None
 
     try:
-        client, model = get_text_auxiliary_client("goal_judge")
-    except Exception as exc:
-        logger.debug("goal draft: get_text_auxiliary_client failed: %s", exc)
-        return None
-
-    if client is None or not model:
-        return None
-
-    try:
-        resp = client.chat.completions.create(
-            model=model,
+        # Route through call_llm — same #35566 fix as the judge call above.
+        resp = call_llm(
+            task="goal_judge",
             messages=[
                 {"role": "system", "content": DRAFT_CONTRACT_SYSTEM_PROMPT},
                 {"role": "user", "content": f"Objective:\n{_truncate(objective, 4000)}"},
@@ -1023,7 +1008,6 @@ def draft_contract(objective: str, *, timeout: float = DEFAULT_JUDGE_TIMEOUT) ->
             temperature=0,
             max_tokens=_goal_judge_max_tokens(),
             timeout=timeout,
-            extra_body=get_auxiliary_extra_body() or None,
         )
     except Exception as exc:
         logger.info("goal draft: API call failed (%s)", exc)
    • hermes_cli/kanban_decompose.py
      @@ -298,23 +298,11 @@ def decompose_task(
     roster, valid_names = _build_roster()
 
     try:
-        from agent.auxiliary_client import (  # type: ignore
-            get_auxiliary_extra_body,
-            get_text_auxiliary_client,
-        )
+        from agent.auxiliary_client import call_llm  # type: ignore
     except Exception as exc:
         logger.debug("decompose: auxiliary client import failed: %s", exc)
         return DecomposeOutcome(task_id, False, "auxiliary client unavailable")
 
-    try:
-        client, model = get_text_auxiliary_client("kanban_decomposer")
-    except Exception as exc:
-        logger.debug("decompose: get_text_auxiliary_client failed: %s", exc)
-        return DecomposeOutcome(task_id, False, "auxiliary client unavailable")
-
-    if client is None or not model:
-        return DecomposeOutcome(task_id, False, "no auxiliary client configured")
-
     user_msg = _USER_TEMPLATE.format(
         task_id=task.id,
         title=_truncate(task.title or "", 400),
@@ -324,8 +312,12 @@ def decompose_task(
     )
 
     try:
-        resp = client.chat.completions.create(
-            model=model,
+        # Route through call_llm so auxiliary.kanban_decomposer.* config
+        # (provider/model/base_url, extra_body, reasoning_effort, retries)
+        # all apply — the previous direct client.chat.completions.create()
+        # path dropped auxiliary.<task>.extra_body entirely (#35566).
+        resp = call_llm(
+            task="kanban_decomposer",
             messages=[
                 {"role": "system", "content": _SYSTEM_PROMPT},
                 {"role": "user", "content": user_msg},
@@ -333,7 +325,6 @@ def decompose_task(
             temperature=0.3,
             max_tokens=4000,
             timeout=timeout or 180,
-            extra_body=get_auxiliary_extra_body() or None,
         )
     except Exception as exc:
         logger.info(
    • hermes_cli/kanban_specify.py
      @@ -162,22 +162,11 @@ def specify_task(
         )
 
     try:
-        from agent.auxiliary_client import get_auxiliary_extra_body, get_text_auxiliary_client
+        from agent.auxiliary_client import call_llm
     except Exception as exc:  # pragma: no cover — import smoke test
         logger.debug("specify: auxiliary client import failed: %s", exc)
         return SpecifyOutcome(task_id, False, "auxiliary client unavailable")
 
-    try:
-        client, model = get_text_auxiliary_client("triage_specifier")
-    except Exception as exc:
-        logger.debug("specify: get_text_auxiliary_client failed: %s", exc)
-        return SpecifyOutcome(task_id, False, "auxiliary client unavailable")
-
-    if client is None or not model:
-        return SpecifyOutcome(
-            task_id, False, "no auxiliary client configured"
-        )
-
     user_msg = _USER_TEMPLATE.format(
         task_id=task.id,
         title=_truncate(task.title or "", 400),
@@ -185,8 +174,11 @@ def specify_task(
     )
 
     try:
-        resp = client.chat.completions.create(
-            model=model,
+        # Route through call_llm so auxiliary.triage_specifier.* config
+        # (provider/model/base_url, extra_body, reasoning_effort, retries)
+        # all apply — the direct-create path dropped extra_body (#35566).
+        resp = call_llm(
+            task="triage_specifier",
             messages=[
                 {"role": "system", "content": _SYSTEM_PROMPT},
                 {"role": "user", "content": user_msg},
@@ -194,7 +186,6 @@ def specify_task(
             temperature=0.3,
             max_tokens=[redacted:secret-assignment]
             timeout=timeout or 120,
-            extra_body=get_auxiliary_extra_body() or None,
         )
     except Exception as exc:
         logger.info(
    • hermes_cli/profile_describer.py
      @@ -210,23 +210,11 @@ def describe_profile(
         model, provider = None, None
 
     try:
-        from agent.auxiliary_client import (  # type: ignore
-            get_auxiliary_extra_body,
-            get_text_auxiliary_client,
-        )
+        from agent.auxiliary_client import call_llm  # type: ignore
     except Exception as exc:
         logger.debug("describe: auxiliary client import failed: %s", exc)
         return DescribeOutcome(canon, False, "auxiliary client unavailable")
 
-    try:
-        client, aux_model = get_text_auxiliary_client("profile_describer")
-    except Exception as exc:
-        logger.debug("describe: get_text_auxiliary_client failed: %s", exc)
-        return DescribeOutcome(canon, False, "auxiliary client unavailable")
-
-    if client is None or not aux_model:
-        return DescribeOutcome(canon, False, "no auxiliary client configured")
-
     user_msg = _USER_TEMPLATE.format(
         name=canon,
         model=(model or "(unset)"),
@@ -237,8 +225,11 @@ def describe_profile(
     )
 
     try:
-        resp = client.chat.completions.create(
-            model=aux_model,
+        # Route through call_llm so auxiliary.profile_describer.* config
+        # (provider/model/base_url, extra_body, reasoning_effort, retries)
+        # all apply — the direct-create path dropped extra_body (#35566).
+        resp = call_llm(
+            task="profile_describer",
             messages=[
                 {"role": "system", "content": _SYSTEM_PROMPT},
                 {"role": "user", "content": user_msg},
@@ -246,7 +237,6 @@ def describe_profile(
             temperature=0.3,
             max_tokens=400,
             timeout=timeout or 60,
-            extra_body=get_auxiliary_extra_body() or None,
         )
     except Exception as exc:
         logger.info("describe: API call failed for %s (%s)", canon, exc)
    • tests/hermes_cli/test_goals.py
      @@ -166,8 +166,8 @@ def test_no_aux_client_continues(self):
         from hermes_cli import goals
 
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(None, None),
+            "agent.auxiliary_client.call_llm",
+            side_effect=RuntimeError("No LLM provider configured"),
         ):
             verdict, _, _, _wd = goals.judge_goal("my goal", "my response")
         assert verdict == "continue"
@@ -176,11 +176,9 @@ def test_api_error_continues(self):
         """Judge exception → fail-open continue (don't wedge progress on judge bugs)."""
         from hermes_cli import goals
 
-        fake_client = MagicMock()
-        fake_client.chat.completions.create.side_effect = RuntimeError("boom")
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(fake_client, "judge-model"),
+            "agent.auxiliary_client.call_llm",
+            side_effect=RuntimeError("boom"),
         ):
             verdict, reason, _, _wd = goals.judge_goal("goal", "response")
         assert verdict == "continue"
@@ -189,17 +187,11 @@ def test_api_error_continues(self):
     def test_judge_says_done(self):
         from hermes_cli import goals
 
-        fake_client = MagicMock()
-        fake_client.chat.completions.create.return_value = MagicMock(
-            choices=[
-                MagicMock(
-                    message=MagicMock(content='{"done": true, "reason": "achieved"}')
-                )
-            ]
-        )
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(fake_client, "judge-model"),
+            "agent.auxiliary_client.call_llm",
+            return_value=MagicMock(
+                choices=[MagicMock(message=MagicMock(content='{"done": true, "reason": "achieved"}'))]
+            ),
         ):
             verdict, reason, _, _wd = goals.judge_goal("goal", "agent response")
         assert verdict == "done"
@@ -208,17 +200,11 @@ def test_judge_says_done(self):
     def test_judge_says_continue(self):
         from hermes_cli import goals
 
-        fake_client = MagicMock()
-        fake_client.chat.completions.create.return_value = MagicMock(
-            choices=[
-                MagicMock(
-                    message=MagicMock(content='{"done": false, "reason": "not yet"}')
-                )
-            ]
-        )
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(fake_client, "judge-model"),
+            "agent.auxiliary_client.call_llm",
+            return_value=MagicMock(
+                choices=[MagicMock(message=MagicMock(content='{"done": false, "reason": "not yet"}'))]
+            ),
         ):
             verdict, reason, _, _wd = goals.judge_goal("goal", "agent response")
         assert verdict == "continue"
@@ -448,11 +434,9 @@ def test_api_error_does_not_count_as_parse_failure(self):
         """Transient network/API errors must not trip the auto-pause guard."""
         from hermes_cli import goals
 
-        fake_client = MagicMock()
-        fake_client.chat.completions.create.side_effect = RuntimeError("connection reset")
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(fake_client, "judge-model"),
+            "agent.auxiliary_client.call_llm",
+            side_effect=RuntimeError("connection reset"),
         ):
             verdict, _, parse_failed, _wd = goals.judge_goal("goal", "response")
         assert verdict == "continue"
@@ -462,13 +446,9 @@ def test_empty_judge_reply_flagged_as_parse_failure(self):
         """End-to-end: judge returns empty content → parse_failed=True."""
         from hermes_cli import goals
 
-        fake_client = MagicMock()
-        fake_client.chat.completions.create.return_value = MagicMock(
-            choices=[MagicMock(message=MagicMock(content=""))]
-        )
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(fake_client, "judge-model"),
+            "agent.auxiliary_client.call_llm",
+            return_value=MagicMock(choices=[MagicMock(message=MagicMock(content=""))]),
         ):
             verdict, _, parse_failed, _wd = goals.judge_goal("goal", "response")
         assert verdict == "continue"
@@ -747,22 +727,11 @@ class _FakeChoice:
             message = _FakeMsg()
         class _FakeResp:
             choices = [_FakeChoice()]
-        class _FakeClient:
-            class chat:
-                class completions:
-                    @staticmethod
-                    def create(**kwargs):
-                        captured.update(kwargs)
-                        return _FakeResp()
-
-        with patch.object(goals, "get_text_auxiliary_client",
-                          return_value=(_FakeClient, "fake-model"), create=True), \
-             patch.object(goals, "get_auxiliary_extra_body",
-                          return_value=None, create=True), \
-             patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(_FakeClient, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body",
-                   return_value=None):
+        def _fake_call_llm(**kwargs):
+            captured.update(kwargs)
+            return _FakeResp()
+
+        with patch("agent.auxiliary_client.call_llm", side_effect=_fake_call_llm):
             verdict, reason, parse_failed, _wd = goals.judge_goal(
                 "ship the feature",
                 "ok shipped",
@@ -790,18 +759,11 @@ class _FakeChoice:
             message = _FakeMsg()
         class _FakeResp:
             choices = [_FakeChoice()]
-        class _FakeClient:
-            class chat:
-                class completions:
-                    @staticmethod
-                    def create(**kwargs):
-                        captured.update(kwargs)
-                        return _FakeResp()
-
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(_FakeClient, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body",
-                   return_value=None):
+        def _fake_call_llm(**kwargs):
+            captured.update(kwargs)
+            return _FakeResp()
+
+        with patch("agent.auxiliary_client.call_llm", side_effect=_fake_call_llm):
             goals.judge_goal("ship it", "done", subgoals=None)
 
         sent_messages = captured.get("messages") or []
@@ -1363,7 +1325,8 @@ def test_persistence_roundtrip(self, hermes_home):
 
 
 class TestJudgeWithContract:
-    def _fake_client(self, captured, content='{"done": false, "reason": "more"}'):
+    def _fake_call_llm(self, captured, content='{"done": false, "reason": "more"}'):
+        """judge_goal routes through call_llm (#35566) — capture its kwargs."""
         class _FakeMsg:
             pass
         _FakeMsg.content = content
@@ -1371,14 +1334,11 @@ class _FakeChoice:
             message = _FakeMsg()
         class _FakeResp:
             choices = [_FakeChoice()]
-        class _FakeClient:
-            class chat:
-                class completions:
-                    @staticmethod
-                    def create(**kwargs):
-                        captured.update(kwargs)
-                        return _FakeResp()
-        return _FakeClient
+
+        def _fake(**kwargs):
+            captured.update(kwargs)
+            return _FakeResp()
+        return _fake
 
     def test_judge_uses_contract_template(self, hermes_home):
         from unittest.mock import patch
@@ -1386,10 +1346,8 @@ def test_judge_uses_contract_template(self, hermes_home):
         from hermes_cli.goals import GoalContract
 
         captured = {}
-        client = self._fake_client(captured)
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(client, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value=None):
+        with patch("agent.auxiliary_client.call_llm",
+                   side_effect=self._fake_call_llm(captured)):
             goals.judge_goal(
                 "ship it", "I think it's done",
                 contract=GoalContract(verification="pytest -q passes"),
@@ -1407,10 +1365,8 @@ def test_contract_plus_subgoals_combine(self, hermes_home):
         from hermes_cli.goals import GoalContract
 
         captured = {}
-        client = self._fake_client(captured)
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(client, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value=None):
+        with patch("agent.auxiliary_client.call_llm",
+                   side_effect=self._fake_call_llm(captured)):
             goals.judge_goal(
                 "ship it", "done",
                 subgoals=["write changelog"],
@@ -1438,16 +1394,8 @@ class _FakeChoice:
             message = _FakeMsg()
         class _FakeResp:
             choices = [_FakeChoice()]
-        class _FakeClient:
-            class chat:
-                class completions:
-                    @staticmethod
-                    def create(**kwargs):
-                        return _FakeResp()
-
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(_FakeClient, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value=None):
+        with patch("agent.auxiliary_client.call_llm",
+                   return_value=_FakeResp()):
             contract = goals.draft_contract("Migrate auth to JWT")
         assert contract is not None
         assert contract.outcome == "auth on JWT"
@@ -1464,24 +1412,16 @@ class _FakeChoice:
             message = _FakeMsg()
         class _FakeResp:
             choices = [_FakeChoice()]
-        class _FakeClient:
-            class chat:
-                class completions:
-                    @staticmethod
-                    def create(**kwargs):
-                        return _FakeResp()
-
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(_FakeClient, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value=None):
+        with patch("agent.auxiliary_client.call_llm",
+                   return_value=_FakeResp()):
             assert goals.draft_contract("anything") is None
 
     def test_draft_returns_none_when_no_client(self, hermes_home):
         from unittest.mock import patch
         from hermes_cli import goals
 
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(None, None)):
+        with patch("agent.auxiliary_client.call_llm",
+                   side_effect=RuntimeError("No LLM provider configured")):
             assert goals.draft_contract("anything") is None
 
 
@@ -1495,7 +1435,8 @@ class TestContractAndBackgroundCompose:
     the contract block and the background-process list to the judge, so it
     can return either done (evidence met) or wait (parked on the poller)."""
 
-    def _capture_client(self, captured, content='{"verdict": "wait", "wait_on_pid": 4242, "reason": "CI still running"}'):
+    def _capture_call_llm(self, captured, content='{"verdict": "wait", "wait_on_pid": 4242, "reason": "CI still running"}'):
+        """judge_goal routes through call_llm (#35566) — capture its kwargs."""
         class _FakeMsg:
             pass
         _FakeMsg.content = content
@@ -1503,14 +1444,11 @@ class _FakeChoice:
             message = _FakeMsg()
         class _FakeResp:
             choices = [_FakeChoice()]
-        class _FakeClient:
-            class chat:
-                class completions:
-                    @staticmethod
-                    def create(**kwargs):
-                        captured.update(kwargs)
-                        return _FakeResp()
-        return _FakeClient
+
+        def _fake(**kwargs):
+            captured.update(kwargs)
+            return _FakeResp()
+        return _fake
 
     def test_judge_prompt_carries_contract_and_background(self, hermes_home):
         from unittest.mock import patch
@@ -1518,14 +1456,12 @@ def test_judge_prompt_carries_contract_and_background(self, hermes_home):
         from hermes_cli.goals import GoalContract
 
         captured = {}
-        client = self._capture_client(captured)
         bg = [{
             "session_id": "ci-watch", "pid": 4242, "status": "running",
             "command": "wait_for_pr_green.sh 50501", "trigger": "exit",
         }]
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(client, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value=None):
+        with patch("agent.auxiliary_client.call_llm",
+                   side_effect=self._capture_call_llm(captured)):
             verdict, reason, parse_failed, wait_directive = goals.judge_goal(
                 "ship the PR",
                 "I pushed and started the CI watcher; waiting on it now.",
@@ -1550,14 +1486,12 @@ def test_contract_goal_can_still_complete_on_evidence(self, hermes_home):
         from hermes_cli.goals import GoalContract
 
         captured = {}
-        client = self._capture_client(
-            captured,
-            content='{"verdict": "done", "reason": "CI is green, evidence shown"}',
-        )
         bg = [{"session_id": "ci", "pid": 4242, "status": "running", "command": "ci", "trigger": "exit"}]
-        with patch("agent.auxiliary_client.get_text_auxiliary_client",
-                   return_value=(client, "fake-model")), \
-             patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value=None):
+        with patch("agent.auxiliary_client.call_llm",
+                   side_effect=self._capture_call_llm(
+                       captured,
+                       content='{"verdict": "done", "reason": "CI is green, evidence shown"}',
+                   )):
             verdict, reason, parse_failed, wait_directive = goals.judge_goal(
                 "ship the PR",
                 "CI finished: 30 passed, 0 failed. Done.",
    • tests/hermes_cli/test_kanban_cli.py
      @@ -465,11 +465,10 @@ def test_run_slash_specify_end_to_end(kanban_home, monkeypatch):
     resp.choices[0].message.content = (
         '{"title": "Spec: rough idea", "body": "**Goal**\\nShip it."}'
     )
-    fake_client = MagicMock()
-    fake_client.chat.completions.create = MagicMock(return_value=resp)
+    # specify_task routes through call_llm now (#35566) — mock it directly.
     monkeypatch.setattr(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        lambda *a, **kw: (fake_client, "test-model"),
+        "agent.auxiliary_client.call_llm",
+        MagicMock(return_value=resp),
     )
 
     # Specify via slash.
    • tests/hermes_cli/test_kanban_decompose.py
      @@ -41,18 +41,19 @@ def _mock_client_returning(content: str):
 
 
 def _patch_aux_client(content: str, *, model: str = "test-model"):
-    client = _mock_client_returning(content)
+    # decompose_task now routes through call_llm (see #35566) — mock it at
+    # the source module so task config, extra_body, and retries stay out of
+    # unit-test scope.
     return patch(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        return_value=(client, model),
+        "agent.auxiliary_client.call_llm",
+        return_value=_fake_aux_response(content),
     )
 
 
 def _patch_extra_body():
-    return patch(
-        "agent.auxiliary_client.get_auxiliary_extra_body",
-        return_value={},
-    )
+    # No-op shim retained for call-site compatibility: extra_body plumbing
+    # now lives inside call_llm, which _patch_aux_client already mocks.
+    return patch("agent.auxiliary_client.get_auxiliary_extra_body", return_value={})
 
 
 def _patch_list_profiles(names: list[str]):
@@ -334,9 +335,11 @@ def test_decompose_no_aux_client_configured(kanban_home):
     for p in patches:
         p.start()
     try:
+        # call_llm raises RuntimeError when no provider is configured; the
+        # decomposer must convert that into a failed outcome, not a crash.
         with patch(
-            "agent.auxiliary_client.get_text_auxiliary_client",
-            return_value=(None, ""),
+            "agent.auxiliary_client.call_llm",
+            side_effect=RuntimeError("No LLM provider configured"),
         ):
             outcome = decomp.decompose_task(tid, author="me")
     finally:
@@ -344,4 +347,5 @@ def test_decompose_no_aux_client_configured(kanban_home):
             p.stop()
 
     assert outcome.ok is False
-    assert "no auxiliary client" in outcome.reason
+    # call_llm's no-provider RuntimeError surfaces via the LLM-error branch.
+    assert "LLM error" in outcome.reason
    • tests/hermes_cli/test_kanban_specify.py
      @@ -48,15 +48,12 @@ def _mock_client_returning(content: str):
 
 
 def _patch_aux_client(content: str, *, model: str = "test-model"):
-    """Patch get_text_auxiliary_client at its source + at the module that
-    imported it lazily inside specify_task. Both patches are needed
-    because kanban_specify imports the function inside the function body.
+    """Patch call_llm at its source module — specify_task now routes through
+    it (#35566) instead of building a raw client. Returns (patcher, mock) so
+    callers can still assert on the call.
     """
-    client = _mock_client_returning(content)
-    return patch(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        return_value=(client, model),
-    ), client
+    mock_fn = MagicMock(return_value=_fake_aux_response(content))
+    return patch("agent.auxiliary_client.call_llm", mock_fn), mock_fn
 
 
 # ---------------------------------------------------------------------------
@@ -159,13 +156,14 @@ def test_specify_task_no_aux_client_configured(kanban_home):
         tid = kb.create_task(conn, title="rough", triage=True)
 
     with patch(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        return_value=(None, ""),
+        "agent.auxiliary_client.call_llm",
+        side_effect=RuntimeError("No LLM provider configured"),
     ):
         outcome = spec.specify_task(tid)
 
     assert outcome.ok is False
-    assert "auxiliary client" in outcome.reason
+    # call_llm's no-provider RuntimeError surfaces via the LLM-error branch.
+    assert "LLM error" in outcome.reason
     # Task must stay in triage — we never touched it.
     with kb.connect() as conn:
         assert kb.get_task(conn, tid).status == "triage"
@@ -176,10 +174,9 @@ def test_specify_task_llm_api_error_keeps_task_in_triage(kanban_home):
         tid = kb.create_task(conn, title="rough", triage=True)
 
     client = MagicMock()
-    client.chat.completions.create = MagicMock(side_effect=RuntimeError("429 rate limited"))
     with patch(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        return_value=(client, "test-model"),
+        "agent.auxiliary_client.call_llm",
+        side_effect=RuntimeError("429 rate limited"),
     ):
         outcome = spec.specify_task(tid)
 
@@ -288,8 +285,8 @@ def test_cli_specify_all_returns_1_when_every_task_fails(kanban_home, capsys):
         kb.create_task(conn, title="b", triage=True)
 
     with patch(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        return_value=(None, ""),  # no aux client → every task fails
+        "agent.auxiliary_client.call_llm",
+        side_effect=RuntimeError("No LLM provider configured"),  # every task fails
     ):
         rc = _run_cli("specify", "--all")
    • tests/hermes_cli/test_profile_describer.py
      @@ -79,11 +79,11 @@ def _fake_aux_response(content: str):
 
 
 def _patch_aux_client(content: str):
-    client = MagicMock()
-    client.chat.completions.create = MagicMock(return_value=_fake_aux_response(content))
+    # describe_profile now routes through call_llm (#35566) — mock it at the
+    # source module.
     return patch(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        return_value=(client, "test-model"),
+        "agent.auxiliary_client.call_llm",
+        return_value=_fake_aux_response(content),
     )
    • tests/plugins/test_kanban_dashboard_plugin.py
      @@ -2182,13 +2182,10 @@ def _patch_specifier_response(monkeypatch, *, content, model="test-model"):
     resp = MagicMock()
     resp.choices = [MagicMock()]
     resp.choices[0].message.content = content
-    fake_client = MagicMock()
-    fake_client.chat.completions.create = MagicMock(return_value=resp)
-    monkeypatch.setattr(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        lambda *a, **kw: (fake_client, model),
-    )
-    return fake_client
+    # specify_task routes through call_llm now (#35566) — mock it directly.
+    fake_call = MagicMock(return_value=resp)
+    monkeypatch.setattr("agent.auxiliary_client.call_llm", fake_call)
+    return fake_call
 
 
 def test_specify_happy_path(client, monkeypatch):
@@ -2250,11 +2247,11 @@ def test_specify_no_aux_client_surfaces_reason(client, monkeypatch):
         json={"title": "rough", "triage": True},
     ).json()["task"]
 
-    # Simulate "no auxiliary client configured".
-    monkeypatch.setattr(
-        "agent.auxiliary_client.get_text_auxiliary_client",
-        lambda *a, **kw: (None, ""),
-    )
+    # Simulate "no auxiliary client configured" — call_llm raises when
+    # no provider resolves (#35566 routing).
+    def _no_provider(**kwargs):
+        raise RuntimeError("No LLM provider configured")
+    monkeypatch.setattr("agent.auxiliary_client.call_llm", _no_provider)
 
     r = client.post(
         f"/api/plugins/kanban/tasks/{t['id']}/specify",
@@ -2263,7 +2260,8 @@ def test_specify_no_aux_client_surfaces_reason(client, monkeypatch):
     assert r.status_code == 200
     body = r.json()
     assert body["ok"] is False
-    assert "auxiliary client" in body["reason"]
+    # call_llm's no-provider RuntimeError surfaces via the LLM-error branch.
+    assert "LLM error" in body["reason"]
 
     # Task must stay in triage — nothing was touched.
     detail = client.get(f"/api/plugins/kanban/tasks/{t['id']}").json()["task"]

1 PR closed without merging

From the close comments: done elsewhere 1

  • #60356 fix: use max_completion_tokens (not max_tokens) for OpenAI models in auxiliary LLM calls · done elsewhere → #65029, #60355
Closed issues · 1
  • Closed#60355Auxiliary LLM calls fail with `max_tokens` on gpt-4o/o-series/gpt-5 models (kanban decompose, specify, goals, profile describe)
How this cluster was grouped
  • Link evidence (structural links throughout): every member is connected by structural evidence: shared changed files, or a closing/superseded reference.
  • Pairwise link density 0.67: the share of item pairs that carry a link, counting a structural link in full and a relevance-only link at 0.36. 1.00 means every pair is linked.

Pick up a Lab task

Check whether a fix covers the open issues here, with your own agent. Bring back evidence for maintainers.

Open the Lab

At a glance

Subsystem
agent
Merged PRs
#65029
Closed issues
1
Attempts
1 PR closed without merging
Fix link checked
15 Jul 2026

For your agent

The same facts, typed.

cluster/60355.json
cluster/60355.md