fix(desktop): keep model runtime state per session (#43702)

* fix(desktop): keep model runtime state per session

(cherry picked from commit f72ee87d99ee38cb7b5badeb9a8af869bb92073a)

* fix(desktop): keep footer model state scoped to active session

(cherry picked from commit d91942ebd4671ff857b5c8526dbf133f04782ecb)

* fix(desktop): restore stored runtime when resuming sessions

(cherry picked from commit 32b3793418257617b8da57e26151f079c2620d00)

* fix(desktop): persist live runtime changes for resume

(cherry picked from commit c58467779436dcef44a80ad55b52664752dc0837)

* fix(desktop): persist resumed endpoint runtime

* chore(attribution): map pinguarmy's commit email in AUTHOR_MAP

The salvaged commits on this branch preserve @pinguarmy's authorship
(郝鹏宇 / peterhao@Peters-MacBook-Air.local). Add the mapping so the
check-attribution CI gate resolves the email to the GitHub username.

---------

Co-authored-by: 郝鹏宇 <peterhao@Peters-MacBook-Air.local>

tests/test_tui_gateway_server.py

@@ -878,6 +878,97 @@ def test_session_resume_uses_parent_lineage_for_display(monkeypatch):
     assert captured["history_calls"] == [("tip", False), ("tip", True)]
 
 
+def test_session_resume_passes_stored_runtime_to_agent(monkeypatch):
+    captured = {}
+
+    class FakeDB:
+        def get_session(self, target):
+            return {
+                "id": target,
+                "model": "gpt-5.4",
+                "billing_provider": "openai-codex",
+                "model_config": '{"reasoning_config":{"enabled":true,"effort":"high"},"service_tier":"priority","base_url":"https://custom.example/v1","api_mode":"chat_completions"}',
+            }
+
+        def reopen_session(self, target):
+            pass
+
+        def get_messages_as_conversation(self, target, include_ancestors=False):
+            return [{"role": "user", "content": "hello"}]
+
+    def fake_make_agent(sid, key, session_id=None, session_db=None, **kwargs):
+        captured.update(kwargs)
+        return types.SimpleNamespace(model="gpt-5.4", provider="openai-codex")
+
+    monkeypatch.setattr(server, "_get_db", lambda: FakeDB())
+    monkeypatch.setattr(server, "_enable_gateway_prompts", lambda: None)
+    monkeypatch.setattr(server, "_set_session_context", lambda target: [])
+    monkeypatch.setattr(server, "_clear_session_context", lambda tokens: None)
+    monkeypatch.setattr(server, "_make_agent", fake_make_agent)
+    monkeypatch.setattr(server, "_session_info", lambda agent, *a: {"model": agent.model, "provider": agent.provider})
+
+    def fake_init_session(sid, key, agent, history, cols=80):
+        server._sessions[sid] = {"agent": agent, "session_key": key}
+
+    monkeypatch.setattr(server, "_init_session", fake_init_session)
+
+    resp = server.handle_request(
+        {"id": "1", "method": "session.resume", "params": {"session_id": "stored-session"}}
+    )
+
+    assert resp["result"]["info"] == {"model": "gpt-5.4", "provider": "openai-codex"}
+    assert captured["model_override"] == {
+        "model": "gpt-5.4",
+        "provider": "openai-codex",
+        "base_url": "https://custom.example/v1",
+        "api_mode": "chat_completions",
+    }
+    assert captured["provider_override"] == "openai-codex"
+    assert captured["reasoning_config_override"] == {"enabled": True, "effort": "high"}
+    assert captured["service_tier_override"] == "priority"
+    runtime_sid = resp["result"]["session_id"]
+    assert server._sessions[runtime_sid]["model_override"] == captured["model_override"]
+
+
+def test_persist_live_session_runtime_preserves_resume_metadata(monkeypatch):
+    updates = {}
+
+    class FakeDB:
+        def get_session(self, session_id):
+            assert session_id == "stored-session"
+            return {"model_config": '{"_branched_from":"root"}'}
+
+        def update_session_meta(self, session_id, model_config_json, model=None):
+            updates["meta"] = (session_id, json.loads(model_config_json), model)
+
+    agent = types.SimpleNamespace(
+        model="gpt-5.4",
+        provider="openai-codex",
+        base_url="https://custom.example/v1",
+        api_mode="chat_completions",
+        reasoning_config={"enabled": True, "effort": "high"},
+        service_tier="priority",
+        _session_db=FakeDB(),
+    )
+
+    server._persist_live_session_runtime({"agent": agent, "session_key": "stored-session"})
+
+    assert "model" not in updates
+    assert updates["meta"] == (
+        "stored-session",
+        {
+            "_branched_from": "root",
+            "model": "gpt-5.4",
+            "provider": "openai-codex",
+            "base_url": "https://custom.example/v1",
+            "api_mode": "chat_completions",
+            "reasoning_config": {"enabled": True, "effort": "high"},
+            "service_tier": "priority",
+        },
+        "gpt-5.4",
+    )
+
+
 def test_status_callback_emits_kind_and_text():
     with patch("tui_gateway.server._emit") as emit:
         cb = server._agent_cbs("sid")["status_callback"]
@@ -3593,8 +3684,9 @@ def test_session_info_includes_mcp_servers(monkeypatch):
     fake_mod.get_mcp_status = lambda: fake_status
     monkeypatch.setitem(sys.modules, "tools.mcp_tool", fake_mod)
 
-    info = server._session_info(types.SimpleNamespace(tools=[], model=""))
+    info = server._session_info(types.SimpleNamespace(tools=[], model="", provider="openai-codex"))
 
+    assert info["provider"] == "openai-codex"
     assert info["mcp_servers"] == fake_status
 
 
@@ -5983,6 +6075,45 @@ def test_make_agent_defaults_to_90(monkeypatch):
     assert mock_agent.call_args.kwargs["max_iterations"] == 90
 
 
+def test_make_agent_uses_session_runtime_overrides(monkeypatch):
+    _setup_make_agent_mocks(monkeypatch, {})
+    resolved = {}
+
+    def fake_resolve_runtime_provider(requested=None, target_model=None):
+        resolved["requested"] = requested
+        resolved["target_model"] = target_model
+        return {
+            "provider": requested,
+            "base_url": None,
+            "api_key": None,
+            "api_mode": None,
+            "command": None,
+            "args": None,
+            "credential_pool": None,
+        }
+
+    monkeypatch.setattr(
+        "hermes_cli.runtime_provider.resolve_runtime_provider",
+        fake_resolve_runtime_provider,
+    )
+
+    with patch("run_agent.AIAgent") as mock_agent:
+        server._make_agent(
+            "sid1",
+            "key1",
+            model_override="gpt-5.4",
+            provider_override="openai-codex",
+            reasoning_config_override={"enabled": True, "effort": "high"},
+            service_tier_override="priority",
+        )
+
+    assert resolved == {"requested": "openai-codex", "target_model": "gpt-5.4"}
+    assert mock_agent.call_args.kwargs["model"] == "gpt-5.4"
+    assert mock_agent.call_args.kwargs["provider"] == "openai-codex"
+    assert mock_agent.call_args.kwargs["reasoning_config"] == {"enabled": True, "effort": "high"}
+    assert mock_agent.call_args.kwargs["service_tier"] == "priority"
+
+
 def test_make_agent_handles_null_agent_config(monkeypatch):
     _setup_make_agent_mocks(monkeypatch, {"agent": None, "max_turns": 80})