Run pylint workflow only on pull requests

.github/workflows/pylint.yml

@@ -1,6 +1,6 @@
 name: Pylint
 
-on: [ push ]
+on: [ pull_request ]
 
 jobs:
   build:

tests/test_local_llm.py

@@ -2,17 +2,28 @@
 
 
 class DummyModel:
-    def __call__(self, *_args, **_kwargs):
-        return {"choices": [{"text": "ok"}]}
+    """Simple callable model stub for local LLM tests."""
 
+    def __init__(self):
+        self.last_kwargs = {}
 
-def test_normalize_max_tokens():
-    assert local_llm._normalize_max_tokens(-1) == local_llm.LOCAL_LLM_MAX_TOKENS
-    assert local_llm._normalize_max_tokens("bad") == local_llm.LOCAL_LLM_MAX_TOKENS
-    assert local_llm._normalize_max_tokens(1_000_000) == local_llm.LOCAL_LLM_CONTEXT
+    def __call__(self, *_args, **kwargs):
+        self.last_kwargs = kwargs
+        return {"choices": [{"text": "ok"}]}
 
 
 def test_ask_local_llm_empty_prompt(monkeypatch):
-    monkeypatch.setattr(local_llm, "LLM_MODEL", DummyModel())
+    dummy_model = DummyModel()
+    monkeypatch.setattr(local_llm, "LLM_MODEL", dummy_model)
     result = local_llm.ask_local_llm("   ")
     assert result == "[Local LLM error: Empty prompt]"
+
+
+def test_ask_local_llm_normalizes_max_tokens(monkeypatch):
+    dummy_model = DummyModel()
+    monkeypatch.setattr(local_llm, "LLM_MODEL", dummy_model)
+
+    result = local_llm.ask_local_llm("build plan", max_tokens=-1)
+
+    assert result == "ok"
+    assert dummy_model.last_kwargs["max_tokens"] == local_llm.LOCAL_LLM_MAX_TOKENS