refactor: Remove explicit `google_api_key` assignment for Gemini model, relying on automatic environment variable pickup.

agent.py

@@ -66,7 +66,7 @@ openrouter_model = ChatOpenAI(
 # Google AI Studio Fallback Model (Gemini)
 gemini_model = ChatGoogleGenerativeAI(
     model="gemini-1.5-pro",
-    google_api_key=os.getenv("GOOGLE_API_KEY"),
+    # google_api_key is automatically picked up from GOOGLE_API_KEY environment variable
     temperature=0,
 )
 
@@ -392,24 +392,22 @@ def answer_message(state: AgentState) -> AgentState:
     
     TODAY'S EXACT DATE is {current_date}. Keep this in mind for all time-sensitive queries.
 
-    CRITICAL RULES FOR SEARCH & TOOLS:
-    1. If an image, video, or audio file is attached, YOU MUST NOT SAY "I don't have access to analyze..." or "I cannot see". YOU ARE NOT BLIND. You have external APIs (analyze_image, analyze_video, analyze_audio) that will act as your eyes and ears! ALWAYS invoke these tools immediately to get descriptions!
-    2. If a text/data file is attached, use the appropriate tool (run_python_script, read_document) to analyze the file content.
-    3. Use run_python_script freely to process data (pandas), read complex documents (.xlsx, .pdf), or do heavy math calculations.
-    4. When using tools like web_search or wiki_search, do not blindly search the entire question. Extract the core entities.
-    5. If the first search result doesn't contain the answer, THINK step-by-step, refine your search query (e.g., use synonyms, or search for broader concepts), and search again.
-    6. Cross-reference facts if they seem ambiguous.
+    CRITICAL RULES FOR FILES & TOOLS:
+    1. If a message contains a path like `[Attached File Local Path: ...]` followed by an image (.png, .jpg, .jpeg), video, or audio file, YOU MUST USE THE CORRESPONDING TOOL (analyze_image, analyze_video, analyze_audio).
+    2. YOU ARE NOT BLIND. NEVER say "I cannot see images" or "I don't have access to files". Use your tools to see and hear for you!
+    3. If you see a file path, invoke the tool IMMEDIATELY in your first ReAct step.
+    4. For text/data files, use `read_document` or `run_python_script` (especially for .xlsx or .pdf).
+    5. Be thorough. If one tool doesn't give enough info, use another (e.g., search the web for context).
 
-    Do not include any thought process before answering the question, and only response exactly what was being asked of you.
-    If you are not able to provide an answer, use tools or state the limitation that you're facing instead. 
+    Example of Tool Trigger:
+    User: "What is in this image? [Attached File Local Path: /path/to/image.png]"
+    Your Thought: "I need to see the image to answer. I will use the analyze_image tool."
+    Your Action: Call `analyze_image(image_path='/path/to/image.png', question='What is in this image?')`
 
     YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated list of numbers and/or strings. 
     If you are asked for a number, don't use comma to write your number, and don't use units such as $ or percent sign unless specified otherwise. 
     If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. 
     If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list is a number or a string.
-
-    Example question: How many hours are there in a day?
-    Response: 24    
     """)]
     messages = prompt + messages