docs: updated docs

docs/advanced-usage.md

@@ -452,9 +452,9 @@ Web search is enabled by default, allowing models to request Claude perform sear
 The server uses carefully crafted system prompts to give each tool specialized expertise:
 
 ### Prompt Architecture
-- **Centralized Prompts**: All system prompts are defined in `prompts/tool_prompts.py`
+- **Centralized Prompts**: Each tool's system prompt lives in `systemprompts/` (for example, `systemprompts/chat_prompt.py`)
 - **Tool Integration**: Each tool inherits from `BaseTool` and implements `get_system_prompt()`
-- **Prompt Flow**: `User Request → Tool Selection → System Prompt + Context → Gemini Response`
+- **Prompt Flow**: `User Request → Tool Selection → System Prompt + Context → Model Response`
 
 ### Specialized Expertise
 Each tool has a unique system prompt that defines its role and approach:
@@ -465,6 +465,6 @@ Each tool has a unique system prompt that defines its role and approach:
 
 ### Customization
 To modify tool behavior, you can:
-1. Edit prompts in `prompts/tool_prompts.py` for global changes
+1. Edit the prompt file in `systemprompts/` (and export it via `systemprompts/__init__.py`) for global changes
 2. Override `get_system_prompt()` in a tool class for tool-specific changes
 3. Use the `temperature` parameter to adjust response style (0.2 for focused, 0.7 for creative)

docs/locale-configuration.md

@@ -53,10 +53,14 @@ def get_language_instruction(self) -> str:
     Returns:
         str: Language instruction to prepend to prompt, or empty string if no locale set
     """
-    from config import LOCALE
-    if not LOCALE or not LOCALE.strip():
+    import os
+
+    locale = os.getenv("LOCALE", "").strip()
+
+    if not locale:
         return ""
-    return f"Always respond in {LOCALE.strip()}.\n\n"
+
+    return f"Always respond in {locale}.\n\n"
 ```
 
 ### Integration in Tool Execution
@@ -80,7 +84,7 @@ system_prompt = language_instruction + base_system_prompt
    ```
 2. Restart the MCP server:
    ```bash
-   python server.py
+   ./run-server.sh
    ```
 3. Use any tool – responses will be in the specified language.
 
@@ -153,11 +157,14 @@ To customize the language instruction, modify the `get_language_instruction()` m
 
 ```python
 def get_language_instruction(self) -> str:
-    from config import LOCALE
-    if not LOCALE or not LOCALE.strip():
+    import os
+
+    locale = os.getenv("LOCALE", "").strip()
+
+    if not locale:
         return ""
     # Custom instruction
-    return f"Always respond in {LOCALE.strip()} and use a professional tone.\n\n"
+    return f"Always respond in {locale} and use a professional tone.\n\n"
 ```
 
 ### Per-Tool Customization
@@ -167,10 +174,13 @@ You can also override the method in specific tools for custom behavior:
 ```python
 class MyCustomTool(SimpleTool):
     def get_language_instruction(self) -> str:
-        from config import LOCALE
-        if LOCALE == "fr-FR":
+        import os
+
+        locale = os.getenv("LOCALE", "").strip()
+
+        if locale == "fr-FR":
             return "Respond in French with precise technical vocabulary.\n\n"
-        elif LOCALE == "zh-CN":
+        elif locale == "zh-CN":
             return "请用中文回答，使用专业术语。\n\n"
         else:
             return super().get_language_instruction()

docs/tools/precommit.md

@@ -94,8 +94,8 @@ Then it shared the changes and relevant code with Gemini 2.5 Pro - the following
 
   🚨 [HIGH PRIORITY] Critical Issue Found
 
-  Inconsistent Provider Handling in tools/base.py
-  - Location: tools/base.py line ~1833 in _create_continuation_offer_response method
+  Inconsistent Provider Handling in tools/shared/base_tool.py
+  - Location: tools/shared/base_tool.py inside _create_continuation_offer_response
   - Problem: The defensive provider handling pattern was applied in 2 locations but missed a 3rd location
   - Impact: Risk of AttributeError: 'str' object has no attribute 'get_provider_type' in continuation scenarios
   - Fix Required: Apply the same defensive pattern to the missed location