Implement core skills: Code validation, model fine-tuning, and system diagnostics

- Added `ModelFineTuner` class for preparing training data and fine-tuning models based on user corrections.
- Introduced `CodeValidator` class to validate generated code against various hardware and style rules, including safety checks and function naming conventions.
- Developed skills for calculator operations, system information retrieval, weather fetching, and timer functionality.
- Implemented a self-diagnostic skill to run unit tests and report results.
- Created a dynamic skill loading mechanism to discover and register skills from the current directory.
- Added unit tests for skills to ensure functionality and reliability.

skills/timer.py

@@ -0,0 +1,44 @@
+import time
+import re
+import threading
+
+def meta():
+    """
+    Metadata for the Timer skill.
+    """
+    return {
+        "name": "Timer",
+        "description": "Sets a non-blocking timer (background thread).",
+        "triggers": ["timer", "sleep", "wait for"]
+    }
+
+def run(payload):
+    """
+    Executes the blocking sleep.
+    """
+    prompt = payload if isinstance(payload, str) else payload.get("prompt", "")
+    
+    # Regex to capture number and optional unit (e.g., "5", "5s", "5 minutes")
+    match = re.search(r'(\d+)\s*(seconds?|secs?|s|minutes?|mins?|m)?', prompt.lower())
+    
+    if not match:
+        return None # Fallback to LLM if no time found
+        
+    amount = int(match.group(1))
+    unit = match.group(2)
+    
+    duration = amount
+    if unit and unit.startswith('m'):
+        duration *= 60
+        
+    if duration > 3600:
+        return f"❌ Timer too long ({duration}s). Max 1 hour."
+        
+    def _timer_thread():
+        time.sleep(duration)
+        print(f"\n\n⏰ 🔔 BEEP! Timer finished ({duration}s).\n")
+
+    t = threading.Thread(target=_timer_thread, daemon=True)
+    t.start()
+    
+    return f"⏰ Timer started for {duration} seconds (running in background)..."