diff --git a/plugins/skill_scanner/plugin.py b/plugins/skill_scanner/plugin.py
index 1cd6190..f1401ac 100644
--- a/plugins/skill_scanner/plugin.py
+++ b/plugins/skill_scanner/plugin.py
@@ -51,25 +51,91 @@ class SkillOCRThread(QThread):
             self.scan_error.emit(str(e))
     
     def _parse_skills(self, text):
-        """Parse skill data from OCR text."""
+        """Parse skill data from OCR text with improved handling for 3-column layout."""
         skills = {}
+        
+        # Ranks in Entropia Universe (in order)
+        RANKS = [
+            'Newbie', 'Inept', 'Beginner', 'Amateur', 'Average',
+            'Skilled', 'Expert', 'Professional', 'Master', 'Grand Master',
+            'Champion', 'Legendary', 'Guru', 'Astonishing', 'Remarkable',
+            'Outstanding', 'Marvelous', 'Prodigious', 'Amazing', 'Incredible', 'Awesome'
+        ]
+        rank_pattern = '|'.join(RANKS)
+        
+        # Clean up the text - remove common headers and junk
+        text = text.replace('SKILLS', '').replace('ALL CATEGORIES', '')
+        text = text.replace('SKILL NAME', '').replace('RANK', '').replace('POINTS', '')
+        text = text.replace('Attributes', '').replace('COMBAT', '').replace('Design', '')
+        text = text.replace('Construction', '').replace('Defense', '').replace('General', '')
+        text = text.replace('Handgun', '').replace('Heavy Melee Weapons', '')
+        text = text.replace('Information', '').replace('Inflict Melee Damage', '')
+        text = text.replace('Inflict Ranged Damage', '').replace('Light Melee Weapons', '')
+        text = text.replace('Longblades', '').replace('Medical', '').replace('Mining', '')
+        text = text.replace('Science', '').replace('Social', '').replace('Beauty', '')
+        text = text.replace('Mindforce', '')
+        
         lines = text.split('\n')
         
         for line in lines:
-            # Pattern: SkillName Rank Points
+            line = line.strip()
+            if not line:
+                continue
+            
+            # Skip category headers and short lines
+            if len(line) < 10:
+                continue
+            
+            # Try pattern: SkillName Rank Points
+            # More flexible pattern to handle merged text
             match = re.search(
-                r'(\w+(?:\s+\w+)*)\s+(Newbie|Inept|Beginner|Amateur|Average|Skilled|Expert|Professional|Master|Grand Master|Champion|Legendary|Guru|Astonishing|Remarkable|Outstanding|Marvelous|Prodigious|Amazing|Incredible|Awesome)\s+(\d+)',
+                rf'([A-Za-z][A-Za-z\s]{{2,50}}?)\s+({rank_pattern})\s+(\d{{1,6}})(?:\s|$)',
                 line, re.IGNORECASE
             )
+            
             if match:
                 skill_name = match.group(1).strip()
                 rank = match.group(2)
                 points = int(match.group(3))
-                skills[skill_name] = {
-                    'rank': rank,
-                    'points': points,
-                    'scanned_at': datetime.now().isoformat()
-                }
+                
+                # Clean up skill name
+                skill_name = skill_name.strip()
+                
+                # Validate - points should be reasonable (not too small)
+                if points > 0:
+                    skills[skill_name] = {
+                        'rank': rank,
+                        'points': points,
+                        'scanned_at': datetime.now().isoformat()
+                    }
+                    print(f"[SkillScanner] Parsed: {skill_name} = {rank} ({points})")
+        
+        # Alternative parsing: try to find skill-rank-points triplets
+        if not skills:
+            skills = self._parse_skills_alternative(text, RANKS)
+        
+        return skills
+    
+    def _parse_skills_alternative(self, text, ranks):
+        """Alternative parser for when text is heavily merged."""
+        skills = {}
+        
+        # Find all rank positions in the text
+        for rank in ranks:
+            # Look for pattern: [text] [Rank] [number]
+            pattern = rf'([A-Z][a-z]{{2,}}(?:\s+[A-Z][a-z]{{2,}}){{0,3}})\s+{rank}\s+(\d{{1,6}})'
+            matches = re.finditer(pattern, text, re.IGNORECASE)
+            
+            for match in matches:
+                skill_name = match.group(1).strip()
+                points = int(match.group(2))
+                
+                if points > 0 and len(skill_name) > 2:
+                    skills[skill_name] = {
+                        'rank': rank,
+                        'points': points,
+                        'scanned_at': datetime.now().isoformat()
+                    }
         
         return skills