diff --git a/lollms/app.py b/lollms/app.py
index fc5a679..60f6e85 100644
--- a/lollms/app.py
+++ b/lollms/app.py
@@ -529,6 +529,7 @@ class LollmsApplication(LoLLMsCom):
         internet_search_infos = []
         documentation = ""
         knowledge = ""
+        knowledge_infos = {"titles":[],"contents":[]}
 
 
         # boosting information
@@ -633,15 +634,17 @@ class LollmsApplication(LoLLMsCom):
                     self.personality.step_start("Building skills library")
                     if discussion is None:
                         discussion = self.recover_discussion(client_id)
-                    query = self.personality.fast_gen(f"!@>discussion:\n{discussion[-2048:]}\n!@>system: Read the discussion and craft a short skills database search query suited to recover needed information to reply to last {self.config.user_name} message.\nDo not answer the prompt. Do not add explanations.\n!@>search query: ", max_generation_size=256, show_progress=True, callback=self.personality.sink)
+                    query = self.personality.fast_gen(f"!@>system: Read the discussion and reformulate {self.config.user_name}'s request.\nDo not answer the request.\nDo not add explanations.\n!@>discussion:\n{discussion[-2048:]}\n!@>search query: ", max_generation_size=256, show_progress=True, callback=self.personality.sink)
                     # skills = self.skills_library.query_entry(query)
-                    skills, sorted_similarities, document_ids = self.skills_library.query_vector_db(query, top_k=3, max_dist=1000)#query_entry_fts(query)
-                    
+                    if self.config.debug:
+                        ASCIIColors.info(f"Query : {query}")
+                    skill_titles, skills = self.skills_library.query_vector_db(query, top_k=3, max_dist=1000)#query_entry_fts(query)
+                    knowledge_infos={"titles":skill_titles,"contents":skills}
                     if len(skills)>0:
                         if knowledge=="":
-                            knowledge=f"!@>knowledge:\n!@>instructions: Use the knowledge to answer {self.config.user_name}'s message. If you don't have enough information or you don't know how to answer, just say you do not know.\n"
-                        for i,(category, title, content) in enumerate(skills):
-                            knowledge += f"!@>knowledge {i}:\n!@>category:\n{category}\n!@>title:\n{title}\ncontent:\n{content}"
+                            knowledge=f"!@>knowledge:\n"
+                        for i,(title, content) in enumerate(zip(skill_titles,skills)):
+                            knowledge += f"!@>knowledge {i}:\n!@>title:\n{title}\ncontent:\n{content}"
                     self.personality.step_end("Building skills library")
                 except Exception as ex:
                     ASCIIColors.error(ex)
@@ -803,6 +806,7 @@ class LollmsApplication(LoLLMsCom):
             "internet_search_results":internet_search_results,
             "documentation":documentation,
             "knowledge":knowledge,
+            "knowledge_infos":knowledge_infos,
             "user_description":user_description,
             "discussion_messages":discussion_messages,
             "positive_boost":positive_boost,
diff --git a/lollms/databases/skills_database.py b/lollms/databases/skills_database.py
index 261ef2f..5971ae6 100644
--- a/lollms/databases/skills_database.py
+++ b/lollms/databases/skills_database.py
@@ -1,11 +1,11 @@
 import sqlite3
 from safe_store.text_vectorizer import TextVectorizer, VectorizationMethod, VisualizationMethod
+import numpy as np
 class SkillsLibrary:
         
     def __init__(self, db_path):
         self.db_path =db_path
         self._initialize_db()
-        self.vectorizer = TextVectorizer(VectorizationMethod.TFIDF_VECTORIZER)
        
 
     def _initialize_db(self):
@@ -120,33 +120,33 @@ class SkillsLibrary:
         conn.close()
         return res
 
-    def query_vector_db(self, query, top_k=3, max_dist=1000):
+    def query_vector_db(self, query_, top_k=3, max_dist=1000):
+        vectorizer = TextVectorizer(VectorizationMethod.TFIDF_VECTORIZER)
         conn = sqlite3.connect(self.db_path)
         cursor = conn.cursor()
         # Use direct string concatenation for the MATCH expression.
         # Ensure text is safely escaped to avoid SQL injection.
-        query = "SELECT title FROM skills_library"
+        query = "SELECT id, title FROM skills_library"
         cursor.execute(query)
         res = cursor.fetchall()
         cursor.close()
         conn.close()
         for entry in res:
-            self.vectorizer.add_document(entry[0])
-        self.vectorizer.index()
+            vectorizer.add_document(entry[0],entry[1])
+        vectorizer.index()
         
-        skill_titles, sorted_similarities, document_ids = self.vectorizer.recover_text(query, top_k)
+        skill_titles, sorted_similarities, document_ids = vectorizer.recover_text(query_, top_k)
         skills = []
-        for skill, sim in zip(skill_titles, sorted_similarities):
-            if sim>max_dist:
+        for skill_title, sim, id in zip(skill_titles, sorted_similarities, document_ids):
+            if  np.linalg.norm(sim[1])<max_dist:
                 conn = sqlite3.connect(self.db_path)
                 cursor = conn.cursor()
                 # Use direct string concatenation for the MATCH expression.
                 # Ensure text is safely escaped to avoid SQL injection.
-                query = "SELECT content FROM skills_library WHERE title LIKE ?"
-                res = cursor.execute(query, (skill,))
-                skills.append(res[0])
-                cursor.execute(query)
+                query = "SELECT content FROM skills_library WHERE id = ?"
+                cursor.execute(query, (id,))
                 res = cursor.fetchall()
+                skills.append(res[0])
                 cursor.close()
                 conn.close()
         return skill_titles, skills