diff --git a/agents-api/agents_api/queries/docs/search_docs_by_embedding.py b/agents-api/agents_api/queries/docs/search_docs_by_embedding.py
index 61a0506f2..e9e3380e1 100644
--- a/agents-api/agents_api/queries/docs/search_docs_by_embedding.py
+++ b/agents-api/agents_api/queries/docs/search_docs_by_embedding.py
@@ -78,4 +78,3 @@ async def search_docs_by_embedding(
             metadata_filter,
         ],
     )
-
diff --git a/agents-api/tests/test_docs_queries.py b/agents-api/tests/test_docs_queries.py
index dfd48858b..03ef50aba 100644
--- a/agents-api/tests/test_docs_queries.py
+++ b/agents-api/tests/test_docs_queries.py
@@ -22,34 +22,36 @@
 
 import math
 
+
 def make_vector_with_similarity(n: int, d: float):
     """
     Returns a list `v` of length `n` such that the cosine similarity
     between `v` and the all-ones vector of length `n` is approximately d.
     """
     if not -1.0 <= d <= 1.0:
-        raise ValueError("d must lie in [-1, 1].")
-    
+        msg = "d must lie in [-1, 1]."
+        raise ValueError(msg)
+
     # Handle special cases exactly:
     if abs(d - 1.0) < 1e-12:  # d ~ +1
         return [1.0] * n
     if abs(d + 1.0) < 1e-12:  # d ~ -1
         return [-1.0] * n
-    if abs(d) < 1e-12:        # d ~ 0
-        v = [0.0]*n
+    if abs(d) < 1e-12:  # d ~ 0
+        v = [0.0] * n
         if n >= 2:
             v[0] = 1.0
             v[1] = -1.0
         return v
 
     sign_d = 1.0 if d >= 0 else -1.0
-    
+
     # Base part: sign(d)*[1,1,...,1]
-    base = [sign_d]*n
-    
+    base = [sign_d] * n
+
     # Orthogonal unit vector u with sum(u)=0; for simplicity:
     #   u = [1/sqrt(2), -1/sqrt(2), 0, 0, ..., 0]
-    u = [0.0]*n
+    u = [0.0] * n
     if n >= 2:
         u[0] = 1.0 / math.sqrt(2)
         u[1] = -1.0 / math.sqrt(2)
@@ -57,15 +59,16 @@ def make_vector_with_similarity(n: int, d: float):
 
     # Solve for alpha:
     # alpha^2 = n*(1 - d^2)/d^2
-    alpha = math.sqrt(n*(1 - d*d)) / abs(d)
+    alpha = math.sqrt(n * (1 - d * d)) / abs(d)
 
     # Construct v
-    v = [0.0]*n
+    v = [0.0] * n
     for i in range(n):
         v[i] = base[i] + alpha * u[i]
-    
+
     return v
 
+
 @test("query: create user doc")
 async def _(dsn=pg_dsn, developer=test_developer, user=test_user):
     pool = await create_db_pool(dsn=dsn)
@@ -329,7 +332,10 @@ async def _(
 
 @test("query: search docs by embedding with confidence")
 async def _(
-    dsn=pg_dsn, agent=test_agent, developer=test_developer, doc=test_doc_with_embedding, 
+    dsn=pg_dsn,
+    agent=test_agent,
+    developer=test_developer,
+    doc=test_doc_with_embedding,
 ):
     pool = await create_db_pool(dsn=dsn)
 
@@ -344,7 +350,7 @@ async def _(
         developer_id=developer.id,
         owners=[("agent", agent.id)],
         embedding=query_embedding,
-        confidence=confidence*0.9,
+        confidence=confidence * 0.9,
         k=3,  # Add k parameter
         metadata_filter={"test": "test"},  # Add metadata filter
         connection_pool=pool,
@@ -358,7 +364,7 @@ async def _(
         developer_id=developer.id,
         owners=[("agent", agent.id)],
         embedding=query_embedding,
-        confidence=confidence*1.1,
+        confidence=confidence * 1.1,
         k=3,  # Add k parameter
         metadata_filter={"test": "test"},  # Add metadata filter
         connection_pool=pool,