small fixes to prompt

vision_agent/tools/tools.py

@@ -199,7 +199,7 @@ def extract_frames(
 def ocr(image: np.ndarray) -> List[Dict[str, Any]]:
  """'ocr' extracts text from an image. It returns a list of detected text, bounding
  boxes with normalized coordinates, and confidence scores. The results are sorted
- from top-left to bottom right. Each element in the return list represents a
+ from top-left to bottom right. Each element in the returned list represents a
  contiguous text region. Text on separate lines is considered separate regions and
  would be returned as separate elements.