diff --git a/bnd/config.py b/bnd/config.py
index 5a2da76..9746cf3 100644
--- a/bnd/config.py
+++ b/bnd/config.py
@@ -149,30 +149,29 @@ def _load_config() -> Config:
     return Config()
 
 
-def find_file(
-    main_path: str | Path, extension: tuple[str | Path] = (".txt",)
-) -> list[Path]:
+def find_file(path: str | Path, extension: tuple[str] = ('.raw.kwd',)) -> list[Path]:
     """
-    This function finds all the file types specified by 'extension' in the 'main_path' directory
-    and all its subdirectories and their sub-subdirectories etc.,
-    and returns a list of all file paths
-    'extension' is a list of desired file extensions: ['.dat','.prm']
+    Recursively finds files with the specified extensions within the given path.
+    `path` (str or Path): The directory in which to search for files.
+    `extension`: A tuple of file extensions e.g., ('.dat', '.prm').
     """
-    if isinstance(main_path, str):
-        path = Path(main_path)
-    else:
-        path = main_path
+    p = Path(path)
+    if not p.exists():
+        raise FileNotFoundError(f"Path does not exist: {p}")
 
+    # Convert extension to list if it is a string.
     if isinstance(extension, str):
-        extension = extension.split()  # turning extension into a list with a single element
-
-    return [
-        Path(walking[0] / goodfile)
-        for walking in path.walk()
-        for goodfile in walking[2]
-        for ext in extension
-        if goodfile.endswith(ext)
-    ]
+        extension = extension.split()
+
+    # Normalize extensions to ensure they start with a dot.
+    normalized_exts = [ext if ext.startswith('.') else '.' + ext for ext in extension]
+
+    found_files = []
+    for ext in normalized_exts:
+        for file in p.rglob(f"*{ext}"):
+            if file.is_file():
+                found_files.append(file)
+    return found_files
 
 
 def list_dirs(main_path: str | Path) -> list[str]:
diff --git a/bnd/pipeline/kilosort.py b/bnd/pipeline/kilosort.py
index f765093..38aeb58 100644
--- a/bnd/pipeline/kilosort.py
+++ b/bnd/pipeline/kilosort.py
@@ -1,4 +1,6 @@
 from pathlib import Path
+from configparser import ConfigParser
+import os
 
 import torch
 from kilosort import run_kilosort
@@ -6,29 +8,57 @@
 
 from bnd import set_logging
 from bnd.config import Config, _load_config
+from ..config import find_file
 
 logger = set_logging(__name__)
 
 
+def read_metadata(filepath: Path) -> dict:
+    """Parse a section-less INI file (eg NPx metadata file) and return a dictionary of key-value pairs."""
+    with open(filepath, 'r') as f:
+        content = f.read()
+        # Inject a dummy section header
+        content_with_section = '[dummy_section]\n' + content
+
+    config = ConfigParser()
+    config.optionxform = str  # disables lowercasing
+    config.read_string(content_with_section)
+
+    return dict(config.items('dummy_section'))
+
+
+def add_entry_to_metadata(filepath: Path, tag: str, value: str) -> None:
+    """
+    Add or update a tag=value entry in the NPx metadata.
+    """
+    with open(filepath, 'a') as f:  # append mode
+        f.write(f"{tag}={value}\n")
+
 def _read_probe_type(meta_file_path: str) -> str:
-    with open(meta_file_path, "r") as meta_file:
-        for line in meta_file:
-            if "imDatPrb_type" in line:
-                _, value = line.strip().split("=")
-                break
-
-        if int(value) == 0:
-            probe_type = (
-                "neuropixPhase3B1_kilosortChanMap.mat"  # Neuropixels Phase3B1 (staggered)
-            )
-        elif int(value) == 21:
-            probe_type = "NP2_kilosortChanMap.mat"
-        else:
-            raise ValueError(
-                "Probe type not recogised. It appears to be different from Npx 1.0 or 2.0"
-            )
+    meta = read_metadata(meta_file_path)
+    probe_type_val = meta["imDatPrb_type"]
+    if int(probe_type_val) == 0:
+        probe_type = (
+            "neuropixPhase3B1_kilosortChanMap.mat"  # Neuropixels Phase3B1 (staggered)
+        )
+    elif int(probe_type_val) == 21:
+        probe_type = "NP2_kilosortChanMap.mat"
+    else:
+        raise ValueError(
+            "Probe type not recogised. It appears to be different from Npx 1.0 or 2.0"
+        )
     return probe_type
 
+def _fix_session_metadata(meta_file_path: Path) -> None:
+    """ to inject `fileSizeBytes` and `fileTimeSecs` if they are missing"""
+    meta = read_metadata(meta_file_path)
+    if "fileSizeBytes" not in meta:
+        datafile_path = find_file(meta_file_path.parent, 'ap.bin')[0]
+        data_size = os.path.getsize(datafile_path)
+        add_entry_to_metadata(meta_file_path, "fileSizeBytes", str(data_size))
+        data_duration = data_size / int(meta['nSavedChans']) / 2 / int(meta["imSampRate"])
+        add_entry_to_metadata(meta_file_path, "fileTimeSecs", str(data_duration))
+        logger.warning(f"Metadata missing values: Injected fileSizeBytes: {data_size} and fileTimeSecs: {data_duration}")
 
 def run_kilosort_on_stream(
     config: Config,
@@ -55,9 +85,10 @@ def run_kilosort_on_stream(
     -------
 
     """
+    meta_file_path = config.get_subdirectories_from_pattern(probe_folder_path, "*ap.meta")[0]
 
     sorter_params = {
-        "n_chan_bin": 385,
+        "n_chan_bin": int(read_metadata(meta_file_path)["nSavedChans"]),
     }
 
     ksort_output_path = (
@@ -76,9 +107,12 @@ def run_kilosort_on_stream(
         # Sometimes the gateway can throw an error so just double check.
         download_probes()
 
+    
+    # Check if the metadata file is complete
+    # when SpikeGLX crashes, metadata misses some values.
+    _fix_session_metadata(meta_file_path)
     # Find out which probe type we have
-    meta_file_path = config.get_subdirectories_from_pattern(probe_folder_path, "*ap.meta")
-    probe_name = _read_probe_type(str(meta_file_path[0]))
+    probe_name = _read_probe_type(meta_file_path)
 
     _ = run_kilosort(
         settings=sorter_params,