Address GHSA-9726-w42j-3qjr

requirements.txt

@@ -1,9 +1,9 @@
 twine==4.0.1
 flake8==5.0.4
-pytest==7.1.3
-pytest-cov==3.0.0
+pytest==8.3.5
+pytest-cov==6.0.0
 requests==2.32.5
-aiohttp==3.13.2
-black==22.8.0
+aiohttp==3.13.3
+black==24.10.0
 numpy>1.24.0,<2.0.0
 py7zr==0.22.0

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = picklescan
-version = 0.0.34
+version = 0.0.35
 author = Matthieu Maitre
 author_email = mmaitre314@users.noreply.github.com
 description = Security scanner detecting Python Pickle files performing suspicious actions

src/picklescan/scanner.py

@@ -125,6 +125,7 @@ def __str__(self) -> str:
     "ctypes": "*",  # Foreign function interface, can load DLLs, call C functions, manipulate raw memory
     "functools": "partial",  # functools.partial(os.system, "echo pwned")
     "httplib": "*",  # Includes http.client.HTTPSConnection()
+    "_io": {"FileIO"},  # io.FileIO is stored as _io.FileIO, can read arbitrary files bypassing builtins.open blocklist
     "numpy.f2py": "*",  # Multiple unsafe functions (e.g., getlincoef, _eval_length) that call eval on arbitrary strings
     "numpy.testing._private.utils": "*",  # runstring() in this module is a synonym for exec()
     "nt": "*",  # Alias for 'os' on Windows. Includes os.system()
@@ -181,6 +182,7 @@ def __str__(self) -> str:
         "basichandlers"
     },  # allows storing a pickle inside a pickle (if this has valid use cases, scan the input bytes instead of flagging the global)
     "trace": {"Trace.run", "Trace.runctx"},
+    "urllib.request": "*",  # urllib.request.urlopen can be used for SSRF and data exfiltration
     "venv": "*",
     "webbrowser": "*",  # Includes webbrowser.open()
 }

tests/init_data_files.py

@@ -355,6 +355,18 @@ def reduce_GHSA_r8g5_cgf2_4m4m():
     return getlincoef, (_payload, [])
 
 
+def reduce_io_FileIO():
+    import io
+
+    return io.FileIO, ("/etc/hosts", "r")
+
+
+def reduce_urllib_request_urlopen():
+    import urllib.request
+
+    return urllib.request.urlopen, ("https://example.invalid",)
+
+
 def initialize_pickle_file(path: str, obj: Any, version: int):
     if os.path.exists(path):
         print(f"File {path} already exists, skipping initialization.")
@@ -781,6 +793,8 @@ def initialize_pickle_files():
     initialize_pickle_file_from_reduce("GHSA-84r2-jw7c-4r5q.pkl", reduce_GHSA_84r2_jw7c_4r5q)
     initialize_pickle_file_from_reduce("GHSA-vqmv-47xg-9wpr.pkl", reduce_GHSA_vqmv_47xg_9wpr)
     initialize_pickle_file_from_reduce("GHSA-r8g5-cgf2-4m4m.pkl", reduce_GHSA_r8g5_cgf2_4m4m)
+    initialize_pickle_file_from_reduce("io_FileIO.pkl", reduce_io_FileIO)
+    initialize_pickle_file_from_reduce("urllib_request_urlopen.pkl", reduce_urllib_request_urlopen)
 
 
 def initialize_numpy_files():

tests/test_scanner.py

@@ -393,6 +393,8 @@ def test_scan_file_path():
         "GHSA-46h3-79wf-xr6c.pkl",
         [Global("_operator", "attrgetter", SafetyLevel.Dangerous), Global("builtins", "__import__", SafetyLevel.Suspicious)],
     )
+    assert_scan("io_FileIO.pkl", [Global("_io", "FileIO", SafetyLevel.Dangerous)])
+    assert_scan("urllib_request_urlopen.pkl", [Global("urllib.request", "urlopen", SafetyLevel.Dangerous)])
 
 
 def test_scan_file_path_npz():