New check: B113: TrojanSource - Bidirectional control characters

bandit/core/issue.py

@@ -83,7 +83,7 @@ def __init__(
         ident=None,
         lineno=None,
         test_id="",
-        col_offset=0,
+        col_offset=-1,
     ):
         self.severity = severity
         self.cwe = Cwe(cwe)

bandit/core/node_visitor.py

@@ -296,4 +296,14 @@ def process(self, data):
         """
         f_ast = ast.parse(data)
         self.generic_visit(f_ast)
+        # Run tests that do not require access to the AST,
+        # but only to the whole file source:
+        self.context = {
+            "file_data": self.fdata,
+            "filename": self.fname,
+            "lineno": 0,
+            "linerange": [0, 1],
+            "col_offset": 0,
+        }
+        self.update_scores(self.tester.run_tests(self.context, "File"))
         return self.scores

bandit/core/test_properties.py

@@ -15,7 +15,11 @@ def checks(*args):
     def wrapper(func):
         if not hasattr(func, "_checks"):
             func._checks = []
-        func._checks.extend(utils.check_ast_node(a) for a in args)
+        for arg in args:
+            if arg == "File":
+                func._checks.append("File")
+            else:
+                func._checks.append(utils.check_ast_node(arg))
 
         LOG.debug("checks() decorator executed")
         LOG.debug("  func._checks: %s", func._checks)

bandit/core/tester.py

@@ -43,7 +43,7 @@ def run_tests(self, raw_context, checktype):
         tests = self.testset.get_tests(checktype)
         for test in tests:
             name = test.__name__
-            # execute test with the an instance of the context class
+            # execute test with an instance of the context class
             temp_context = copy.copy(raw_context)
             context = b_context.Context(temp_context)
             try:
@@ -66,7 +66,8 @@ def run_tests(self, raw_context, checktype):
                     if result.lineno is None:
                         result.lineno = temp_context["lineno"]
                     result.linerange = temp_context["linerange"]
-                    result.col_offset = temp_context["col_offset"]
+                    if result.col_offset == -1:
+                        result.col_offset = temp_context["col_offset"]
                     result.test = name
                     if result.test_id == "":
                         result.test_id = test._test_id

bandit/plugins/trojansource.py

@@ -0,0 +1,70 @@
+#
+# SPDX-License-Identifier: Apache-2.0
+r"""
+=====================================================
+B113: TrojanSource - Bidirectional control characters
+=====================================================
+
+This plugin checks for the presence of unicode bidirectional control characters
+in Python source files. Those characters can be embedded in comments and strings
+to reorder source code characters in a way that changes its logic.
+
+:Example:
+
+.. code-block:: none
+
+    >> Issue: [B113:trojansource] A Python source file contains bidirectional control characters ('\u202e').
+       Severity: High   Confidence: Medium
+       Location: examples/trojansource.py:0:0
+
+.. seealso::
+
+ .. [1] https://trojansource.codes/
+ .. [2] https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-42574
+
+.. versionadded:: 1.7.2
+
+"""  # noqa: E501
+from tokenize import detect_encoding
+
+import bandit
+from bandit.core import test_properties as test
+
+
+BIDI_CHARACTERS = (
+    "\u202A",
+    "\u202B",
+    "\u202C",
+    "\u202D",
+    "\u202E",
+    "\u2066",
+    "\u2067",
+    "\u2068",
+    "\u2069",
+    "\u200F",
+)
+
+
+@test.test_id("B113")
+@test.checks("File")
+def trojansource(context):
+    with open(context.filename, "rb") as src_file:
+        encoding, _ = detect_encoding(src_file.readline)
+    with open(context.filename, encoding=encoding) as src_file:
+        for lineno, line in enumerate(src_file.readlines(), start=1):
+            for char in BIDI_CHARACTERS:
+                try:
+                    col_offset = line.index(char) + 1
+                except ValueError:
+                    continue
+                text = (
+                    "A Python source file contains bidirectional"
+                    " control characters (%r)." % char
+                )
+                return bandit.Issue(
+                    severity=bandit.HIGH,
+                    confidence=bandit.MEDIUM,
+                    text=text,
+                    lineno=lineno,
+                    col_offset=col_offset,
+                )

doc/source/plugins/trojansource.rst

@@ -0,0 +1,5 @@
+------------------
+B113: trojansource
+------------------
+
+.. automodule:: bandit.plugins.trojansource

examples/trojansource.py

@@ -0,0 +1,5 @@
+#!/usr/bin/env python3
+# cf. https://trojansource.codes/ & https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-42574
+access_level = "user"
+if access_level != 'none‮⁦': # Check if admin ⁩⁦' and access_level != 'user
+    print("You are an admin.\n")

examples/trojansource_latin1.py

@@ -0,0 +1,7 @@
+#!/usr/bin/env python3
+# -*- coding: latin-1 -*-
+# cf. https://trojansource.codes & https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-42574
+# Some special characters: ������
+access_level = "user"
+if access_level != 'none??': # Check if admin ??' and access_level != 'user
+    print("You are an admin.\n")

setup.cfg

@@ -133,6 +133,9 @@ bandit.plugins =
     snmp_insecure_version = bandit.plugins.snmp_security_check:snmp_insecure_version_check
     snmp_weak_cryptography = bandit.plugins.snmp_security_check:snmp_crypto_check
 
+    # bandit/plugins/trojansource.py
+    trojansource = bandit.plugins.trojansource:trojansource
+
 [build_sphinx]
 all_files = 1
 build-dir = doc/build

tests/functional/test_functional.py

@@ -888,3 +888,17 @@ def test_snmp_security_check(self):
             "CONFIDENCE": {"UNDEFINED": 0, "LOW": 0, "MEDIUM": 0, "HIGH": 3},
         }
         self.check_example("snmp.py", expect)
+
+    def test_trojansource(self):
+        expect = {
+            "SEVERITY": {"UNDEFINED": 0, "LOW": 0, "MEDIUM": 0, "HIGH": 1},
+            "CONFIDENCE": {"UNDEFINED": 0, "LOW": 0, "MEDIUM": 1, "HIGH": 0},
+        }
+        self.check_example("trojansource.py", expect)
+
+    def test_trojansource_latin1(self):
+        expect = {
+            "SEVERITY": {"UNDEFINED": 0, "LOW": 0, "MEDIUM": 0, "HIGH": 0},
+            "CONFIDENCE": {"UNDEFINED": 0, "LOW": 0, "MEDIUM": 0, "HIGH": 0},
+        }
+        self.check_example("trojansource_latin1.py", expect)