This commit is contained in:
Iliyan Angelov
2025-12-01 06:50:10 +02:00
parent 91f51bc6fe
commit 62c1fe5951
4682 changed files with 544807 additions and 31208 deletions

View File

@@ -0,0 +1,20 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
from importlib import metadata
from bandit.core import config # noqa
from bandit.core import context # noqa
from bandit.core import manager # noqa
from bandit.core import meta_ast # noqa
from bandit.core import node_visitor # noqa
from bandit.core import test_set # noqa
from bandit.core import tester # noqa
from bandit.core import utils # noqa
from bandit.core.constants import * # noqa
from bandit.core.issue import * # noqa
from bandit.core.test_properties import * # noqa
__author__ = metadata.metadata("bandit")["Author"]
__version__ = metadata.version("bandit")

View File

@@ -0,0 +1,17 @@
#!/usr/bin/env python
# SPDX-License-Identifier: Apache-2.0
"""Bandit is a tool designed to find common security issues in Python code.
Bandit is a tool designed to find common security issues in Python code.
To do this Bandit processes each file, builds an AST from it, and runs
appropriate plugins against the AST nodes. Once Bandit has finished
scanning all the files it generates a report.
Bandit was originally developed within the OpenStack Security Project and
later rehomed to PyCQA.
https://bandit.readthedocs.io/
"""
from bandit.cli import main
main.main()

View File

@@ -0,0 +1,670 @@
#
# Copyright 2016 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
r"""
====================================================
Blacklist various Python calls known to be dangerous
====================================================
This blacklist data checks for a number of Python calls known to have possible
security implications. The following blacklist tests are run against any
function calls encountered in the scanned code base, triggered by encountering
ast.Call nodes.
B301: pickle
------------
Pickle and modules that wrap it can be unsafe when used to
deserialize untrusted data, possible security issue.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B301 | pickle | - pickle.loads | Medium |
| | | - pickle.load | |
| | | - pickle.Unpickler | |
| | | - dill.loads | |
| | | - dill.load | |
| | | - dill.Unpickler | |
| | | - shelve.open | |
| | | - shelve.DbfilenameShelf | |
| | | - jsonpickle.decode | |
| | | - jsonpickle.unpickler.decode | |
| | | - jsonpickle.unpickler.Unpickler | |
| | | - pandas.read_pickle | |
+------+---------------------+------------------------------------+-----------+
B302: marshal
-------------
Deserialization with the marshal module is possibly dangerous.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B302 | marshal | - marshal.load | Medium |
| | | - marshal.loads | |
+------+---------------------+------------------------------------+-----------+
B303: md5
---------
Use of insecure MD2, MD4, MD5, or SHA1 hash function.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B303 | md5 | - hashlib.md5 | Medium |
| | | - hashlib.sha1 | |
| | | - Crypto.Hash.MD2.new | |
| | | - Crypto.Hash.MD4.new | |
| | | - Crypto.Hash.MD5.new | |
| | | - Crypto.Hash.SHA.new | |
| | | - Cryptodome.Hash.MD2.new | |
| | | - Cryptodome.Hash.MD4.new | |
| | | - Cryptodome.Hash.MD5.new | |
| | | - Cryptodome.Hash.SHA.new | |
| | | - cryptography.hazmat.primitives | |
| | | .hashes.MD5 | |
| | | - cryptography.hazmat.primitives | |
| | | .hashes.SHA1 | |
+------+---------------------+------------------------------------+-----------+
B304 - B305: ciphers and modes
------------------------------
Use of insecure cipher or cipher mode. Replace with a known secure cipher such
as AES.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B304 | ciphers | - Crypto.Cipher.ARC2.new | High |
| | | - Crypto.Cipher.ARC4.new | |
| | | - Crypto.Cipher.Blowfish.new | |
| | | - Crypto.Cipher.DES.new | |
| | | - Crypto.Cipher.XOR.new | |
| | | - Cryptodome.Cipher.ARC2.new | |
| | | - Cryptodome.Cipher.ARC4.new | |
| | | - Cryptodome.Cipher.Blowfish.new | |
| | | - Cryptodome.Cipher.DES.new | |
| | | - Cryptodome.Cipher.XOR.new | |
| | | - cryptography.hazmat.primitives | |
| | | .ciphers.algorithms.ARC4 | |
| | | - cryptography.hazmat.primitives | |
| | | .ciphers.algorithms.Blowfish | |
| | | - cryptography.hazmat.primitives | |
| | | .ciphers.algorithms.IDEA | |
| | | - cryptography.hazmat.primitives | |
| | | .ciphers.algorithms.CAST5 | |
| | | - cryptography.hazmat.primitives | |
| | | .ciphers.algorithms.SEED | |
| | | - cryptography.hazmat.primitives | |
| | | .ciphers.algorithms.TripleDES | |
+------+---------------------+------------------------------------+-----------+
| B305 | cipher_modes | - cryptography.hazmat.primitives | Medium |
| | | .ciphers.modes.ECB | |
+------+---------------------+------------------------------------+-----------+
B306: mktemp_q
--------------
Use of insecure and deprecated function (mktemp).
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B306 | mktemp_q | - tempfile.mktemp | Medium |
+------+---------------------+------------------------------------+-----------+
B307: eval
----------
Use of possibly insecure function - consider using safer ast.literal_eval.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B307 | eval | - eval | Medium |
+------+---------------------+------------------------------------+-----------+
B308: mark_safe
---------------
Use of mark_safe() may expose cross-site scripting vulnerabilities and should
be reviewed.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B308 | mark_safe | - django.utils.safestring.mark_safe| Medium |
+------+---------------------+------------------------------------+-----------+
B309: httpsconnection
---------------------
The check for this call has been removed.
Use of HTTPSConnection on older versions of Python prior to 2.7.9 and 3.4.3 do
not provide security, see https://wiki.openstack.org/wiki/OSSN/OSSN-0033
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B309 | httpsconnection | - httplib.HTTPSConnection | Medium |
| | | - http.client.HTTPSConnection | |
| | | - six.moves.http_client | |
| | | .HTTPSConnection | |
+------+---------------------+------------------------------------+-----------+
B310: urllib_urlopen
--------------------
Audit url open for permitted schemes. Allowing use of 'file:'' or custom
schemes is often unexpected.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B310 | urllib_urlopen | - urllib.urlopen | Medium |
| | | - urllib.request.urlopen | |
| | | - urllib.urlretrieve | |
| | | - urllib.request.urlretrieve | |
| | | - urllib.URLopener | |
| | | - urllib.request.URLopener | |
| | | - urllib.FancyURLopener | |
| | | - urllib.request.FancyURLopener | |
| | | - urllib2.urlopen | |
| | | - urllib2.Request | |
| | | - six.moves.urllib.request.urlopen | |
| | | - six.moves.urllib.request | |
| | | .urlretrieve | |
| | | - six.moves.urllib.request | |
| | | .URLopener | |
| | | - six.moves.urllib.request | |
| | | .FancyURLopener | |
+------+---------------------+------------------------------------+-----------+
B311: random
------------
Standard pseudo-random generators are not suitable for security/cryptographic
purposes. Consider using the secrets module instead:
https://docs.python.org/library/secrets.html
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B311 | random | - random.Random | Low |
| | | - random.random | |
| | | - random.randrange | |
| | | - random.randint | |
| | | - random.choice | |
| | | - random.choices | |
| | | - random.uniform | |
| | | - random.triangular | |
| | | - random.randbytes | |
| | | - random.randrange | |
| | | - random.sample | |
| | | - random.getrandbits | |
+------+---------------------+------------------------------------+-----------+
B312: telnetlib
---------------
Telnet-related functions are being called. Telnet is considered insecure. Use
SSH or some other encrypted protocol.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B312 | telnetlib | - telnetlib.\* | High |
+------+---------------------+------------------------------------+-----------+
B313 - B319: XML
----------------
Most of this is based off of Christian Heimes' work on defusedxml:
https://pypi.org/project/defusedxml/#defusedxml-sax
Using various XLM methods to parse untrusted XML data is known to be vulnerable
to XML attacks. Methods should be replaced with their defusedxml equivalents.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B313 | xml_bad_cElementTree| - xml.etree.cElementTree.parse | Medium |
| | | - xml.etree.cElementTree.iterparse | |
| | | - xml.etree.cElementTree.fromstring| |
| | | - xml.etree.cElementTree.XMLParser | |
+------+---------------------+------------------------------------+-----------+
| B314 | xml_bad_ElementTree | - xml.etree.ElementTree.parse | Medium |
| | | - xml.etree.ElementTree.iterparse | |
| | | - xml.etree.ElementTree.fromstring | |
| | | - xml.etree.ElementTree.XMLParser | |
+------+---------------------+------------------------------------+-----------+
| B315 | xml_bad_expatreader | - xml.sax.expatreader.create_parser| Medium |
+------+---------------------+------------------------------------+-----------+
| B316 | xml_bad_expatbuilder| - xml.dom.expatbuilder.parse | Medium |
| | | - xml.dom.expatbuilder.parseString | |
+------+---------------------+------------------------------------+-----------+
| B317 | xml_bad_sax | - xml.sax.parse | Medium |
| | | - xml.sax.parseString | |
| | | - xml.sax.make_parser | |
+------+---------------------+------------------------------------+-----------+
| B318 | xml_bad_minidom | - xml.dom.minidom.parse | Medium |
| | | - xml.dom.minidom.parseString | |
+------+---------------------+------------------------------------+-----------+
| B319 | xml_bad_pulldom | - xml.dom.pulldom.parse | Medium |
| | | - xml.dom.pulldom.parseString | |
+------+---------------------+------------------------------------+-----------+
B320: xml_bad_etree
-------------------
The check for this call has been removed.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B320 | xml_bad_etree | - lxml.etree.parse | Medium |
| | | - lxml.etree.fromstring | |
| | | - lxml.etree.RestrictedElement | |
| | | - lxml.etree.GlobalParserTLS | |
| | | - lxml.etree.getDefaultParser | |
| | | - lxml.etree.check_docinfo | |
+------+---------------------+------------------------------------+-----------+
B321: ftplib
------------
FTP-related functions are being called. FTP is considered insecure. Use
SSH/SFTP/SCP or some other encrypted protocol.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B321 | ftplib | - ftplib.\* | High |
+------+---------------------+------------------------------------+-----------+
B322: input
-----------
The check for this call has been removed.
The input method in Python 2 will read from standard input, evaluate and
run the resulting string as python source code. This is similar, though in
many ways worse, than using eval. On Python 2, use raw_input instead, input
is safe in Python 3.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B322 | input | - input | High |
+------+---------------------+------------------------------------+-----------+
B323: unverified_context
------------------------
By default, Python will create a secure, verified ssl context for use in such
classes as HTTPSConnection. However, it still allows using an insecure
context via the _create_unverified_context that reverts to the previous
behavior that does not validate certificates or perform hostname checks.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B323 | unverified_context | - ssl._create_unverified_context | Medium |
+------+---------------------+------------------------------------+-----------+
B325: tempnam
--------------
The check for this call has been removed.
Use of os.tempnam() and os.tmpnam() is vulnerable to symlink attacks. Consider
using tmpfile() instead.
For further information:
https://docs.python.org/2.7/library/os.html#os.tempnam
https://docs.python.org/3/whatsnew/3.0.html?highlight=tempnam
https://bugs.python.org/issue17880
+------+---------------------+------------------------------------+-----------+
| ID | Name | Calls | Severity |
+======+=====================+====================================+===========+
| B325 | tempnam | - os.tempnam | Medium |
| | | - os.tmpnam | |
+------+---------------------+------------------------------------+-----------+
"""
from bandit.blacklists import utils
from bandit.core import issue
def gen_blacklist():
"""Generate a list of items to blacklist.
Methods of this type, "bandit.blacklist" plugins, are used to build a list
of items that bandit's built in blacklisting tests will use to trigger
issues. They replace the older blacklist* test plugins and allow
blacklisted items to have a unique bandit ID for filtering and profile
usage.
:return: a dictionary mapping node types to a list of blacklist data
"""
sets = []
sets.append(
utils.build_conf_dict(
"pickle",
"B301",
issue.Cwe.DESERIALIZATION_OF_UNTRUSTED_DATA,
[
"pickle.loads",
"pickle.load",
"pickle.Unpickler",
"dill.loads",
"dill.load",
"dill.Unpickler",
"shelve.open",
"shelve.DbfilenameShelf",
"jsonpickle.decode",
"jsonpickle.unpickler.decode",
"jsonpickle.unpickler.Unpickler",
"pandas.read_pickle",
],
"Pickle and modules that wrap it can be unsafe when used to "
"deserialize untrusted data, possible security issue.",
)
)
sets.append(
utils.build_conf_dict(
"marshal",
"B302",
issue.Cwe.DESERIALIZATION_OF_UNTRUSTED_DATA,
["marshal.load", "marshal.loads"],
"Deserialization with the marshal module is possibly dangerous.",
)
)
sets.append(
utils.build_conf_dict(
"md5",
"B303",
issue.Cwe.BROKEN_CRYPTO,
[
"Crypto.Hash.MD2.new",
"Crypto.Hash.MD4.new",
"Crypto.Hash.MD5.new",
"Crypto.Hash.SHA.new",
"Cryptodome.Hash.MD2.new",
"Cryptodome.Hash.MD4.new",
"Cryptodome.Hash.MD5.new",
"Cryptodome.Hash.SHA.new",
"cryptography.hazmat.primitives.hashes.MD5",
"cryptography.hazmat.primitives.hashes.SHA1",
],
"Use of insecure MD2, MD4, MD5, or SHA1 hash function.",
)
)
sets.append(
utils.build_conf_dict(
"ciphers",
"B304",
issue.Cwe.BROKEN_CRYPTO,
[
"Crypto.Cipher.ARC2.new",
"Crypto.Cipher.ARC4.new",
"Crypto.Cipher.Blowfish.new",
"Crypto.Cipher.DES.new",
"Crypto.Cipher.XOR.new",
"Cryptodome.Cipher.ARC2.new",
"Cryptodome.Cipher.ARC4.new",
"Cryptodome.Cipher.Blowfish.new",
"Cryptodome.Cipher.DES.new",
"Cryptodome.Cipher.XOR.new",
"cryptography.hazmat.primitives.ciphers.algorithms.ARC4",
"cryptography.hazmat.primitives.ciphers.algorithms.Blowfish",
"cryptography.hazmat.primitives.ciphers.algorithms.CAST5",
"cryptography.hazmat.primitives.ciphers.algorithms.IDEA",
"cryptography.hazmat.primitives.ciphers.algorithms.SEED",
"cryptography.hazmat.primitives.ciphers.algorithms.TripleDES",
],
"Use of insecure cipher {name}. Replace with a known secure"
" cipher such as AES.",
"HIGH",
)
)
sets.append(
utils.build_conf_dict(
"cipher_modes",
"B305",
issue.Cwe.BROKEN_CRYPTO,
["cryptography.hazmat.primitives.ciphers.modes.ECB"],
"Use of insecure cipher mode {name}.",
)
)
sets.append(
utils.build_conf_dict(
"mktemp_q",
"B306",
issue.Cwe.INSECURE_TEMP_FILE,
["tempfile.mktemp"],
"Use of insecure and deprecated function (mktemp).",
)
)
sets.append(
utils.build_conf_dict(
"eval",
"B307",
issue.Cwe.OS_COMMAND_INJECTION,
["eval"],
"Use of possibly insecure function - consider using safer "
"ast.literal_eval.",
)
)
sets.append(
utils.build_conf_dict(
"mark_safe",
"B308",
issue.Cwe.XSS,
["django.utils.safestring.mark_safe"],
"Use of mark_safe() may expose cross-site scripting "
"vulnerabilities and should be reviewed.",
)
)
# skipped B309 as the check for a call to httpsconnection has been removed
sets.append(
utils.build_conf_dict(
"urllib_urlopen",
"B310",
issue.Cwe.PATH_TRAVERSAL,
[
"urllib.request.urlopen",
"urllib.request.urlretrieve",
"urllib.request.URLopener",
"urllib.request.FancyURLopener",
"six.moves.urllib.request.urlopen",
"six.moves.urllib.request.urlretrieve",
"six.moves.urllib.request.URLopener",
"six.moves.urllib.request.FancyURLopener",
],
"Audit url open for permitted schemes. Allowing use of file:/ or "
"custom schemes is often unexpected.",
)
)
sets.append(
utils.build_conf_dict(
"random",
"B311",
issue.Cwe.INSUFFICIENT_RANDOM_VALUES,
[
"random.Random",
"random.random",
"random.randrange",
"random.randint",
"random.choice",
"random.choices",
"random.uniform",
"random.triangular",
"random.randbytes",
"random.sample",
"random.randrange",
"random.getrandbits",
],
"Standard pseudo-random generators are not suitable for "
"security/cryptographic purposes.",
"LOW",
)
)
sets.append(
utils.build_conf_dict(
"telnetlib",
"B312",
issue.Cwe.CLEARTEXT_TRANSMISSION,
["telnetlib.Telnet"],
"Telnet-related functions are being called. Telnet is considered "
"insecure. Use SSH or some other encrypted protocol.",
"HIGH",
)
)
# Most of this is based off of Christian Heimes' work on defusedxml:
# https://pypi.org/project/defusedxml/#defusedxml-sax
xml_msg = (
"Using {name} to parse untrusted XML data is known to be "
"vulnerable to XML attacks. Replace {name} with its "
"defusedxml equivalent function or make sure "
"defusedxml.defuse_stdlib() is called"
)
sets.append(
utils.build_conf_dict(
"xml_bad_cElementTree",
"B313",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
[
"xml.etree.cElementTree.parse",
"xml.etree.cElementTree.iterparse",
"xml.etree.cElementTree.fromstring",
"xml.etree.cElementTree.XMLParser",
],
xml_msg,
)
)
sets.append(
utils.build_conf_dict(
"xml_bad_ElementTree",
"B314",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
[
"xml.etree.ElementTree.parse",
"xml.etree.ElementTree.iterparse",
"xml.etree.ElementTree.fromstring",
"xml.etree.ElementTree.XMLParser",
],
xml_msg,
)
)
sets.append(
utils.build_conf_dict(
"xml_bad_expatreader",
"B315",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.sax.expatreader.create_parser"],
xml_msg,
)
)
sets.append(
utils.build_conf_dict(
"xml_bad_expatbuilder",
"B316",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.dom.expatbuilder.parse", "xml.dom.expatbuilder.parseString"],
xml_msg,
)
)
sets.append(
utils.build_conf_dict(
"xml_bad_sax",
"B317",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.sax.parse", "xml.sax.parseString", "xml.sax.make_parser"],
xml_msg,
)
)
sets.append(
utils.build_conf_dict(
"xml_bad_minidom",
"B318",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.dom.minidom.parse", "xml.dom.minidom.parseString"],
xml_msg,
)
)
sets.append(
utils.build_conf_dict(
"xml_bad_pulldom",
"B319",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.dom.pulldom.parse", "xml.dom.pulldom.parseString"],
xml_msg,
)
)
# skipped B320 as the check for a call to lxml.etree has been removed
# end of XML tests
sets.append(
utils.build_conf_dict(
"ftplib",
"B321",
issue.Cwe.CLEARTEXT_TRANSMISSION,
["ftplib.FTP"],
"FTP-related functions are being called. FTP is considered "
"insecure. Use SSH/SFTP/SCP or some other encrypted protocol.",
"HIGH",
)
)
# skipped B322 as the check for a call to input() has been removed
sets.append(
utils.build_conf_dict(
"unverified_context",
"B323",
issue.Cwe.IMPROPER_CERT_VALIDATION,
["ssl._create_unverified_context"],
"By default, Python will create a secure, verified ssl context for"
" use in such classes as HTTPSConnection. However, it still allows"
" using an insecure context via the _create_unverified_context "
"that reverts to the previous behavior that does not validate "
"certificates or perform hostname checks.",
)
)
# skipped B324 (used in bandit/plugins/hashlib_new_insecure_functions.py)
# skipped B325 as the check for a call to os.tempnam and os.tmpnam have
# been removed
return {"Call": sets}

View File

@@ -0,0 +1,425 @@
#
# Copyright 2016 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
r"""
======================================================
Blacklist various Python imports known to be dangerous
======================================================
This blacklist data checks for a number of Python modules known to have
possible security implications. The following blacklist tests are run against
any import statements or calls encountered in the scanned code base.
Note that the XML rules listed here are mostly based off of Christian Heimes'
work on defusedxml: https://pypi.org/project/defusedxml/
B401: import_telnetlib
----------------------
A telnet-related module is being imported. Telnet is considered insecure. Use
SSH or some other encrypted protocol.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B401 | import_telnetlib | - telnetlib | high |
+------+---------------------+------------------------------------+-----------+
B402: import_ftplib
-------------------
A FTP-related module is being imported. FTP is considered insecure. Use
SSH/SFTP/SCP or some other encrypted protocol.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B402 | import_ftplib | - ftplib | high |
+------+---------------------+------------------------------------+-----------+
B403: import_pickle
-------------------
Consider possible security implications associated with these modules.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B403 | import_pickle | - pickle | low |
| | | - cPickle | |
| | | - dill | |
| | | - shelve | |
+------+---------------------+------------------------------------+-----------+
B404: import_subprocess
-----------------------
Consider possible security implications associated with these modules.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B404 | import_subprocess | - subprocess | low |
+------+---------------------+------------------------------------+-----------+
B405: import_xml_etree
----------------------
Using various methods to parse untrusted XML data is known to be vulnerable to
XML attacks. Replace vulnerable imports with the equivalent defusedxml package,
or make sure defusedxml.defuse_stdlib() is called.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B405 | import_xml_etree | - xml.etree.cElementTree | low |
| | | - xml.etree.ElementTree | |
+------+---------------------+------------------------------------+-----------+
B406: import_xml_sax
--------------------
Using various methods to parse untrusted XML data is known to be vulnerable to
XML attacks. Replace vulnerable imports with the equivalent defusedxml package,
or make sure defusedxml.defuse_stdlib() is called.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B406 | import_xml_sax | - xml.sax | low |
+------+---------------------+------------------------------------+-----------+
B407: import_xml_expat
----------------------
Using various methods to parse untrusted XML data is known to be vulnerable to
XML attacks. Replace vulnerable imports with the equivalent defusedxml package,
or make sure defusedxml.defuse_stdlib() is called.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B407 | import_xml_expat | - xml.dom.expatbuilder | low |
+------+---------------------+------------------------------------+-----------+
B408: import_xml_minidom
------------------------
Using various methods to parse untrusted XML data is known to be vulnerable to
XML attacks. Replace vulnerable imports with the equivalent defusedxml package,
or make sure defusedxml.defuse_stdlib() is called.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B408 | import_xml_minidom | - xml.dom.minidom | low |
+------+---------------------+------------------------------------+-----------+
B409: import_xml_pulldom
------------------------
Using various methods to parse untrusted XML data is known to be vulnerable to
XML attacks. Replace vulnerable imports with the equivalent defusedxml package,
or make sure defusedxml.defuse_stdlib() is called.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B409 | import_xml_pulldom | - xml.dom.pulldom | low |
+------+---------------------+------------------------------------+-----------+
B410: import_lxml
-----------------
This import blacklist has been removed. The information here has been
left for historical purposes.
Using various methods to parse untrusted XML data is known to be vulnerable to
XML attacks. Replace vulnerable imports with the equivalent defusedxml package.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B410 | import_lxml | - lxml | low |
+------+---------------------+------------------------------------+-----------+
B411: import_xmlrpclib
----------------------
XMLRPC is particularly dangerous as it is also concerned with communicating
data over a network. Use defusedxml.xmlrpc.monkey_patch() function to
monkey-patch xmlrpclib and mitigate remote XML attacks.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B411 | import_xmlrpclib | - xmlrpc | high |
+------+---------------------+------------------------------------+-----------+
B412: import_httpoxy
--------------------
httpoxy is a set of vulnerabilities that affect application code running in
CGI, or CGI-like environments. The use of CGI for web applications should be
avoided to prevent this class of attack. More details are available
at https://httpoxy.org/.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B412 | import_httpoxy | - wsgiref.handlers.CGIHandler | high |
| | | - twisted.web.twcgi.CGIScript | |
+------+---------------------+------------------------------------+-----------+
B413: import_pycrypto
---------------------
pycrypto library is known to have publicly disclosed buffer overflow
vulnerability https://github.com/dlitz/pycrypto/issues/176. It is no longer
actively maintained and has been deprecated in favor of pyca/cryptography
library.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B413 | import_pycrypto | - Crypto.Cipher | high |
| | | - Crypto.Hash | |
| | | - Crypto.IO | |
| | | - Crypto.Protocol | |
| | | - Crypto.PublicKey | |
| | | - Crypto.Random | |
| | | - Crypto.Signature | |
| | | - Crypto.Util | |
+------+---------------------+------------------------------------+-----------+
B414: import_pycryptodome
-------------------------
This import blacklist has been removed. The information here has been
left for historical purposes.
pycryptodome is a direct fork of pycrypto that has not fully addressed
the issues inherent in PyCrypto. It seems to exist, mainly, as an API
compatible continuation of pycrypto and should be deprecated in favor
of pyca/cryptography which has more support among the Python community.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B414 | import_pycryptodome | - Cryptodome.Cipher | high |
| | | - Cryptodome.Hash | |
| | | - Cryptodome.IO | |
| | | - Cryptodome.Protocol | |
| | | - Cryptodome.PublicKey | |
| | | - Cryptodome.Random | |
| | | - Cryptodome.Signature | |
| | | - Cryptodome.Util | |
+------+---------------------+------------------------------------+-----------+
B415: import_pyghmi
-------------------
An IPMI-related module is being imported. IPMI is considered insecure. Use
an encrypted protocol.
+------+---------------------+------------------------------------+-----------+
| ID | Name | Imports | Severity |
+======+=====================+====================================+===========+
| B415 | import_pyghmi | - pyghmi | high |
+------+---------------------+------------------------------------+-----------+
"""
from bandit.blacklists import utils
from bandit.core import issue
def gen_blacklist():
"""Generate a list of items to blacklist.
Methods of this type, "bandit.blacklist" plugins, are used to build a list
of items that bandit's built in blacklisting tests will use to trigger
issues. They replace the older blacklist* test plugins and allow
blacklisted items to have a unique bandit ID for filtering and profile
usage.
:return: a dictionary mapping node types to a list of blacklist data
"""
sets = []
sets.append(
utils.build_conf_dict(
"import_telnetlib",
"B401",
issue.Cwe.CLEARTEXT_TRANSMISSION,
["telnetlib"],
"A telnet-related module is being imported. Telnet is "
"considered insecure. Use SSH or some other encrypted protocol.",
"HIGH",
)
)
sets.append(
utils.build_conf_dict(
"import_ftplib",
"B402",
issue.Cwe.CLEARTEXT_TRANSMISSION,
["ftplib"],
"A FTP-related module is being imported. FTP is considered "
"insecure. Use SSH/SFTP/SCP or some other encrypted protocol.",
"HIGH",
)
)
sets.append(
utils.build_conf_dict(
"import_pickle",
"B403",
issue.Cwe.DESERIALIZATION_OF_UNTRUSTED_DATA,
["pickle", "cPickle", "dill", "shelve"],
"Consider possible security implications associated with "
"{name} module.",
"LOW",
)
)
sets.append(
utils.build_conf_dict(
"import_subprocess",
"B404",
issue.Cwe.OS_COMMAND_INJECTION,
["subprocess"],
"Consider possible security implications associated with the "
"subprocess module.",
"LOW",
)
)
# Most of this is based off of Christian Heimes' work on defusedxml:
# https://pypi.org/project/defusedxml/#defusedxml-sax
xml_msg = (
"Using {name} to parse untrusted XML data is known to be "
"vulnerable to XML attacks. Replace {name} with the equivalent "
"defusedxml package, or make sure defusedxml.defuse_stdlib() "
"is called."
)
sets.append(
utils.build_conf_dict(
"import_xml_etree",
"B405",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.etree.cElementTree", "xml.etree.ElementTree"],
xml_msg,
"LOW",
)
)
sets.append(
utils.build_conf_dict(
"import_xml_sax",
"B406",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.sax"],
xml_msg,
"LOW",
)
)
sets.append(
utils.build_conf_dict(
"import_xml_expat",
"B407",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.dom.expatbuilder"],
xml_msg,
"LOW",
)
)
sets.append(
utils.build_conf_dict(
"import_xml_minidom",
"B408",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.dom.minidom"],
xml_msg,
"LOW",
)
)
sets.append(
utils.build_conf_dict(
"import_xml_pulldom",
"B409",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xml.dom.pulldom"],
xml_msg,
"LOW",
)
)
# skipped B410 as the check for import_lxml has been removed
sets.append(
utils.build_conf_dict(
"import_xmlrpclib",
"B411",
issue.Cwe.IMPROPER_INPUT_VALIDATION,
["xmlrpc"],
"Using {name} to parse untrusted XML data is known to be "
"vulnerable to XML attacks. Use defusedxml.xmlrpc.monkey_patch() "
"function to monkey-patch xmlrpclib and mitigate XML "
"vulnerabilities.",
"HIGH",
)
)
sets.append(
utils.build_conf_dict(
"import_httpoxy",
"B412",
issue.Cwe.IMPROPER_ACCESS_CONTROL,
[
"wsgiref.handlers.CGIHandler",
"twisted.web.twcgi.CGIScript",
"twisted.web.twcgi.CGIDirectory",
],
"Consider possible security implications associated with "
"{name} module.",
"HIGH",
)
)
sets.append(
utils.build_conf_dict(
"import_pycrypto",
"B413",
issue.Cwe.BROKEN_CRYPTO,
[
"Crypto.Cipher",
"Crypto.Hash",
"Crypto.IO",
"Crypto.Protocol",
"Crypto.PublicKey",
"Crypto.Random",
"Crypto.Signature",
"Crypto.Util",
],
"The pyCrypto library and its module {name} are no longer actively"
" maintained and have been deprecated. "
"Consider using pyca/cryptography library.",
"HIGH",
)
)
sets.append(
utils.build_conf_dict(
"import_pyghmi",
"B415",
issue.Cwe.CLEARTEXT_TRANSMISSION,
["pyghmi"],
"An IPMI-related module is being imported. IPMI is considered "
"insecure. Use an encrypted protocol.",
"HIGH",
)
)
return {"Import": sets, "ImportFrom": sets, "Call": sets}

View File

@@ -0,0 +1,17 @@
#
# Copyright 2016 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
r"""Utils module."""
def build_conf_dict(name, bid, cwe, qualnames, message, level="MEDIUM"):
"""Build and return a blacklist configuration dict."""
return {
"name": name,
"id": bid,
"cwe": cwe,
"message": message,
"qualnames": qualnames,
"level": level,
}

View File

@@ -0,0 +1,252 @@
#
# Copyright 2015 Hewlett-Packard Enterprise
#
# SPDX-License-Identifier: Apache-2.0
# #############################################################################
# Bandit Baseline is a tool that runs Bandit against a Git commit, and compares
# the current commit findings to the parent commit findings.
# To do this it checks out the parent commit, runs Bandit (with any provided
# filters or profiles), checks out the current commit, runs Bandit, and then
# reports on any new findings.
# #############################################################################
"""Bandit is a tool designed to find common security issues in Python code."""
import argparse
import contextlib
import logging
import os
import shutil
import subprocess # nosec: B404
import sys
import tempfile
try:
import git
except ImportError:
git = None
bandit_args = sys.argv[1:]
baseline_tmp_file = "_bandit_baseline_run.json_"
current_commit = None
default_output_format = "terminal"
LOG = logging.getLogger(__name__)
repo = None
report_basename = "bandit_baseline_result"
valid_baseline_formats = ["txt", "html", "json"]
"""baseline.py"""
def main():
"""Execute Bandit."""
# our cleanup function needs this and can't be passed arguments
global current_commit
global repo
parent_commit = None
output_format = None
repo = None
report_fname = None
init_logger()
output_format, repo, report_fname = initialize()
if not repo:
sys.exit(2)
# #################### Find current and parent commits ####################
try:
commit = repo.commit()
current_commit = commit.hexsha
LOG.info("Got current commit: [%s]", commit.name_rev)
commit = commit.parents[0]
parent_commit = commit.hexsha
LOG.info("Got parent commit: [%s]", commit.name_rev)
except git.GitCommandError:
LOG.error("Unable to get current or parent commit")
sys.exit(2)
except IndexError:
LOG.error("Parent commit not available")
sys.exit(2)
# #################### Run Bandit against both commits ####################
output_type = (
["-f", "txt"]
if output_format == default_output_format
else ["-o", report_fname]
)
with baseline_setup() as t:
bandit_tmpfile = f"{t}/{baseline_tmp_file}"
steps = [
{
"message": "Getting Bandit baseline results",
"commit": parent_commit,
"args": bandit_args + ["-f", "json", "-o", bandit_tmpfile],
},
{
"message": "Comparing Bandit results to baseline",
"commit": current_commit,
"args": bandit_args + ["-b", bandit_tmpfile] + output_type,
},
]
return_code = None
for step in steps:
repo.head.reset(commit=step["commit"], working_tree=True)
LOG.info(step["message"])
bandit_command = ["bandit"] + step["args"]
try:
output = subprocess.check_output(bandit_command) # nosec: B603
except subprocess.CalledProcessError as e:
output = e.output
return_code = e.returncode
else:
return_code = 0
output = output.decode("utf-8") # subprocess returns bytes
if return_code not in [0, 1]:
LOG.error(
"Error running command: %s\nOutput: %s\n",
bandit_args,
output,
)
# #################### Output and exit ####################################
# print output or display message about written report
if output_format == default_output_format:
print(output)
else:
LOG.info("Successfully wrote %s", report_fname)
# exit with the code the last Bandit run returned
sys.exit(return_code)
# #################### Clean up before exit ###################################
@contextlib.contextmanager
def baseline_setup():
"""Baseline setup by creating temp folder and resetting repo."""
d = tempfile.mkdtemp()
yield d
shutil.rmtree(d, True)
if repo:
repo.head.reset(commit=current_commit, working_tree=True)
# #################### Setup logging ##########################################
def init_logger():
"""Init logger."""
LOG.handlers = []
log_level = logging.INFO
log_format_string = "[%(levelname)7s ] %(message)s"
logging.captureWarnings(True)
LOG.setLevel(log_level)
handler = logging.StreamHandler(sys.stdout)
handler.setFormatter(logging.Formatter(log_format_string))
LOG.addHandler(handler)
# #################### Perform initialization and validate assumptions ########
def initialize():
"""Initialize arguments and output formats."""
valid = True
# #################### Parse Args #########################################
parser = argparse.ArgumentParser(
description="Bandit Baseline - Generates Bandit results compared to "
"a baseline",
formatter_class=argparse.RawDescriptionHelpFormatter,
epilog="Additional Bandit arguments such as severity filtering (-ll) "
"can be added and will be passed to Bandit.",
)
if sys.version_info >= (3, 14):
parser.suggest_on_error = True
parser.color = False
parser.add_argument(
"targets",
metavar="targets",
type=str,
nargs="+",
help="source file(s) or directory(s) to be tested",
)
parser.add_argument(
"-f",
dest="output_format",
action="store",
default="terminal",
help="specify output format",
choices=valid_baseline_formats,
)
args, _ = parser.parse_known_args()
# #################### Setup Output #######################################
# set the output format, or use a default if not provided
output_format = (
args.output_format if args.output_format else default_output_format
)
if output_format == default_output_format:
LOG.info("No output format specified, using %s", default_output_format)
# set the report name based on the output format
report_fname = f"{report_basename}.{output_format}"
# #################### Check Requirements #################################
if git is None:
LOG.error("Git not available, reinstall with baseline extra")
valid = False
return (None, None, None)
try:
repo = git.Repo(os.getcwd())
except git.exc.InvalidGitRepositoryError:
LOG.error("Bandit baseline must be called from a git project root")
valid = False
except git.exc.GitCommandNotFound:
LOG.error("Git command not found")
valid = False
else:
if repo.is_dirty():
LOG.error(
"Current working directory is dirty and must be " "resolved"
)
valid = False
# if output format is specified, we need to be able to write the report
if output_format != default_output_format and os.path.exists(report_fname):
LOG.error("File %s already exists, aborting", report_fname)
valid = False
# Bandit needs to be able to create this temp file
if os.path.exists(baseline_tmp_file):
LOG.error(
"Temporary file %s needs to be removed prior to running",
baseline_tmp_file,
)
valid = False
# we must validate -o is not provided, as it will mess up Bandit baseline
if "-o" in bandit_args:
LOG.error("Bandit baseline must not be called with the -o option")
valid = False
return (output_format, repo, report_fname) if valid else (None, None, None)
if __name__ == "__main__":
main()

View File

@@ -0,0 +1,207 @@
# Copyright 2015 Red Hat Inc.
#
# SPDX-License-Identifier: Apache-2.0
"""Bandit is a tool designed to find common security issues in Python code."""
import argparse
import importlib
import logging
import os
import sys
import yaml
from bandit.core import extension_loader
PROG_NAME = "bandit_conf_generator"
LOG = logging.getLogger(__name__)
template = """
### Bandit config file generated from:
# '{cli}'
### This config may optionally select a subset of tests to run or skip by
### filling out the 'tests' and 'skips' lists given below. If no tests are
### specified for inclusion then it is assumed all tests are desired. The skips
### set will remove specific tests from the include set. This can be controlled
### using the -t/-s CLI options. Note that the same test ID should not appear
### in both 'tests' and 'skips', this would be nonsensical and is detected by
### Bandit at runtime.
# Available tests:
{test_list}
# (optional) list included test IDs here, eg '[B101, B406]':
{test}
# (optional) list skipped test IDs here, eg '[B101, B406]':
{skip}
### (optional) plugin settings - some test plugins require configuration data
### that may be given here, per-plugin. All bandit test plugins have a built in
### set of sensible defaults and these will be used if no configuration is
### provided. It is not necessary to provide settings for every (or any) plugin
### if the defaults are acceptable.
{settings}
"""
def init_logger():
"""Init logger."""
LOG.handlers = []
log_level = logging.INFO
log_format_string = "[%(levelname)5s]: %(message)s"
logging.captureWarnings(True)
LOG.setLevel(log_level)
handler = logging.StreamHandler(sys.stdout)
handler.setFormatter(logging.Formatter(log_format_string))
LOG.addHandler(handler)
def parse_args():
"""Parse arguments."""
help_description = """Bandit Config Generator
This tool is used to generate an optional profile. The profile may be used
to include or skip tests and override values for plugins.
When used to store an output profile, this tool will output a template that
includes all plugins and their default settings. Any settings which aren't
being overridden can be safely removed from the profile and default values
will be used. Bandit will prefer settings from the profile over the built
in values."""
parser = argparse.ArgumentParser(
description=help_description,
formatter_class=argparse.RawTextHelpFormatter,
)
if sys.version_info >= (3, 14):
parser.suggest_on_error = True
parser.color = False
parser.add_argument(
"--show-defaults",
dest="show_defaults",
action="store_true",
help="show the default settings values for each "
"plugin but do not output a profile",
)
parser.add_argument(
"-o",
"--out",
dest="output_file",
action="store",
help="output file to save profile",
)
parser.add_argument(
"-t",
"--tests",
dest="tests",
action="store",
default=None,
type=str,
help="list of test names to run",
)
parser.add_argument(
"-s",
"--skip",
dest="skips",
action="store",
default=None,
type=str,
help="list of test names to skip",
)
args = parser.parse_args()
if not args.output_file and not args.show_defaults:
parser.print_help()
parser.exit(1)
return args
def get_config_settings():
"""Get configuration settings."""
config = {}
for plugin in extension_loader.MANAGER.plugins:
fn_name = plugin.name
function = plugin.plugin
# if a function takes config...
if hasattr(function, "_takes_config"):
fn_module = importlib.import_module(function.__module__)
# call the config generator if it exists
if hasattr(fn_module, "gen_config"):
config[fn_name] = fn_module.gen_config(function._takes_config)
return yaml.safe_dump(config, default_flow_style=False)
def main():
"""Config generator to write configuration file."""
init_logger()
args = parse_args()
yaml_settings = get_config_settings()
if args.show_defaults:
print(yaml_settings)
if args.output_file:
if os.path.exists(os.path.abspath(args.output_file)):
LOG.error("File %s already exists, exiting", args.output_file)
sys.exit(2)
try:
with open(args.output_file, "w") as f:
skips = args.skips.split(",") if args.skips else []
tests = args.tests.split(",") if args.tests else []
for skip in skips:
if not extension_loader.MANAGER.check_id(skip):
raise RuntimeError(f"unknown ID in skips: {skip}")
for test in tests:
if not extension_loader.MANAGER.check_id(test):
raise RuntimeError(f"unknown ID in tests: {test}")
tpl = "# {0} : {1}"
test_list = [
tpl.format(t.plugin._test_id, t.name)
for t in extension_loader.MANAGER.plugins
]
others = [
tpl.format(k, v["name"])
for k, v in (
extension_loader.MANAGER.blacklist_by_id.items()
)
]
test_list.extend(others)
test_list.sort()
contents = template.format(
cli=" ".join(sys.argv),
settings=yaml_settings,
test_list="\n".join(test_list),
skip="skips: " + str(skips) if skips else "skips:",
test="tests: " + str(tests) if tests else "tests:",
)
f.write(contents)
except OSError:
LOG.error("Unable to open %s for writing", args.output_file)
except Exception as e:
LOG.error("Error: %s", e)
else:
LOG.info("Successfully wrote profile: %s", args.output_file)
return 0
if __name__ == "__main__":
sys.exit(main())

View File

@@ -0,0 +1,701 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
"""Bandit is a tool designed to find common security issues in Python code."""
import argparse
import fnmatch
import logging
import os
import sys
import textwrap
import bandit
from bandit.core import config as b_config
from bandit.core import constants
from bandit.core import manager as b_manager
from bandit.core import utils
BASE_CONFIG = "bandit.yaml"
LOG = logging.getLogger()
def _init_logger(log_level=logging.INFO, log_format=None):
"""Initialize the logger.
:param debug: Whether to enable debug mode
:return: An instantiated logging instance
"""
LOG.handlers = []
if not log_format:
# default log format
log_format_string = constants.log_format_string
else:
log_format_string = log_format
logging.captureWarnings(True)
LOG.setLevel(log_level)
handler = logging.StreamHandler(sys.stderr)
handler.setFormatter(logging.Formatter(log_format_string))
LOG.addHandler(handler)
LOG.debug("logging initialized")
def _get_options_from_ini(ini_path, target):
"""Return a dictionary of config options or None if we can't load any."""
ini_file = None
if ini_path:
ini_file = ini_path
else:
bandit_files = []
for t in target:
for root, _, filenames in os.walk(t):
for filename in fnmatch.filter(filenames, ".bandit"):
bandit_files.append(os.path.join(root, filename))
if len(bandit_files) > 1:
LOG.error(
"Multiple .bandit files found - scan separately or "
"choose one with --ini\n\t%s",
", ".join(bandit_files),
)
sys.exit(2)
elif len(bandit_files) == 1:
ini_file = bandit_files[0]
LOG.info("Found project level .bandit file: %s", bandit_files[0])
if ini_file:
return utils.parse_ini_file(ini_file)
else:
return None
def _init_extensions():
from bandit.core import extension_loader as ext_loader
return ext_loader.MANAGER
def _log_option_source(default_val, arg_val, ini_val, option_name):
"""It's useful to show the source of each option."""
# When default value is not defined, arg_val and ini_val is deterministic
if default_val is None:
if arg_val:
LOG.info("Using command line arg for %s", option_name)
return arg_val
elif ini_val:
LOG.info("Using ini file for %s", option_name)
return ini_val
else:
return None
# No value passed to commad line and default value is used
elif default_val == arg_val:
return ini_val if ini_val else arg_val
# Certainly a value is passed to commad line
else:
return arg_val
def _running_under_virtualenv():
if hasattr(sys, "real_prefix"):
return True
elif sys.prefix != getattr(sys, "base_prefix", sys.prefix):
return True
def _get_profile(config, profile_name, config_path):
profile = {}
if profile_name:
profiles = config.get_option("profiles") or {}
profile = profiles.get(profile_name)
if profile is None:
raise utils.ProfileNotFound(config_path, profile_name)
LOG.debug("read in legacy profile '%s': %s", profile_name, profile)
else:
profile["include"] = set(config.get_option("tests") or [])
profile["exclude"] = set(config.get_option("skips") or [])
return profile
def _log_info(args, profile):
inc = ",".join([t for t in profile["include"]]) or "None"
exc = ",".join([t for t in profile["exclude"]]) or "None"
LOG.info("profile include tests: %s", inc)
LOG.info("profile exclude tests: %s", exc)
LOG.info("cli include tests: %s", args.tests)
LOG.info("cli exclude tests: %s", args.skips)
def main():
"""Bandit CLI."""
# bring our logging stuff up as early as possible
debug = (
logging.DEBUG
if "-d" in sys.argv or "--debug" in sys.argv
else logging.INFO
)
_init_logger(debug)
extension_mgr = _init_extensions()
baseline_formatters = [
f.name
for f in filter(
lambda x: hasattr(x.plugin, "_accepts_baseline"),
extension_mgr.formatters,
)
]
# now do normal startup
parser = argparse.ArgumentParser(
description="Bandit - a Python source code security analyzer",
formatter_class=argparse.RawDescriptionHelpFormatter,
)
if sys.version_info >= (3, 14):
parser.suggest_on_error = True
parser.color = False
parser.add_argument(
"targets",
metavar="targets",
type=str,
nargs="*",
help="source file(s) or directory(s) to be tested",
)
parser.add_argument(
"-r",
"--recursive",
dest="recursive",
action="store_true",
help="find and process files in subdirectories",
)
parser.add_argument(
"-a",
"--aggregate",
dest="agg_type",
action="store",
default="file",
type=str,
choices=["file", "vuln"],
help="aggregate output by vulnerability (default) or by filename",
)
parser.add_argument(
"-n",
"--number",
dest="context_lines",
action="store",
default=3,
type=int,
help="maximum number of code lines to output for each issue",
)
parser.add_argument(
"-c",
"--configfile",
dest="config_file",
action="store",
default=None,
type=str,
help="optional config file to use for selecting plugins and "
"overriding defaults",
)
parser.add_argument(
"-p",
"--profile",
dest="profile",
action="store",
default=None,
type=str,
help="profile to use (defaults to executing all tests)",
)
parser.add_argument(
"-t",
"--tests",
dest="tests",
action="store",
default=None,
type=str,
help="comma-separated list of test IDs to run",
)
parser.add_argument(
"-s",
"--skip",
dest="skips",
action="store",
default=None,
type=str,
help="comma-separated list of test IDs to skip",
)
severity_group = parser.add_mutually_exclusive_group(required=False)
severity_group.add_argument(
"-l",
"--level",
dest="severity",
action="count",
default=1,
help="report only issues of a given severity level or "
"higher (-l for LOW, -ll for MEDIUM, -lll for HIGH)",
)
severity_group.add_argument(
"--severity-level",
dest="severity_string",
action="store",
help="report only issues of a given severity level or higher."
' "all" and "low" are likely to produce the same results, but it'
" is possible for rules to be undefined which will"
' not be listed in "low".',
choices=["all", "low", "medium", "high"],
)
confidence_group = parser.add_mutually_exclusive_group(required=False)
confidence_group.add_argument(
"-i",
"--confidence",
dest="confidence",
action="count",
default=1,
help="report only issues of a given confidence level or "
"higher (-i for LOW, -ii for MEDIUM, -iii for HIGH)",
)
confidence_group.add_argument(
"--confidence-level",
dest="confidence_string",
action="store",
help="report only issues of a given confidence level or higher."
' "all" and "low" are likely to produce the same results, but it'
" is possible for rules to be undefined which will"
' not be listed in "low".',
choices=["all", "low", "medium", "high"],
)
output_format = (
"screen"
if (
sys.stdout.isatty()
and os.getenv("NO_COLOR") is None
and os.getenv("TERM") != "dumb"
)
else "txt"
)
parser.add_argument(
"-f",
"--format",
dest="output_format",
action="store",
default=output_format,
help="specify output format",
choices=sorted(extension_mgr.formatter_names),
)
parser.add_argument(
"--msg-template",
action="store",
default=None,
help="specify output message template"
" (only usable with --format custom),"
" see CUSTOM FORMAT section"
" for list of available values",
)
parser.add_argument(
"-o",
"--output",
dest="output_file",
action="store",
nargs="?",
type=argparse.FileType("w", encoding="utf-8"),
default=sys.stdout,
help="write report to filename",
)
group = parser.add_mutually_exclusive_group(required=False)
group.add_argument(
"-v",
"--verbose",
dest="verbose",
action="store_true",
help="output extra information like excluded and included files",
)
parser.add_argument(
"-d",
"--debug",
dest="debug",
action="store_true",
help="turn on debug mode",
)
group.add_argument(
"-q",
"--quiet",
"--silent",
dest="quiet",
action="store_true",
help="only show output in the case of an error",
)
parser.add_argument(
"--ignore-nosec",
dest="ignore_nosec",
action="store_true",
help="do not skip lines with # nosec comments",
)
parser.add_argument(
"-x",
"--exclude",
dest="excluded_paths",
action="store",
default=",".join(constants.EXCLUDE),
help="comma-separated list of paths (glob patterns "
"supported) to exclude from scan "
"(note that these are in addition to the excluded "
"paths provided in the config file) (default: "
+ ",".join(constants.EXCLUDE)
+ ")",
)
parser.add_argument(
"-b",
"--baseline",
dest="baseline",
action="store",
default=None,
help="path of a baseline report to compare against "
"(only JSON-formatted files are accepted)",
)
parser.add_argument(
"--ini",
dest="ini_path",
action="store",
default=None,
help="path to a .bandit file that supplies command line arguments",
)
parser.add_argument(
"--exit-zero",
action="store_true",
dest="exit_zero",
default=False,
help="exit with 0, " "even with results found",
)
python_ver = sys.version.replace("\n", "")
parser.add_argument(
"--version",
action="version",
version=f"%(prog)s {bandit.__version__}\n"
f" python version = {python_ver}",
)
parser.set_defaults(debug=False)
parser.set_defaults(verbose=False)
parser.set_defaults(quiet=False)
parser.set_defaults(ignore_nosec=False)
plugin_info = [
f"{a[0]}\t{a[1].name}" for a in extension_mgr.plugins_by_id.items()
]
blacklist_info = []
for a in extension_mgr.blacklist.items():
for b in a[1]:
blacklist_info.append(f"{b['id']}\t{b['name']}")
plugin_list = "\n\t".join(sorted(set(plugin_info + blacklist_info)))
dedent_text = textwrap.dedent(
"""
CUSTOM FORMATTING
-----------------
Available tags:
{abspath}, {relpath}, {line}, {col}, {test_id},
{severity}, {msg}, {confidence}, {range}
Example usage:
Default template:
bandit -r examples/ --format custom --msg-template \\
"{abspath}:{line}: {test_id}[bandit]: {severity}: {msg}"
Provides same output as:
bandit -r examples/ --format custom
Tags can also be formatted in python string.format() style:
bandit -r examples/ --format custom --msg-template \\
"{relpath:20.20s}: {line:03}: {test_id:^8}: DEFECT: {msg:>20}"
See python documentation for more information about formatting style:
https://docs.python.org/3/library/string.html
The following tests were discovered and loaded:
-----------------------------------------------
"""
)
parser.epilog = dedent_text + f"\t{plugin_list}"
# setup work - parse arguments, and initialize BanditManager
args = parser.parse_args()
# Check if `--msg-template` is not present without custom formatter
if args.output_format != "custom" and args.msg_template is not None:
parser.error("--msg-template can only be used with --format=custom")
# Check if confidence or severity level have been specified with strings
if args.severity_string is not None:
if args.severity_string == "all":
args.severity = 1
elif args.severity_string == "low":
args.severity = 2
elif args.severity_string == "medium":
args.severity = 3
elif args.severity_string == "high":
args.severity = 4
# Other strings will be blocked by argparse
if args.confidence_string is not None:
if args.confidence_string == "all":
args.confidence = 1
elif args.confidence_string == "low":
args.confidence = 2
elif args.confidence_string == "medium":
args.confidence = 3
elif args.confidence_string == "high":
args.confidence = 4
# Other strings will be blocked by argparse
# Handle .bandit files in projects to pass cmdline args from file
ini_options = _get_options_from_ini(args.ini_path, args.targets)
if ini_options:
# prefer command line, then ini file
args.config_file = _log_option_source(
parser.get_default("configfile"),
args.config_file,
ini_options.get("configfile"),
"config file",
)
args.excluded_paths = _log_option_source(
parser.get_default("excluded_paths"),
args.excluded_paths,
ini_options.get("exclude"),
"excluded paths",
)
args.skips = _log_option_source(
parser.get_default("skips"),
args.skips,
ini_options.get("skips"),
"skipped tests",
)
args.tests = _log_option_source(
parser.get_default("tests"),
args.tests,
ini_options.get("tests"),
"selected tests",
)
ini_targets = ini_options.get("targets")
if ini_targets:
ini_targets = ini_targets.split(",")
args.targets = _log_option_source(
parser.get_default("targets"),
args.targets,
ini_targets,
"selected targets",
)
# TODO(tmcpeak): any other useful options to pass from .bandit?
args.recursive = _log_option_source(
parser.get_default("recursive"),
args.recursive,
ini_options.get("recursive"),
"recursive scan",
)
args.agg_type = _log_option_source(
parser.get_default("agg_type"),
args.agg_type,
ini_options.get("aggregate"),
"aggregate output type",
)
args.context_lines = _log_option_source(
parser.get_default("context_lines"),
args.context_lines,
int(ini_options.get("number") or 0) or None,
"max code lines output for issue",
)
args.profile = _log_option_source(
parser.get_default("profile"),
args.profile,
ini_options.get("profile"),
"profile",
)
args.severity = _log_option_source(
parser.get_default("severity"),
args.severity,
ini_options.get("level"),
"severity level",
)
args.confidence = _log_option_source(
parser.get_default("confidence"),
args.confidence,
ini_options.get("confidence"),
"confidence level",
)
args.output_format = _log_option_source(
parser.get_default("output_format"),
args.output_format,
ini_options.get("format"),
"output format",
)
args.msg_template = _log_option_source(
parser.get_default("msg_template"),
args.msg_template,
ini_options.get("msg-template"),
"output message template",
)
args.output_file = _log_option_source(
parser.get_default("output_file"),
args.output_file,
ini_options.get("output"),
"output file",
)
args.verbose = _log_option_source(
parser.get_default("verbose"),
args.verbose,
ini_options.get("verbose"),
"output extra information",
)
args.debug = _log_option_source(
parser.get_default("debug"),
args.debug,
ini_options.get("debug"),
"debug mode",
)
args.quiet = _log_option_source(
parser.get_default("quiet"),
args.quiet,
ini_options.get("quiet"),
"silent mode",
)
args.ignore_nosec = _log_option_source(
parser.get_default("ignore_nosec"),
args.ignore_nosec,
ini_options.get("ignore-nosec"),
"do not skip lines with # nosec",
)
args.baseline = _log_option_source(
parser.get_default("baseline"),
args.baseline,
ini_options.get("baseline"),
"path of a baseline report",
)
try:
b_conf = b_config.BanditConfig(config_file=args.config_file)
except utils.ConfigError as e:
LOG.error(e)
sys.exit(2)
if not args.targets:
parser.print_usage()
sys.exit(2)
# if the log format string was set in the options, reinitialize
if b_conf.get_option("log_format"):
log_format = b_conf.get_option("log_format")
_init_logger(log_level=logging.DEBUG, log_format=log_format)
if args.quiet:
_init_logger(log_level=logging.WARN)
try:
profile = _get_profile(b_conf, args.profile, args.config_file)
_log_info(args, profile)
profile["include"].update(args.tests.split(",") if args.tests else [])
profile["exclude"].update(args.skips.split(",") if args.skips else [])
extension_mgr.validate_profile(profile)
except (utils.ProfileNotFound, ValueError) as e:
LOG.error(e)
sys.exit(2)
b_mgr = b_manager.BanditManager(
b_conf,
args.agg_type,
args.debug,
profile=profile,
verbose=args.verbose,
quiet=args.quiet,
ignore_nosec=args.ignore_nosec,
)
if args.baseline is not None:
try:
with open(args.baseline) as bl:
data = bl.read()
b_mgr.populate_baseline(data)
except OSError:
LOG.warning("Could not open baseline report: %s", args.baseline)
sys.exit(2)
if args.output_format not in baseline_formatters:
LOG.warning(
"Baseline must be used with one of the following "
"formats: " + str(baseline_formatters)
)
sys.exit(2)
if args.output_format != "json":
if args.config_file:
LOG.info("using config: %s", args.config_file)
LOG.info(
"running on Python %d.%d.%d",
sys.version_info.major,
sys.version_info.minor,
sys.version_info.micro,
)
# initiate file discovery step within Bandit Manager
b_mgr.discover_files(args.targets, args.recursive, args.excluded_paths)
if not b_mgr.b_ts.tests:
LOG.error("No tests would be run, please check the profile.")
sys.exit(2)
# initiate execution of tests within Bandit Manager
b_mgr.run_tests()
LOG.debug(b_mgr.b_ma)
LOG.debug(b_mgr.metrics)
# trigger output of results by Bandit Manager
sev_level = constants.RANKING[args.severity - 1]
conf_level = constants.RANKING[args.confidence - 1]
b_mgr.output_results(
args.context_lines,
sev_level,
conf_level,
args.output_file,
args.output_format,
args.msg_template,
)
if (
b_mgr.results_count(sev_filter=sev_level, conf_filter=conf_level) > 0
and not args.exit_zero
):
sys.exit(1)
else:
sys.exit(0)
if __name__ == "__main__":
main()

View File

@@ -0,0 +1,15 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
from bandit.core import config # noqa
from bandit.core import context # noqa
from bandit.core import manager # noqa
from bandit.core import meta_ast # noqa
from bandit.core import node_visitor # noqa
from bandit.core import test_set # noqa
from bandit.core import tester # noqa
from bandit.core import utils # noqa
from bandit.core.constants import * # noqa
from bandit.core.issue import * # noqa
from bandit.core.test_properties import * # noqa

View File

@@ -0,0 +1,72 @@
#
# Copyright 2016 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import ast
from bandit.core import issue
def report_issue(check, name):
return issue.Issue(
severity=check.get("level", "MEDIUM"),
confidence="HIGH",
cwe=check.get("cwe", issue.Cwe.NOTSET),
text=check["message"].replace("{name}", name),
ident=name,
test_id=check.get("id", "LEGACY"),
)
def blacklist(context, config):
"""Generic blacklist test, B001.
This generic blacklist test will be called for any encountered node with
defined blacklist data available. This data is loaded via plugins using
the 'bandit.blacklists' entry point. Please see the documentation for more
details. Each blacklist datum has a unique bandit ID that may be used for
filtering purposes, or alternatively all blacklisting can be filtered using
the id of this built in test, 'B001'.
"""
blacklists = config
node_type = context.node.__class__.__name__
if node_type == "Call":
func = context.node.func
if isinstance(func, ast.Name) and func.id == "__import__":
if len(context.node.args):
if isinstance(
context.node.args[0], ast.Constant
) and isinstance(context.node.args[0].value, str):
name = context.node.args[0].value
else:
# TODO(??): import through a variable, need symbol tab
name = "UNKNOWN"
else:
name = "" # handle '__import__()'
else:
name = context.call_function_name_qual
# In the case the Call is an importlib.import, treat the first
# argument name as an actual import module name.
# Will produce None if argument is not a literal or identifier
if name in ["importlib.import_module", "importlib.__import__"]:
if context.call_args_count > 0:
name = context.call_args[0]
else:
name = context.call_keywords["name"]
for check in blacklists[node_type]:
for qn in check["qualnames"]:
if name is not None and name == qn:
return report_issue(check, name)
if node_type.startswith("Import"):
prefix = ""
if node_type == "ImportFrom":
if context.node.module is not None:
prefix = context.node.module + "."
for check in blacklists[node_type]:
for name in context.node.names:
for qn in check["qualnames"]:
if (prefix + name.name).startswith(qn):
return report_issue(check, name.name)

View File

@@ -0,0 +1,271 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import logging
import sys
import yaml
if sys.version_info >= (3, 11):
import tomllib
else:
try:
import tomli as tomllib
except ImportError:
tomllib = None
from bandit.core import constants
from bandit.core import extension_loader
from bandit.core import utils
LOG = logging.getLogger(__name__)
class BanditConfig:
def __init__(self, config_file=None):
"""Attempt to initialize a config dictionary from a yaml file.
Error out if loading the yaml file fails for any reason.
:param config_file: The Bandit yaml config file
:raises bandit.utils.ConfigError: If the config is invalid or
unreadable.
"""
self.config_file = config_file
self._config = {}
if config_file:
try:
f = open(config_file, "rb")
except OSError:
raise utils.ConfigError(
"Could not read config file.", config_file
)
if config_file.endswith(".toml"):
if tomllib is None:
raise utils.ConfigError(
"toml parser not available, reinstall with toml extra",
config_file,
)
try:
with f:
self._config = (
tomllib.load(f).get("tool", {}).get("bandit", {})
)
except tomllib.TOMLDecodeError as err:
LOG.error(err)
raise utils.ConfigError("Error parsing file.", config_file)
else:
try:
with f:
self._config = yaml.safe_load(f)
except yaml.YAMLError as err:
LOG.error(err)
raise utils.ConfigError("Error parsing file.", config_file)
self.validate(config_file)
# valid config must be a dict
if not isinstance(self._config, dict):
raise utils.ConfigError("Error parsing file.", config_file)
self.convert_legacy_config()
else:
# use sane defaults
self._config["plugin_name_pattern"] = "*.py"
self._config["include"] = ["*.py", "*.pyw"]
self._init_settings()
def get_option(self, option_string):
"""Returns the option from the config specified by the option_string.
'.' can be used to denote levels, for example to retrieve the options
from the 'a' profile you can use 'profiles.a'
:param option_string: The string specifying the option to retrieve
:return: The object specified by the option_string, or None if it can't
be found.
"""
option_levels = option_string.split(".")
cur_item = self._config
for level in option_levels:
if cur_item and (level in cur_item):
cur_item = cur_item[level]
else:
return None
return cur_item
def get_setting(self, setting_name):
if setting_name in self._settings:
return self._settings[setting_name]
else:
return None
@property
def config(self):
"""Property to return the config dictionary
:return: Config dictionary
"""
return self._config
def _init_settings(self):
"""This function calls a set of other functions (one per setting)
This function calls a set of other functions (one per setting) to build
out the _settings dictionary. Each other function will set values from
the config (if set), otherwise use defaults (from constants if
possible).
:return: -
"""
self._settings = {}
self._init_plugin_name_pattern()
def _init_plugin_name_pattern(self):
"""Sets settings['plugin_name_pattern'] from default or config file."""
plugin_name_pattern = constants.plugin_name_pattern
if self.get_option("plugin_name_pattern"):
plugin_name_pattern = self.get_option("plugin_name_pattern")
self._settings["plugin_name_pattern"] = plugin_name_pattern
def convert_legacy_config(self):
updated_profiles = self.convert_names_to_ids()
bad_calls, bad_imports = self.convert_legacy_blacklist_data()
if updated_profiles:
self.convert_legacy_blacklist_tests(
updated_profiles, bad_calls, bad_imports
)
self._config["profiles"] = updated_profiles
def convert_names_to_ids(self):
"""Convert test names to IDs, unknown names are left unchanged."""
extman = extension_loader.MANAGER
updated_profiles = {}
for name, profile in (self.get_option("profiles") or {}).items():
# NOTE(tkelsey): can't use default of get() because value is
# sometimes explicitly 'None', for example when the list is given
# in yaml but not populated with any values.
include = {
(extman.get_test_id(i) or i)
for i in (profile.get("include") or [])
}
exclude = {
(extman.get_test_id(i) or i)
for i in (profile.get("exclude") or [])
}
updated_profiles[name] = {"include": include, "exclude": exclude}
return updated_profiles
def convert_legacy_blacklist_data(self):
"""Detect legacy blacklist data and convert it to new format."""
bad_calls_list = []
bad_imports_list = []
bad_calls = self.get_option("blacklist_calls") or {}
bad_calls = bad_calls.get("bad_name_sets", {})
for item in bad_calls:
for key, val in item.items():
val["name"] = key
val["message"] = val["message"].replace("{func}", "{name}")
bad_calls_list.append(val)
bad_imports = self.get_option("blacklist_imports") or {}
bad_imports = bad_imports.get("bad_import_sets", {})
for item in bad_imports:
for key, val in item.items():
val["name"] = key
val["message"] = val["message"].replace("{module}", "{name}")
val["qualnames"] = val["imports"]
del val["imports"]
bad_imports_list.append(val)
if bad_imports_list or bad_calls_list:
LOG.warning(
"Legacy blacklist data found in config, overriding "
"data plugins"
)
return bad_calls_list, bad_imports_list
@staticmethod
def convert_legacy_blacklist_tests(profiles, bad_imports, bad_calls):
"""Detect old blacklist tests, convert to use new builtin."""
def _clean_set(name, data):
if name in data:
data.remove(name)
data.add("B001")
for name, profile in profiles.items():
blacklist = {}
include = profile["include"]
exclude = profile["exclude"]
name = "blacklist_calls"
if name in include and name not in exclude:
blacklist.setdefault("Call", []).extend(bad_calls)
_clean_set(name, include)
_clean_set(name, exclude)
name = "blacklist_imports"
if name in include and name not in exclude:
blacklist.setdefault("Import", []).extend(bad_imports)
blacklist.setdefault("ImportFrom", []).extend(bad_imports)
blacklist.setdefault("Call", []).extend(bad_imports)
_clean_set(name, include)
_clean_set(name, exclude)
_clean_set("blacklist_import_func", include)
_clean_set("blacklist_import_func", exclude)
# This can happen with a legacy config that includes
# blacklist_calls but exclude blacklist_imports for example
if "B001" in include and "B001" in exclude:
exclude.remove("B001")
profile["blacklist"] = blacklist
def validate(self, path):
"""Validate the config data."""
legacy = False
message = (
"Config file has an include or exclude reference "
"to legacy test '{0}' but no configuration data for "
"it. Configuration data is required for this test. "
"Please consider switching to the new config file "
"format, the tool 'bandit-config-generator' can help "
"you with this."
)
def _test(key, block, exclude, include):
if key in exclude or key in include:
if self._config.get(block) is None:
raise utils.ConfigError(message.format(key), path)
if "profiles" in self._config:
legacy = True
for profile in self._config["profiles"].values():
inc = profile.get("include") or set()
exc = profile.get("exclude") or set()
_test("blacklist_imports", "blacklist_imports", inc, exc)
_test("blacklist_import_func", "blacklist_imports", inc, exc)
_test("blacklist_calls", "blacklist_calls", inc, exc)
# show deprecation message
if legacy:
LOG.warning(
"Config file '%s' contains deprecated legacy config "
"data. Please consider upgrading to the new config "
"format. The tool 'bandit-config-generator' can help "
"you with this. Support for legacy configs will be "
"removed in a future bandit version.",
path,
)

View File

@@ -0,0 +1,40 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
# default plugin name pattern
plugin_name_pattern = "*.py"
RANKING = ["UNDEFINED", "LOW", "MEDIUM", "HIGH"]
RANKING_VALUES = {"UNDEFINED": 1, "LOW": 3, "MEDIUM": 5, "HIGH": 10}
CRITERIA = [("SEVERITY", "UNDEFINED"), ("CONFIDENCE", "UNDEFINED")]
# add each ranking to globals, to allow direct access in module name space
for rank in RANKING:
globals()[rank] = rank
CONFIDENCE_DEFAULT = "UNDEFINED"
# A list of values Python considers to be False.
# These can be useful in tests to check if a value is True or False.
# We don't handle the case of user-defined classes being false.
# These are only useful when we have a constant in code. If we
# have a variable we cannot determine if False.
# See https://docs.python.org/3/library/stdtypes.html#truth-value-testing
FALSE_VALUES = [None, False, "False", 0, 0.0, 0j, "", (), [], {}]
# override with "log_format" option in config file
log_format_string = "[%(module)s]\t%(levelname)s\t%(message)s"
# Directories to exclude by default
EXCLUDE = (
".svn",
"CVS",
".bzr",
".hg",
".git",
"__pycache__",
".tox",
".eggs",
"*.egg",
)

View File

@@ -0,0 +1,316 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import ast
from bandit.core import utils
class Context:
def __init__(self, context_object=None):
"""Initialize the class with a context, empty dict otherwise
:param context_object: The context object to create class from
:return: -
"""
if context_object is not None:
self._context = context_object
else:
self._context = dict()
def __repr__(self):
"""Generate representation of object for printing / interactive use
Most likely only interested in non-default properties, so we return
the string version of _context.
Example string returned:
<Context {'node': <_ast.Call object at 0x110252510>, 'function': None,
'name': 'socket', 'imports': set(['socket']), 'module': None,
'filename': 'examples/binding.py',
'call': <_ast.Call object at 0x110252510>, 'lineno': 3,
'import_aliases': {}, 'qualname': 'socket.socket'}>
:return: A string representation of the object
"""
return f"<Context {self._context}>"
@property
def call_args(self):
"""Get a list of function args
:return: A list of function args
"""
args = []
if "call" in self._context and hasattr(self._context["call"], "args"):
for arg in self._context["call"].args:
if hasattr(arg, "attr"):
args.append(arg.attr)
else:
args.append(self._get_literal_value(arg))
return args
@property
def call_args_count(self):
"""Get the number of args a function call has
:return: The number of args a function call has or None
"""
if "call" in self._context and hasattr(self._context["call"], "args"):
return len(self._context["call"].args)
else:
return None
@property
def call_function_name(self):
"""Get the name (not FQ) of a function call
:return: The name (not FQ) of a function call
"""
return self._context.get("name")
@property
def call_function_name_qual(self):
"""Get the FQ name of a function call
:return: The FQ name of a function call
"""
return self._context.get("qualname")
@property
def call_keywords(self):
"""Get a dictionary of keyword parameters
:return: A dictionary of keyword parameters for a call as strings
"""
if "call" in self._context and hasattr(
self._context["call"], "keywords"
):
return_dict = {}
for li in self._context["call"].keywords:
if hasattr(li.value, "attr"):
return_dict[li.arg] = li.value.attr
else:
return_dict[li.arg] = self._get_literal_value(li.value)
return return_dict
else:
return None
@property
def node(self):
"""Get the raw AST node associated with the context
:return: The raw AST node associated with the context
"""
return self._context.get("node")
@property
def string_val(self):
"""Get the value of a standalone unicode or string object
:return: value of a standalone unicode or string object
"""
return self._context.get("str")
@property
def bytes_val(self):
"""Get the value of a standalone bytes object (py3 only)
:return: value of a standalone bytes object
"""
return self._context.get("bytes")
@property
def string_val_as_escaped_bytes(self):
"""Get escaped value of the object.
Turn the value of a string or bytes object into byte sequence with
unknown, control, and \\ characters escaped.
This function should be used when looking for a known sequence in a
potentially badly encoded string in the code.
:return: sequence of printable ascii bytes representing original string
"""
val = self.string_val
if val is not None:
# it's any of str or unicode in py2, or str in py3
return val.encode("unicode_escape")
val = self.bytes_val
if val is not None:
return utils.escaped_bytes_representation(val)
return None
@property
def statement(self):
"""Get the raw AST for the current statement
:return: The raw AST for the current statement
"""
return self._context.get("statement")
@property
def function_def_defaults_qual(self):
"""Get a list of fully qualified default values in a function def
:return: List of defaults
"""
defaults = []
if (
"node" in self._context
and hasattr(self._context["node"], "args")
and hasattr(self._context["node"].args, "defaults")
):
for default in self._context["node"].args.defaults:
defaults.append(
utils.get_qual_attr(
default, self._context["import_aliases"]
)
)
return defaults
def _get_literal_value(self, literal):
"""Utility function to turn AST literals into native Python types
:param literal: The AST literal to convert
:return: The value of the AST literal
"""
if isinstance(literal, ast.Constant):
if isinstance(literal.value, bool):
literal_value = str(literal.value)
elif literal.value is None:
literal_value = str(literal.value)
else:
literal_value = literal.value
elif isinstance(literal, ast.List):
return_list = list()
for li in literal.elts:
return_list.append(self._get_literal_value(li))
literal_value = return_list
elif isinstance(literal, ast.Tuple):
return_tuple = tuple()
for ti in literal.elts:
return_tuple += (self._get_literal_value(ti),)
literal_value = return_tuple
elif isinstance(literal, ast.Set):
return_set = set()
for si in literal.elts:
return_set.add(self._get_literal_value(si))
literal_value = return_set
elif isinstance(literal, ast.Dict):
literal_value = dict(zip(literal.keys, literal.values))
elif isinstance(literal, ast.Name):
literal_value = literal.id
else:
literal_value = None
return literal_value
def get_call_arg_value(self, argument_name):
"""Gets the value of a named argument in a function call.
:return: named argument value
"""
kwd_values = self.call_keywords
if kwd_values is not None and argument_name in kwd_values:
return kwd_values[argument_name]
def check_call_arg_value(self, argument_name, argument_values=None):
"""Checks for a value of a named argument in a function call.
Returns none if the specified argument is not found.
:param argument_name: A string - name of the argument to look for
:param argument_values: the value, or list of values to test against
:return: Boolean True if argument found and matched, False if
found and not matched, None if argument not found at all
"""
arg_value = self.get_call_arg_value(argument_name)
if arg_value is not None:
if not isinstance(argument_values, list):
# if passed a single value, or a tuple, convert to a list
argument_values = list((argument_values,))
for val in argument_values:
if arg_value == val:
return True
return False
else:
# argument name not found, return None to allow testing for this
# eventuality
return None
def get_lineno_for_call_arg(self, argument_name):
"""Get the line number for a specific named argument
In case the call is split over multiple lines, get the correct one for
the argument.
:param argument_name: A string - name of the argument to look for
:return: Integer - the line number of the found argument, or -1
"""
if hasattr(self.node, "keywords"):
for key in self.node.keywords:
if key.arg == argument_name:
return key.value.lineno
def get_call_arg_at_position(self, position_num):
"""Returns positional argument at the specified position (if it exists)
:param position_num: The index of the argument to return the value for
:return: Value of the argument at the specified position if it exists
"""
max_args = self.call_args_count
if max_args and position_num < max_args:
arg = self._context["call"].args[position_num]
return getattr(arg, "attr", None) or self._get_literal_value(arg)
else:
return None
def is_module_being_imported(self, module):
"""Check for the specified module is currently being imported
:param module: The module name to look for
:return: True if the module is found, False otherwise
"""
return self._context.get("module") == module
def is_module_imported_exact(self, module):
"""Check if a specified module has been imported; only exact matches.
:param module: The module name to look for
:return: True if the module is found, False otherwise
"""
return module in self._context.get("imports", [])
def is_module_imported_like(self, module):
"""Check if a specified module has been imported
Check if a specified module has been imported; specified module exists
as part of any import statement.
:param module: The module name to look for
:return: True if the module is found, False otherwise
"""
if "imports" in self._context:
for imp in self._context["imports"]:
if module in imp:
return True
return False
@property
def filename(self):
return self._context.get("filename")
@property
def file_data(self):
return self._context.get("file_data")
@property
def import_aliases(self):
return self._context.get("import_aliases")

View File

@@ -0,0 +1,54 @@
#
# Copyright 2016 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import bandit
def get_url(bid):
# where our docs are hosted
base_url = f"https://bandit.readthedocs.io/en/{bandit.__version__}/"
# NOTE(tkelsey): for some reason this import can't be found when stevedore
# loads up the formatter plugin that imports this file. It is available
# later though.
from bandit.core import extension_loader
info = extension_loader.MANAGER.plugins_by_id.get(bid)
if info is not None:
return f"{base_url}plugins/{bid.lower()}_{info.plugin.__name__}.html"
info = extension_loader.MANAGER.blacklist_by_id.get(bid)
if info is not None:
template = "blacklists/blacklist_{kind}.html#{id}-{name}"
info["name"] = info["name"].replace("_", "-")
if info["id"].startswith("B3"): # B3XX
# Some of the links are combined, so we have exception cases
if info["id"] in ["B304", "B305"]:
info = info.copy()
info["id"] = "b304-b305"
info["name"] = "ciphers-and-modes"
elif info["id"] in [
"B313",
"B314",
"B315",
"B316",
"B317",
"B318",
"B319",
"B320",
]:
info = info.copy()
info["id"] = "b313-b320"
ext = template.format(
kind="calls", id=info["id"], name=info["name"]
)
else:
ext = template.format(
kind="imports", id=info["id"], name=info["name"]
)
return base_url + ext.lower()
return base_url # no idea, give the docs main page

View File

@@ -0,0 +1,114 @@
#
# SPDX-License-Identifier: Apache-2.0
import logging
import sys
from stevedore import extension
from bandit.core import utils
LOG = logging.getLogger(__name__)
class Manager:
# These IDs are for bandit built in tests
builtin = ["B001"] # Built in blacklist test
def __init__(
self,
formatters_namespace="bandit.formatters",
plugins_namespace="bandit.plugins",
blacklists_namespace="bandit.blacklists",
):
# Cache the extension managers, loaded extensions, and extension names
self.load_formatters(formatters_namespace)
self.load_plugins(plugins_namespace)
self.load_blacklists(blacklists_namespace)
def load_formatters(self, formatters_namespace):
self.formatters_mgr = extension.ExtensionManager(
namespace=formatters_namespace,
invoke_on_load=False,
verify_requirements=False,
)
self.formatters = list(self.formatters_mgr)
self.formatter_names = self.formatters_mgr.names()
def load_plugins(self, plugins_namespace):
self.plugins_mgr = extension.ExtensionManager(
namespace=plugins_namespace,
invoke_on_load=False,
verify_requirements=False,
)
def test_has_id(plugin):
if not hasattr(plugin.plugin, "_test_id"):
# logger not setup yet, so using print
print(
f"WARNING: Test '{plugin.name}' has no ID, skipping.",
file=sys.stderr,
)
return False
return True
self.plugins = list(filter(test_has_id, list(self.plugins_mgr)))
self.plugin_names = [plugin.name for plugin in self.plugins]
self.plugins_by_id = {p.plugin._test_id: p for p in self.plugins}
self.plugins_by_name = {p.name: p for p in self.plugins}
def get_test_id(self, test_name):
if test_name in self.plugins_by_name:
return self.plugins_by_name[test_name].plugin._test_id
if test_name in self.blacklist_by_name:
return self.blacklist_by_name[test_name]["id"]
return None
def load_blacklists(self, blacklist_namespace):
self.blacklists_mgr = extension.ExtensionManager(
namespace=blacklist_namespace,
invoke_on_load=False,
verify_requirements=False,
)
self.blacklist = {}
blacklist = list(self.blacklists_mgr)
for item in blacklist:
for key, val in item.plugin().items():
utils.check_ast_node(key)
self.blacklist.setdefault(key, []).extend(val)
self.blacklist_by_id = {}
self.blacklist_by_name = {}
for val in self.blacklist.values():
for b in val:
self.blacklist_by_id[b["id"]] = b
self.blacklist_by_name[b["name"]] = b
def validate_profile(self, profile):
"""Validate that everything in the configured profiles looks good."""
for inc in profile["include"]:
if not self.check_id(inc):
LOG.warning(f"Unknown test found in profile: {inc}")
for exc in profile["exclude"]:
if not self.check_id(exc):
LOG.warning(f"Unknown test found in profile: {exc}")
union = set(profile["include"]) & set(profile["exclude"])
if len(union) > 0:
raise ValueError(
f"Non-exclusive include/exclude test sets: {union}"
)
def check_id(self, test):
return (
test in self.plugins_by_id
or test in self.blacklist_by_id
or test in self.builtin
)
# Using entry-points and pkg_resources *can* be expensive. So let's load these
# once, store them on the object, and have a module global object for
# accessing them. After the first time this module is imported, it should save
# this attribute on the module and not have to reload the entry-points.
MANAGER = Manager()

View File

@@ -0,0 +1,245 @@
#
# Copyright 2015 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import linecache
from bandit.core import constants
class Cwe:
NOTSET = 0
IMPROPER_INPUT_VALIDATION = 20
PATH_TRAVERSAL = 22
OS_COMMAND_INJECTION = 78
XSS = 79
BASIC_XSS = 80
SQL_INJECTION = 89
CODE_INJECTION = 94
IMPROPER_WILDCARD_NEUTRALIZATION = 155
HARD_CODED_PASSWORD = 259
IMPROPER_ACCESS_CONTROL = 284
IMPROPER_CERT_VALIDATION = 295
CLEARTEXT_TRANSMISSION = 319
INADEQUATE_ENCRYPTION_STRENGTH = 326
BROKEN_CRYPTO = 327
INSUFFICIENT_RANDOM_VALUES = 330
INSECURE_TEMP_FILE = 377
UNCONTROLLED_RESOURCE_CONSUMPTION = 400
DOWNLOAD_OF_CODE_WITHOUT_INTEGRITY_CHECK = 494
DESERIALIZATION_OF_UNTRUSTED_DATA = 502
MULTIPLE_BINDS = 605
IMPROPER_CHECK_OF_EXCEPT_COND = 703
INCORRECT_PERMISSION_ASSIGNMENT = 732
INAPPROPRIATE_ENCODING_FOR_OUTPUT_CONTEXT = 838
MITRE_URL_PATTERN = "https://cwe.mitre.org/data/definitions/%s.html"
def __init__(self, id=NOTSET):
self.id = id
def link(self):
if self.id == Cwe.NOTSET:
return ""
return Cwe.MITRE_URL_PATTERN % str(self.id)
def __str__(self):
if self.id == Cwe.NOTSET:
return ""
return "CWE-%i (%s)" % (self.id, self.link())
def as_dict(self):
return (
{"id": self.id, "link": self.link()}
if self.id != Cwe.NOTSET
else {}
)
def as_jsons(self):
return str(self.as_dict())
def from_dict(self, data):
if "id" in data:
self.id = int(data["id"])
else:
self.id = Cwe.NOTSET
def __eq__(self, other):
return self.id == other.id
def __ne__(self, other):
return self.id != other.id
def __hash__(self):
return id(self)
class Issue:
def __init__(
self,
severity,
cwe=0,
confidence=constants.CONFIDENCE_DEFAULT,
text="",
ident=None,
lineno=None,
test_id="",
col_offset=-1,
end_col_offset=0,
):
self.severity = severity
self.cwe = Cwe(cwe)
self.confidence = confidence
if isinstance(text, bytes):
text = text.decode("utf-8")
self.text = text
self.ident = ident
self.fname = ""
self.fdata = None
self.test = ""
self.test_id = test_id
self.lineno = lineno
self.col_offset = col_offset
self.end_col_offset = end_col_offset
self.linerange = []
def __str__(self):
return (
"Issue: '%s' from %s:%s: CWE: %s, Severity: %s Confidence: "
"%s at %s:%i:%i"
) % (
self.text,
self.test_id,
(self.ident or self.test),
str(self.cwe),
self.severity,
self.confidence,
self.fname,
self.lineno,
self.col_offset,
)
def __eq__(self, other):
# if the issue text, severity, confidence, and filename match, it's
# the same issue from our perspective
match_types = [
"text",
"severity",
"cwe",
"confidence",
"fname",
"test",
"test_id",
]
return all(
getattr(self, field) == getattr(other, field)
for field in match_types
)
def __ne__(self, other):
return not self.__eq__(other)
def __hash__(self):
return id(self)
def filter(self, severity, confidence):
"""Utility to filter on confidence and severity
This function determines whether an issue should be included by
comparing the severity and confidence rating of the issue to minimum
thresholds specified in 'severity' and 'confidence' respectively.
Formatters should call manager.filter_results() directly.
This will return false if either the confidence or severity of the
issue are lower than the given threshold values.
:param severity: Severity threshold
:param confidence: Confidence threshold
:return: True/False depending on whether issue meets threshold
"""
rank = constants.RANKING
return rank.index(self.severity) >= rank.index(
severity
) and rank.index(self.confidence) >= rank.index(confidence)
def get_code(self, max_lines=3, tabbed=False):
"""Gets lines of code from a file the generated this issue.
:param max_lines: Max lines of context to return
:param tabbed: Use tabbing in the output
:return: strings of code
"""
lines = []
max_lines = max(max_lines, 1)
lmin = max(1, self.lineno - max_lines // 2)
lmax = lmin + len(self.linerange) + max_lines - 1
if self.fname == "<stdin>":
self.fdata.seek(0)
for line_num in range(1, lmin):
self.fdata.readline()
tmplt = "%i\t%s" if tabbed else "%i %s"
for line in range(lmin, lmax):
if self.fname == "<stdin>":
text = self.fdata.readline()
else:
text = linecache.getline(self.fname, line)
if isinstance(text, bytes):
text = text.decode("utf-8")
if not len(text):
break
lines.append(tmplt % (line, text))
return "".join(lines)
def as_dict(self, with_code=True, max_lines=3):
"""Convert the issue to a dict of values for outputting."""
out = {
"filename": self.fname,
"test_name": self.test,
"test_id": self.test_id,
"issue_severity": self.severity,
"issue_cwe": self.cwe.as_dict(),
"issue_confidence": self.confidence,
"issue_text": self.text.encode("utf-8").decode("utf-8"),
"line_number": self.lineno,
"line_range": self.linerange,
"col_offset": self.col_offset,
"end_col_offset": self.end_col_offset,
}
if with_code:
out["code"] = self.get_code(max_lines=max_lines)
return out
def from_dict(self, data, with_code=True):
self.code = data["code"]
self.fname = data["filename"]
self.severity = data["issue_severity"]
self.cwe = cwe_from_dict(data["issue_cwe"])
self.confidence = data["issue_confidence"]
self.text = data["issue_text"]
self.test = data["test_name"]
self.test_id = data["test_id"]
self.lineno = data["line_number"]
self.linerange = data["line_range"]
self.col_offset = data.get("col_offset", 0)
self.end_col_offset = data.get("end_col_offset", 0)
def cwe_from_dict(data):
cwe = Cwe()
cwe.from_dict(data)
return cwe
def issue_from_dict(data):
i = Issue(severity=data["issue_severity"])
i.from_dict(data)
return i

View File

@@ -0,0 +1,499 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import collections
import fnmatch
import io
import json
import logging
import os
import re
import sys
import tokenize
import traceback
from rich import progress
from bandit.core import constants as b_constants
from bandit.core import extension_loader
from bandit.core import issue
from bandit.core import meta_ast as b_meta_ast
from bandit.core import metrics
from bandit.core import node_visitor as b_node_visitor
from bandit.core import test_set as b_test_set
LOG = logging.getLogger(__name__)
NOSEC_COMMENT = re.compile(r"#\s*nosec:?\s*(?P<tests>[^#]+)?#?")
NOSEC_COMMENT_TESTS = re.compile(r"(?:(B\d+|[a-z\d_]+),?)+", re.IGNORECASE)
PROGRESS_THRESHOLD = 50
class BanditManager:
scope = []
def __init__(
self,
config,
agg_type,
debug=False,
verbose=False,
quiet=False,
profile=None,
ignore_nosec=False,
):
"""Get logger, config, AST handler, and result store ready
:param config: config options object
:type config: bandit.core.BanditConfig
:param agg_type: aggregation type
:param debug: Whether to show debug messages or not
:param verbose: Whether to show verbose output
:param quiet: Whether to only show output in the case of an error
:param profile_name: Optional name of profile to use (from cmd line)
:param ignore_nosec: Whether to ignore #nosec or not
:return:
"""
self.debug = debug
self.verbose = verbose
self.quiet = quiet
if not profile:
profile = {}
self.ignore_nosec = ignore_nosec
self.b_conf = config
self.files_list = []
self.excluded_files = []
self.b_ma = b_meta_ast.BanditMetaAst()
self.skipped = []
self.results = []
self.baseline = []
self.agg_type = agg_type
self.metrics = metrics.Metrics()
self.b_ts = b_test_set.BanditTestSet(config, profile)
self.scores = []
def get_skipped(self):
ret = []
# "skip" is a tuple of name and reason, decode just the name
for skip in self.skipped:
if isinstance(skip[0], bytes):
ret.append((skip[0].decode("utf-8"), skip[1]))
else:
ret.append(skip)
return ret
def get_issue_list(
self, sev_level=b_constants.LOW, conf_level=b_constants.LOW
):
return self.filter_results(sev_level, conf_level)
def populate_baseline(self, data):
"""Populate a baseline set of issues from a JSON report
This will populate a list of baseline issues discovered from a previous
run of bandit. Later this baseline can be used to filter out the result
set, see filter_results.
"""
items = []
try:
jdata = json.loads(data)
items = [issue.issue_from_dict(j) for j in jdata["results"]]
except Exception as e:
LOG.warning("Failed to load baseline data: %s", e)
self.baseline = items
def filter_results(self, sev_filter, conf_filter):
"""Returns a list of results filtered by the baseline
This works by checking the number of results returned from each file we
process. If the number of results is different to the number reported
for the same file in the baseline, then we return all results for the
file. We can't reliably return just the new results, as line numbers
will likely have changed.
:param sev_filter: severity level filter to apply
:param conf_filter: confidence level filter to apply
"""
results = [
i for i in self.results if i.filter(sev_filter, conf_filter)
]
if not self.baseline:
return results
unmatched = _compare_baseline_results(self.baseline, results)
# if it's a baseline we'll return a dictionary of issues and a list of
# candidate issues
return _find_candidate_matches(unmatched, results)
def results_count(
self, sev_filter=b_constants.LOW, conf_filter=b_constants.LOW
):
"""Return the count of results
:param sev_filter: Severity level to filter lower
:param conf_filter: Confidence level to filter
:return: Number of results in the set
"""
return len(self.get_issue_list(sev_filter, conf_filter))
def output_results(
self,
lines,
sev_level,
conf_level,
output_file,
output_format,
template=None,
):
"""Outputs results from the result store
:param lines: How many surrounding lines to show per result
:param sev_level: Which severity levels to show (LOW, MEDIUM, HIGH)
:param conf_level: Which confidence levels to show (LOW, MEDIUM, HIGH)
:param output_file: File to store results
:param output_format: output format plugin name
:param template: Output template with non-terminal tags <N>
(default: {abspath}:{line}:
{test_id}[bandit]: {severity}: {msg})
:return: -
"""
try:
formatters_mgr = extension_loader.MANAGER.formatters_mgr
if output_format not in formatters_mgr:
output_format = (
"screen"
if (
sys.stdout.isatty()
and os.getenv("NO_COLOR") is None
and os.getenv("TERM") != "dumb"
)
else "txt"
)
formatter = formatters_mgr[output_format]
report_func = formatter.plugin
if output_format == "custom":
report_func(
self,
fileobj=output_file,
sev_level=sev_level,
conf_level=conf_level,
template=template,
)
else:
report_func(
self,
fileobj=output_file,
sev_level=sev_level,
conf_level=conf_level,
lines=lines,
)
except Exception as e:
raise RuntimeError(
f"Unable to output report using "
f"'{output_format}' formatter: {str(e)}"
)
def discover_files(self, targets, recursive=False, excluded_paths=""):
"""Add tests directly and from a directory to the test set
:param targets: The command line list of files and directories
:param recursive: True/False - whether to add all files from dirs
:return:
"""
# We'll mantain a list of files which are added, and ones which have
# been explicitly excluded
files_list = set()
excluded_files = set()
excluded_path_globs = self.b_conf.get_option("exclude_dirs") or []
included_globs = self.b_conf.get_option("include") or ["*.py"]
# if there are command line provided exclusions add them to the list
if excluded_paths:
for path in excluded_paths.split(","):
if os.path.isdir(path):
path = os.path.join(path, "*")
excluded_path_globs.append(path)
# build list of files we will analyze
for fname in targets:
# if this is a directory and recursive is set, find all files
if os.path.isdir(fname):
if recursive:
new_files, newly_excluded = _get_files_from_dir(
fname,
included_globs=included_globs,
excluded_path_strings=excluded_path_globs,
)
files_list.update(new_files)
excluded_files.update(newly_excluded)
else:
LOG.warning(
"Skipping directory (%s), use -r flag to "
"scan contents",
fname,
)
else:
# if the user explicitly mentions a file on command line,
# we'll scan it, regardless of whether it's in the included
# file types list
if _is_file_included(
fname,
included_globs,
excluded_path_globs,
enforce_glob=False,
):
if fname != "-":
fname = os.path.join(".", fname)
files_list.add(fname)
else:
excluded_files.add(fname)
self.files_list = sorted(files_list)
self.excluded_files = sorted(excluded_files)
def run_tests(self):
"""Runs through all files in the scope
:return: -
"""
# if we have problems with a file, we'll remove it from the files_list
# and add it to the skipped list instead
new_files_list = list(self.files_list)
if (
len(self.files_list) > PROGRESS_THRESHOLD
and LOG.getEffectiveLevel() <= logging.INFO
):
files = progress.track(self.files_list)
else:
files = self.files_list
for count, fname in enumerate(files):
LOG.debug("working on file : %s", fname)
try:
if fname == "-":
open_fd = os.fdopen(sys.stdin.fileno(), "rb", 0)
fdata = io.BytesIO(open_fd.read())
new_files_list = [
"<stdin>" if x == "-" else x for x in new_files_list
]
self._parse_file("<stdin>", fdata, new_files_list)
else:
with open(fname, "rb") as fdata:
self._parse_file(fname, fdata, new_files_list)
except OSError as e:
self.skipped.append((fname, e.strerror))
new_files_list.remove(fname)
# reflect any files which may have been skipped
self.files_list = new_files_list
# do final aggregation of metrics
self.metrics.aggregate()
def _parse_file(self, fname, fdata, new_files_list):
try:
# parse the current file
data = fdata.read()
lines = data.splitlines()
self.metrics.begin(fname)
self.metrics.count_locs(lines)
# nosec_lines is a dict of line number -> set of tests to ignore
# for the line
nosec_lines = dict()
try:
fdata.seek(0)
tokens = tokenize.tokenize(fdata.readline)
if not self.ignore_nosec:
for toktype, tokval, (lineno, _), _, _ in tokens:
if toktype == tokenize.COMMENT:
nosec_lines[lineno] = _parse_nosec_comment(tokval)
except tokenize.TokenError:
pass
score = self._execute_ast_visitor(fname, fdata, data, nosec_lines)
self.scores.append(score)
self.metrics.count_issues([score])
except KeyboardInterrupt:
sys.exit(2)
except SyntaxError:
self.skipped.append(
(fname, "syntax error while parsing AST from file")
)
new_files_list.remove(fname)
except Exception as e:
LOG.error(
"Exception occurred when executing tests against %s.", fname
)
if not LOG.isEnabledFor(logging.DEBUG):
LOG.error(
'Run "bandit --debug %s" to see the full traceback.', fname
)
self.skipped.append((fname, "exception while scanning file"))
new_files_list.remove(fname)
LOG.debug(" Exception string: %s", e)
LOG.debug(" Exception traceback: %s", traceback.format_exc())
def _execute_ast_visitor(self, fname, fdata, data, nosec_lines):
"""Execute AST parse on each file
:param fname: The name of the file being parsed
:param data: Original file contents
:param lines: The lines of code to process
:return: The accumulated test score
"""
score = []
res = b_node_visitor.BanditNodeVisitor(
fname,
fdata,
self.b_ma,
self.b_ts,
self.debug,
nosec_lines,
self.metrics,
)
score = res.process(data)
self.results.extend(res.tester.results)
return score
def _get_files_from_dir(
files_dir, included_globs=None, excluded_path_strings=None
):
if not included_globs:
included_globs = ["*.py"]
if not excluded_path_strings:
excluded_path_strings = []
files_list = set()
excluded_files = set()
for root, _, files in os.walk(files_dir):
for filename in files:
path = os.path.join(root, filename)
if _is_file_included(path, included_globs, excluded_path_strings):
files_list.add(path)
else:
excluded_files.add(path)
return files_list, excluded_files
def _is_file_included(
path, included_globs, excluded_path_strings, enforce_glob=True
):
"""Determine if a file should be included based on filename
This utility function determines if a file should be included based
on the file name, a list of parsed extensions, excluded paths, and a flag
specifying whether extensions should be enforced.
:param path: Full path of file to check
:param parsed_extensions: List of parsed extensions
:param excluded_paths: List of paths (globbing supported) from which we
should not include files
:param enforce_glob: Can set to false to bypass extension check
:return: Boolean indicating whether a file should be included
"""
return_value = False
# if this is matches a glob of files we look at, and it isn't in an
# excluded path
if _matches_glob_list(path, included_globs) or not enforce_glob:
if not _matches_glob_list(path, excluded_path_strings) and not any(
x in path for x in excluded_path_strings
):
return_value = True
return return_value
def _matches_glob_list(filename, glob_list):
for glob in glob_list:
if fnmatch.fnmatch(filename, glob):
return True
return False
def _compare_baseline_results(baseline, results):
"""Compare a baseline list of issues to list of results
This function compares a baseline set of issues to a current set of issues
to find results that weren't present in the baseline.
:param baseline: Baseline list of issues
:param results: Current list of issues
:return: List of unmatched issues
"""
return [a for a in results if a not in baseline]
def _find_candidate_matches(unmatched_issues, results_list):
"""Returns a dictionary with issue candidates
For example, let's say we find a new command injection issue in a file
which used to have two. Bandit can't tell which of the command injection
issues in the file are new, so it will show all three. The user should
be able to pick out the new one.
:param unmatched_issues: List of issues that weren't present before
:param results_list: main list of current Bandit findings
:return: A dictionary with a list of candidates for each issue
"""
issue_candidates = collections.OrderedDict()
for unmatched in unmatched_issues:
issue_candidates[unmatched] = [
i for i in results_list if unmatched == i
]
return issue_candidates
def _find_test_id_from_nosec_string(extman, match):
test_id = extman.check_id(match)
if test_id:
return match
# Finding by short_id didn't work, let's check the test name
test_id = extman.get_test_id(match)
if not test_id:
# Name and short id didn't work:
LOG.warning(
"Test in comment: %s is not a test name or id, ignoring", match
)
return test_id # We want to return None or the string here regardless
def _parse_nosec_comment(comment):
found_no_sec_comment = NOSEC_COMMENT.search(comment)
if not found_no_sec_comment:
# there was no nosec comment
return None
matches = found_no_sec_comment.groupdict()
nosec_tests = matches.get("tests", set())
# empty set indicates that there was a nosec comment without specific
# test ids or names
test_ids = set()
if nosec_tests:
extman = extension_loader.MANAGER
# lookup tests by short code or name
for test in NOSEC_COMMENT_TESTS.finditer(nosec_tests):
test_match = test.group(1)
test_id = _find_test_id_from_nosec_string(extman, test_match)
if test_id:
test_ids.add(test_id)
return test_ids

View File

@@ -0,0 +1,44 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import collections
import logging
LOG = logging.getLogger(__name__)
class BanditMetaAst:
nodes = collections.OrderedDict()
def __init__(self):
pass
def add_node(self, node, parent_id, depth):
"""Add a node to the AST node collection
:param node: The AST node to add
:param parent_id: The ID of the node's parent
:param depth: The depth of the node
:return: -
"""
node_id = hex(id(node))
LOG.debug("adding node : %s [%s]", node_id, depth)
self.nodes[node_id] = {
"raw": node,
"parent_id": parent_id,
"depth": depth,
}
def __str__(self):
"""Dumps a listing of all of the nodes
Dumps a listing of all of the nodes for debugging purposes
:return: -
"""
tmpstr = ""
for k, v in self.nodes.items():
tmpstr += f"Node: {k}\n"
tmpstr += f"\t{str(v)}\n"
tmpstr += f"Length: {len(self.nodes)}\n"
return tmpstr

View File

@@ -0,0 +1,106 @@
#
# Copyright 2015 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import collections
from bandit.core import constants
class Metrics:
"""Bandit metric gathering.
This class is a singleton used to gather and process metrics collected when
processing a code base with bandit. Metric collection is stateful, that
is, an active metric block will be set when requested and all subsequent
operations will effect that metric block until it is replaced by a setting
a new one.
"""
def __init__(self):
self.data = dict()
self.data["_totals"] = {
"loc": 0,
"nosec": 0,
"skipped_tests": 0,
}
# initialize 0 totals for criteria and rank; this will be reset later
for rank in constants.RANKING:
for criteria in constants.CRITERIA:
self.data["_totals"][f"{criteria[0]}.{rank}"] = 0
def begin(self, fname):
"""Begin a new metric block.
This starts a new metric collection name "fname" and makes is active.
:param fname: the metrics unique name, normally the file name.
"""
self.data[fname] = {
"loc": 0,
"nosec": 0,
"skipped_tests": 0,
}
self.current = self.data[fname]
def note_nosec(self, num=1):
"""Note a "nosec" comment.
Increment the currently active metrics nosec count.
:param num: number of nosecs seen, defaults to 1
"""
self.current["nosec"] += num
def note_skipped_test(self, num=1):
"""Note a "nosec BXXX, BYYY, ..." comment.
Increment the currently active metrics skipped_tests count.
:param num: number of skipped_tests seen, defaults to 1
"""
self.current["skipped_tests"] += num
def count_locs(self, lines):
"""Count lines of code.
We count lines that are not empty and are not comments. The result is
added to our currently active metrics loc count (normally this is 0).
:param lines: lines in the file to process
"""
def proc(line):
tmp = line.strip()
return bool(tmp and not tmp.startswith(b"#"))
self.current["loc"] += sum(proc(line) for line in lines)
def count_issues(self, scores):
self.current.update(self._get_issue_counts(scores))
def aggregate(self):
"""Do final aggregation of metrics."""
c = collections.Counter()
for fname in self.data:
c.update(self.data[fname])
self.data["_totals"] = dict(c)
@staticmethod
def _get_issue_counts(scores):
"""Get issue counts aggregated by confidence/severity rankings.
:param scores: list of scores to aggregate / count
:return: aggregated total (count) of issues identified
"""
issue_counts = {}
for score in scores:
for criteria, _ in constants.CRITERIA:
for i, rank in enumerate(constants.RANKING):
label = f"{criteria}.{rank}"
if label not in issue_counts:
issue_counts[label] = 0
count = (
score[criteria][i]
// constants.RANKING_VALUES[rank]
)
issue_counts[label] += count
return issue_counts

View File

@@ -0,0 +1,297 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import ast
import logging
import operator
from bandit.core import constants
from bandit.core import tester as b_tester
from bandit.core import utils as b_utils
LOG = logging.getLogger(__name__)
class BanditNodeVisitor:
def __init__(
self, fname, fdata, metaast, testset, debug, nosec_lines, metrics
):
self.debug = debug
self.nosec_lines = nosec_lines
self.scores = {
"SEVERITY": [0] * len(constants.RANKING),
"CONFIDENCE": [0] * len(constants.RANKING),
}
self.depth = 0
self.fname = fname
self.fdata = fdata
self.metaast = metaast
self.testset = testset
self.imports = set()
self.import_aliases = {}
self.tester = b_tester.BanditTester(
self.testset, self.debug, nosec_lines, metrics
)
# in some cases we can't determine a qualified name
try:
self.namespace = b_utils.get_module_qualname_from_path(fname)
except b_utils.InvalidModulePath:
LOG.warning(
"Unable to find qualified name for module: %s", self.fname
)
self.namespace = ""
LOG.debug("Module qualified name: %s", self.namespace)
self.metrics = metrics
def visit_ClassDef(self, node):
"""Visitor for AST ClassDef node
Add class name to current namespace for all descendants.
:param node: Node being inspected
:return: -
"""
# For all child nodes, add this class name to current namespace
self.namespace = b_utils.namespace_path_join(self.namespace, node.name)
def visit_FunctionDef(self, node):
"""Visitor for AST FunctionDef nodes
add relevant information about the node to
the context for use in tests which inspect function definitions.
Add the function name to the current namespace for all descendants.
:param node: The node that is being inspected
:return: -
"""
self.context["function"] = node
qualname = self.namespace + "." + b_utils.get_func_name(node)
name = qualname.split(".")[-1]
self.context["qualname"] = qualname
self.context["name"] = name
# For all child nodes and any tests run, add this function name to
# current namespace
self.namespace = b_utils.namespace_path_join(self.namespace, name)
self.update_scores(self.tester.run_tests(self.context, "FunctionDef"))
def visit_Call(self, node):
"""Visitor for AST Call nodes
add relevant information about the node to
the context for use in tests which inspect function calls.
:param node: The node that is being inspected
:return: -
"""
self.context["call"] = node
qualname = b_utils.get_call_name(node, self.import_aliases)
name = qualname.split(".")[-1]
self.context["qualname"] = qualname
self.context["name"] = name
self.update_scores(self.tester.run_tests(self.context, "Call"))
def visit_Import(self, node):
"""Visitor for AST Import nodes
add relevant information about node to
the context for use in tests which inspect imports.
:param node: The node that is being inspected
:return: -
"""
for nodename in node.names:
if nodename.asname:
self.import_aliases[nodename.asname] = nodename.name
self.imports.add(nodename.name)
self.context["module"] = nodename.name
self.update_scores(self.tester.run_tests(self.context, "Import"))
def visit_ImportFrom(self, node):
"""Visitor for AST ImportFrom nodes
add relevant information about node to
the context for use in tests which inspect imports.
:param node: The node that is being inspected
:return: -
"""
module = node.module
if module is None:
return self.visit_Import(node)
for nodename in node.names:
# TODO(ljfisher) Names in import_aliases could be overridden
# by local definitions. If this occurs bandit will see the
# name in import_aliases instead of the local definition.
# We need better tracking of names.
if nodename.asname:
self.import_aliases[nodename.asname] = (
module + "." + nodename.name
)
else:
# Even if import is not aliased we need an entry that maps
# name to module.name. For example, with 'from a import b'
# b should be aliased to the qualified name a.b
self.import_aliases[nodename.name] = (
module + "." + nodename.name
)
self.imports.add(module + "." + nodename.name)
self.context["module"] = module
self.context["name"] = nodename.name
self.update_scores(self.tester.run_tests(self.context, "ImportFrom"))
def visit_Constant(self, node):
"""Visitor for AST Constant nodes
call the appropriate method for the node type.
this maintains compatibility with <3.6 and 3.8+
This code is heavily influenced by Anthony Sottile (@asottile) here:
https://bugs.python.org/msg342486
:param node: The node that is being inspected
:return: -
"""
if isinstance(node.value, str):
self.visit_Str(node)
elif isinstance(node.value, bytes):
self.visit_Bytes(node)
def visit_Str(self, node):
"""Visitor for AST String nodes
add relevant information about node to
the context for use in tests which inspect strings.
:param node: The node that is being inspected
:return: -
"""
self.context["str"] = node.value
if not isinstance(node._bandit_parent, ast.Expr): # docstring
self.context["linerange"] = b_utils.linerange(node._bandit_parent)
self.update_scores(self.tester.run_tests(self.context, "Str"))
def visit_Bytes(self, node):
"""Visitor for AST Bytes nodes
add relevant information about node to
the context for use in tests which inspect strings.
:param node: The node that is being inspected
:return: -
"""
self.context["bytes"] = node.value
if not isinstance(node._bandit_parent, ast.Expr): # docstring
self.context["linerange"] = b_utils.linerange(node._bandit_parent)
self.update_scores(self.tester.run_tests(self.context, "Bytes"))
def pre_visit(self, node):
self.context = {}
self.context["imports"] = self.imports
self.context["import_aliases"] = self.import_aliases
if self.debug:
LOG.debug(ast.dump(node))
self.metaast.add_node(node, "", self.depth)
if hasattr(node, "lineno"):
self.context["lineno"] = node.lineno
if hasattr(node, "col_offset"):
self.context["col_offset"] = node.col_offset
if hasattr(node, "end_col_offset"):
self.context["end_col_offset"] = node.end_col_offset
self.context["node"] = node
self.context["linerange"] = b_utils.linerange(node)
self.context["filename"] = self.fname
self.context["file_data"] = self.fdata
LOG.debug(
"entering: %s %s [%s]", hex(id(node)), type(node), self.depth
)
self.depth += 1
LOG.debug(self.context)
return True
def visit(self, node):
name = node.__class__.__name__
method = "visit_" + name
visitor = getattr(self, method, None)
if visitor is not None:
if self.debug:
LOG.debug("%s called (%s)", method, ast.dump(node))
visitor(node)
else:
self.update_scores(self.tester.run_tests(self.context, name))
def post_visit(self, node):
self.depth -= 1
LOG.debug("%s\texiting : %s", self.depth, hex(id(node)))
# HACK(tkelsey): this is needed to clean up post-recursion stuff that
# gets setup in the visit methods for these node types.
if isinstance(node, (ast.FunctionDef, ast.ClassDef)):
self.namespace = b_utils.namespace_path_split(self.namespace)[0]
def generic_visit(self, node):
"""Drive the visitor."""
for _, value in ast.iter_fields(node):
if isinstance(value, list):
max_idx = len(value) - 1
for idx, item in enumerate(value):
if isinstance(item, ast.AST):
if idx < max_idx:
item._bandit_sibling = value[idx + 1]
else:
item._bandit_sibling = None
item._bandit_parent = node
if self.pre_visit(item):
self.visit(item)
self.generic_visit(item)
self.post_visit(item)
elif isinstance(value, ast.AST):
value._bandit_sibling = None
value._bandit_parent = node
if self.pre_visit(value):
self.visit(value)
self.generic_visit(value)
self.post_visit(value)
def update_scores(self, scores):
"""Score updater
Since we moved from a single score value to a map of scores per
severity, this is needed to update the stored list.
:param score: The score list to update our scores with
"""
# we'll end up with something like:
# SEVERITY: {0, 0, 0, 10} where 10 is weighted by finding and level
for score_type in self.scores:
self.scores[score_type] = list(
map(operator.add, self.scores[score_type], scores[score_type])
)
def process(self, data):
"""Main process loop
Build and process the AST
:param lines: lines code to process
:return score: the aggregated score for the current file
"""
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

View File

@@ -0,0 +1,83 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import logging
from bandit.core import utils
LOG = logging.getLogger(__name__)
def checks(*args):
"""Decorator function to set checks to be run."""
def wrapper(func):
if not hasattr(func, "_checks"):
func._checks = []
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)
return func
return wrapper
def takes_config(*args):
"""Test function takes config
Use of this delegate before a test function indicates that it should be
passed data from the config file. Passing a name parameter allows
aliasing tests and thus sharing config options.
"""
name = ""
def _takes_config(func):
if not hasattr(func, "_takes_config"):
func._takes_config = name
return func
if len(args) == 1 and callable(args[0]):
name = args[0].__name__
return _takes_config(args[0])
else:
name = args[0]
return _takes_config
def test_id(id_val):
"""Test function identifier
Use this decorator before a test function indicates its simple ID
"""
def _has_id(func):
if not hasattr(func, "_test_id"):
func._test_id = id_val
return func
return _has_id
def accepts_baseline(*args):
"""Decorator to indicate formatter accepts baseline results
Use of this decorator before a formatter indicates that it is able to deal
with baseline results. Specifically this means it has a way to display
candidate results and know when it should do so.
"""
def wrapper(func):
if not hasattr(func, "_accepts_baseline"):
func._accepts_baseline = True
LOG.debug("accepts_baseline() decorator executed on %s", func.__name__)
return func
return wrapper(args[0])

View File

@@ -0,0 +1,114 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import importlib
import logging
from bandit.core import blacklisting
from bandit.core import extension_loader
LOG = logging.getLogger(__name__)
class BanditTestSet:
def __init__(self, config, profile=None):
if not profile:
profile = {}
extman = extension_loader.MANAGER
filtering = self._get_filter(config, profile)
self.plugins = [
p for p in extman.plugins if p.plugin._test_id in filtering
]
self.plugins.extend(self._load_builtins(filtering, profile))
self._load_tests(config, self.plugins)
@staticmethod
def _get_filter(config, profile):
extman = extension_loader.MANAGER
inc = set(profile.get("include", []))
exc = set(profile.get("exclude", []))
all_blacklist_tests = set()
for _, tests in extman.blacklist.items():
all_blacklist_tests.update(t["id"] for t in tests)
# this block is purely for backwards compatibility, the rules are as
# follows:
# B001,B401 means B401
# B401 means B401
# B001 means all blacklist tests
if "B001" in inc:
if not inc.intersection(all_blacklist_tests):
inc.update(all_blacklist_tests)
inc.discard("B001")
if "B001" in exc:
if not exc.intersection(all_blacklist_tests):
exc.update(all_blacklist_tests)
exc.discard("B001")
if inc:
filtered = inc
else:
filtered = set(extman.plugins_by_id.keys())
filtered.update(extman.builtin)
filtered.update(all_blacklist_tests)
return filtered - exc
def _load_builtins(self, filtering, profile):
"""loads up builtin functions, so they can be filtered."""
class Wrapper:
def __init__(self, name, plugin):
self.name = name
self.plugin = plugin
extman = extension_loader.MANAGER
blacklist = profile.get("blacklist")
if not blacklist: # not overridden by legacy data
blacklist = {}
for node, tests in extman.blacklist.items():
values = [t for t in tests if t["id"] in filtering]
if values:
blacklist[node] = values
if not blacklist:
return []
# this dresses up the blacklist to look like a plugin, but
# the '_checks' data comes from the blacklist information.
# the '_config' is the filtered blacklist data set.
blacklisting.blacklist._test_id = "B001"
blacklisting.blacklist._checks = blacklist.keys()
blacklisting.blacklist._config = blacklist
return [Wrapper("blacklist", blacklisting.blacklist)]
def _load_tests(self, config, plugins):
"""Builds a dict mapping tests to node types."""
self.tests = {}
for plugin in plugins:
if hasattr(plugin.plugin, "_takes_config"):
# TODO(??): config could come from profile ...
cfg = config.get_option(plugin.plugin._takes_config)
if cfg is None:
genner = importlib.import_module(plugin.plugin.__module__)
cfg = genner.gen_config(plugin.plugin._takes_config)
plugin.plugin._config = cfg
for check in plugin.plugin._checks:
self.tests.setdefault(check, []).append(plugin.plugin)
LOG.debug(
"added function %s (%s) targeting %s",
plugin.name,
plugin.plugin._test_id,
check,
)
def get_tests(self, checktype):
"""Returns all tests that are of type checktype
:param checktype: The type of test to filter on
:return: A list of tests which are of the specified type
"""
return self.tests.get(checktype) or []

View File

@@ -0,0 +1,166 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import copy
import logging
import warnings
from bandit.core import constants
from bandit.core import context as b_context
from bandit.core import utils
warnings.formatwarning = utils.warnings_formatter
LOG = logging.getLogger(__name__)
class BanditTester:
def __init__(self, testset, debug, nosec_lines, metrics):
self.results = []
self.testset = testset
self.last_result = None
self.debug = debug
self.nosec_lines = nosec_lines
self.metrics = metrics
def run_tests(self, raw_context, checktype):
"""Runs all tests for a certain type of check, for example
Runs all tests for a certain type of check, for example 'functions'
store results in results.
:param raw_context: Raw context dictionary
:param checktype: The type of checks to run
:return: a score based on the number and type of test results with
extra metrics about nosec comments
"""
scores = {
"SEVERITY": [0] * len(constants.RANKING),
"CONFIDENCE": [0] * len(constants.RANKING),
}
tests = self.testset.get_tests(checktype)
for test in tests:
name = test.__name__
# execute test with an instance of the context class
temp_context = copy.copy(raw_context)
context = b_context.Context(temp_context)
try:
if hasattr(test, "_config"):
result = test(context, test._config)
else:
result = test(context)
if result is not None:
nosec_tests_to_skip = self._get_nosecs_from_contexts(
temp_context, test_result=result
)
if isinstance(temp_context["filename"], bytes):
result.fname = temp_context["filename"].decode("utf-8")
else:
result.fname = temp_context["filename"]
result.fdata = temp_context["file_data"]
if result.lineno is None:
result.lineno = temp_context["lineno"]
if result.linerange == []:
result.linerange = temp_context["linerange"]
if result.col_offset == -1:
result.col_offset = temp_context["col_offset"]
result.end_col_offset = temp_context.get(
"end_col_offset", 0
)
result.test = name
if result.test_id == "":
result.test_id = test._test_id
# don't skip the test if there was no nosec comment
if nosec_tests_to_skip is not None:
# If the set is empty then it means that nosec was
# used without test number -> update nosecs counter.
# If the test id is in the set of tests to skip,
# log and increment the skip by test count.
if not nosec_tests_to_skip:
LOG.debug("skipped, nosec without test number")
self.metrics.note_nosec()
continue
if result.test_id in nosec_tests_to_skip:
LOG.debug(
f"skipped, nosec for test {result.test_id}"
)
self.metrics.note_skipped_test()
continue
self.results.append(result)
LOG.debug("Issue identified by %s: %s", name, result)
sev = constants.RANKING.index(result.severity)
val = constants.RANKING_VALUES[result.severity]
scores["SEVERITY"][sev] += val
con = constants.RANKING.index(result.confidence)
val = constants.RANKING_VALUES[result.confidence]
scores["CONFIDENCE"][con] += val
else:
nosec_tests_to_skip = self._get_nosecs_from_contexts(
temp_context
)
if (
nosec_tests_to_skip
and test._test_id in nosec_tests_to_skip
):
LOG.warning(
f"nosec encountered ({test._test_id}), but no "
f"failed test on line {temp_context['lineno']}"
)
except Exception as e:
self.report_error(name, context, e)
if self.debug:
raise
LOG.debug("Returning scores: %s", scores)
return scores
def _get_nosecs_from_contexts(self, context, test_result=None):
"""Use context and optional test result to get set of tests to skip.
:param context: temp context
:param test_result: optional test result
:return: set of tests to skip for the line based on contexts
"""
nosec_tests_to_skip = set()
base_tests = (
self.nosec_lines.get(test_result.lineno, None)
if test_result
else None
)
context_tests = utils.get_nosec(self.nosec_lines, context)
# if both are none there were no comments
# this is explicitly different from being empty.
# empty set indicates blanket nosec comment without
# individual test names or ids
if base_tests is None and context_tests is None:
nosec_tests_to_skip = None
# combine tests from current line and context line
if base_tests is not None:
nosec_tests_to_skip.update(base_tests)
if context_tests is not None:
nosec_tests_to_skip.update(context_tests)
return nosec_tests_to_skip
@staticmethod
def report_error(test, context, error):
what = "Bandit internal error running: "
what += f"{test} "
what += "on file %s at line %i: " % (
context._context["filename"],
context._context["lineno"],
)
what += str(error)
import traceback
what += traceback.format_exc()
LOG.error(what)

View File

@@ -0,0 +1,398 @@
#
# Copyright 2014 Hewlett-Packard Development Company, L.P.
#
# SPDX-License-Identifier: Apache-2.0
import ast
import logging
import os.path
import sys
try:
import configparser
except ImportError:
import ConfigParser as configparser
LOG = logging.getLogger(__name__)
"""Various helper functions."""
def _get_attr_qual_name(node, aliases):
"""Get a the full name for the attribute node.
This will resolve a pseudo-qualified name for the attribute
rooted at node as long as all the deeper nodes are Names or
Attributes. This will give you how the code referenced the name but
will not tell you what the name actually refers to. If we
encounter a node without a static name we punt with an
empty string. If this encounters something more complex, such as
foo.mylist[0](a,b) we just return empty string.
:param node: AST Name or Attribute node
:param aliases: Import aliases dictionary
:returns: Qualified name referred to by the attribute or name.
"""
if isinstance(node, ast.Name):
if node.id in aliases:
return aliases[node.id]
return node.id
elif isinstance(node, ast.Attribute):
name = f"{_get_attr_qual_name(node.value, aliases)}.{node.attr}"
if name in aliases:
return aliases[name]
return name
else:
return ""
def get_call_name(node, aliases):
if isinstance(node.func, ast.Name):
if deepgetattr(node, "func.id") in aliases:
return aliases[deepgetattr(node, "func.id")]
return deepgetattr(node, "func.id")
elif isinstance(node.func, ast.Attribute):
return _get_attr_qual_name(node.func, aliases)
else:
return ""
def get_func_name(node):
return node.name # TODO(tkelsey): get that qualname using enclosing scope
def get_qual_attr(node, aliases):
if isinstance(node, ast.Attribute):
try:
val = deepgetattr(node, "value.id")
if val in aliases:
prefix = aliases[val]
else:
prefix = deepgetattr(node, "value.id")
except Exception:
# NOTE(tkelsey): degrade gracefully when we can't get the fully
# qualified name for an attr, just return its base name.
prefix = ""
return f"{prefix}.{node.attr}"
else:
return "" # TODO(tkelsey): process other node types
def deepgetattr(obj, attr):
"""Recurses through an attribute chain to get the ultimate value."""
for key in attr.split("."):
obj = getattr(obj, key)
return obj
class InvalidModulePath(Exception):
pass
class ConfigError(Exception):
"""Raised when the config file fails validation."""
def __init__(self, message, config_file):
self.config_file = config_file
self.message = f"{config_file} : {message}"
super().__init__(self.message)
class ProfileNotFound(Exception):
"""Raised when chosen profile cannot be found."""
def __init__(self, config_file, profile):
self.config_file = config_file
self.profile = profile
message = "Unable to find profile ({}) in config file: {}".format(
self.profile,
self.config_file,
)
super().__init__(message)
def warnings_formatter(
message, category=UserWarning, filename="", lineno=-1, line=""
):
"""Monkey patch for warnings.warn to suppress cruft output."""
return f"{message}\n"
def get_module_qualname_from_path(path):
"""Get the module's qualified name by analysis of the path.
Resolve the absolute pathname and eliminate symlinks. This could result in
an incorrect name if symlinks are used to restructure the python lib
directory.
Starting from the right-most directory component look for __init__.py in
the directory component. If it exists then the directory name is part of
the module name. Move left to the subsequent directory components until a
directory is found without __init__.py.
:param: Path to module file. Relative paths will be resolved relative to
current working directory.
:return: fully qualified module name
"""
(head, tail) = os.path.split(path)
if head == "" or tail == "":
raise InvalidModulePath(
f'Invalid python file path: "{path}" Missing path or file name'
)
qname = [os.path.splitext(tail)[0]]
while head not in ["/", ".", ""]:
if os.path.isfile(os.path.join(head, "__init__.py")):
(head, tail) = os.path.split(head)
qname.insert(0, tail)
else:
break
qualname = ".".join(qname)
return qualname
def namespace_path_join(base, name):
"""Extend the current namespace path with an additional name
Take a namespace path (i.e., package.module.class) and extends it
with an additional name (i.e., package.module.class.subclass).
This is similar to how os.path.join works.
:param base: (String) The base namespace path.
:param name: (String) The new name to append to the base path.
:returns: (String) A new namespace path resulting from combination of
base and name.
"""
return f"{base}.{name}"
def namespace_path_split(path):
"""Split the namespace path into a pair (head, tail).
Tail will be the last namespace path component and head will
be everything leading up to that in the path. This is similar to
os.path.split.
:param path: (String) A namespace path.
:returns: (String, String) A tuple where the first component is the base
path and the second is the last path component.
"""
return tuple(path.rsplit(".", 1))
def escaped_bytes_representation(b):
"""PY3 bytes need escaping for comparison with other strings.
In practice it turns control characters into acceptable codepoints then
encodes them into bytes again to turn unprintable bytes into printable
escape sequences.
This is safe to do for the whole range 0..255 and result matches
unicode_escape on a unicode string.
"""
return b.decode("unicode_escape").encode("unicode_escape")
def calc_linerange(node):
"""Calculate linerange for subtree"""
if hasattr(node, "_bandit_linerange"):
return node._bandit_linerange
lines_min = 9999999999
lines_max = -1
if hasattr(node, "lineno"):
lines_min = node.lineno
lines_max = node.lineno
for n in ast.iter_child_nodes(node):
lines_minmax = calc_linerange(n)
lines_min = min(lines_min, lines_minmax[0])
lines_max = max(lines_max, lines_minmax[1])
node._bandit_linerange = (lines_min, lines_max)
return (lines_min, lines_max)
def linerange(node):
"""Get line number range from a node."""
if hasattr(node, "lineno"):
return list(range(node.lineno, node.end_lineno + 1))
else:
if hasattr(node, "_bandit_linerange_stripped"):
lines_minmax = node._bandit_linerange_stripped
return list(range(lines_minmax[0], lines_minmax[1] + 1))
strip = {
"body": None,
"orelse": None,
"handlers": None,
"finalbody": None,
}
for key in strip.keys():
if hasattr(node, key):
strip[key] = getattr(node, key)
setattr(node, key, [])
lines_min = 9999999999
lines_max = -1
if hasattr(node, "lineno"):
lines_min = node.lineno
lines_max = node.lineno
for n in ast.iter_child_nodes(node):
lines_minmax = calc_linerange(n)
lines_min = min(lines_min, lines_minmax[0])
lines_max = max(lines_max, lines_minmax[1])
for key in strip.keys():
if strip[key] is not None:
setattr(node, key, strip[key])
if lines_max == -1:
lines_min = 0
lines_max = 1
node._bandit_linerange_stripped = (lines_min, lines_max)
lines = list(range(lines_min, lines_max + 1))
"""Try and work around a known Python bug with multi-line strings."""
# deal with multiline strings lineno behavior (Python issue #16806)
if hasattr(node, "_bandit_sibling") and hasattr(
node._bandit_sibling, "lineno"
):
start = min(lines)
delta = node._bandit_sibling.lineno - start
if delta > 1:
return list(range(start, node._bandit_sibling.lineno))
return lines
def concat_string(node, stop=None):
"""Builds a string from a ast.BinOp chain.
This will build a string from a series of ast.Constant nodes wrapped in
ast.BinOp nodes. Something like "a" + "b" + "c" or "a %s" % val etc.
The provided node can be any participant in the BinOp chain.
:param node: (ast.Constant or ast.BinOp) The node to process
:param stop: (ast.Constant or ast.BinOp) Optional base node to stop at
:returns: (Tuple) the root node of the expression, the string value
"""
def _get(node, bits, stop=None):
if node != stop:
bits.append(
_get(node.left, bits, stop)
if isinstance(node.left, ast.BinOp)
else node.left
)
bits.append(
_get(node.right, bits, stop)
if isinstance(node.right, ast.BinOp)
else node.right
)
bits = [node]
while isinstance(node._bandit_parent, ast.BinOp):
node = node._bandit_parent
if isinstance(node, ast.BinOp):
_get(node, bits, stop)
return (
node,
" ".join(
[
x.value
for x in bits
if isinstance(x, ast.Constant) and isinstance(x.value, str)
]
),
)
def get_called_name(node):
"""Get a function name from an ast.Call node.
An ast.Call node representing a method call with present differently to one
wrapping a function call: thing.call() vs call(). This helper will grab the
unqualified call name correctly in either case.
:param node: (ast.Call) the call node
:returns: (String) the function name
"""
func = node.func
try:
return func.attr if isinstance(func, ast.Attribute) else func.id
except AttributeError:
return ""
def get_path_for_function(f):
"""Get the path of the file where the function is defined.
:returns: the path, or None if one could not be found or f is not a real
function
"""
if hasattr(f, "__module__"):
module_name = f.__module__
elif hasattr(f, "im_func"):
module_name = f.im_func.__module__
else:
LOG.warning("Cannot resolve file where %s is defined", f)
return None
module = sys.modules[module_name]
if hasattr(module, "__file__"):
return module.__file__
else:
LOG.warning("Cannot resolve file path for module %s", module_name)
return None
def parse_ini_file(f_loc):
config = configparser.ConfigParser()
try:
config.read(f_loc)
return {k: v for k, v in config.items("bandit")}
except (configparser.Error, KeyError, TypeError):
LOG.warning(
"Unable to parse config file %s or missing [bandit] " "section",
f_loc,
)
return None
def check_ast_node(name):
"Check if the given name is that of a valid AST node."
try:
# These ast Node types don't exist in Python 3.14, but plugins may
# still check on them.
if sys.version_info >= (3, 14) and name in (
"Num",
"Str",
"Ellipsis",
"NameConstant",
"Bytes",
):
return name
node = getattr(ast, name)
if issubclass(node, ast.AST):
return name
except AttributeError: # nosec(tkelsey): catching expected exception
pass
raise TypeError(f"Error: {name} is not a valid node type in AST")
def get_nosec(nosec_lines, context):
for lineno in context["linerange"]:
nosec = nosec_lines.get(lineno, None)
if nosec is not None:
return nosec
return None

View File

@@ -0,0 +1,82 @@
#
# SPDX-License-Identifier: Apache-2.0
r"""
=============
CSV Formatter
=============
This formatter outputs the issues in a comma separated values format.
:Example:
.. code-block:: none
filename,test_name,test_id,issue_severity,issue_confidence,issue_cwe,
issue_text,line_number,line_range,more_info
examples/yaml_load.py,blacklist_calls,B301,MEDIUM,HIGH,
https://cwe.mitre.org/data/definitions/20.html,"Use of unsafe yaml
load. Allows instantiation of arbitrary objects. Consider yaml.safe_load().
",5,[5],https://bandit.readthedocs.io/en/latest/
.. versionadded:: 0.11.0
.. versionchanged:: 1.5.0
New field `more_info` added to output
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
# Necessary for this formatter to work when imported on Python 2. Importing
# the standard library's csv module conflicts with the name of this module.
import csv
import logging
import sys
from bandit.core import docs_utils
LOG = logging.getLogger(__name__)
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Prints issues in CSV format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
results = manager.get_issue_list(
sev_level=sev_level, conf_level=conf_level
)
with fileobj:
fieldnames = [
"filename",
"test_name",
"test_id",
"issue_severity",
"issue_confidence",
"issue_cwe",
"issue_text",
"line_number",
"col_offset",
"end_col_offset",
"line_range",
"more_info",
]
writer = csv.DictWriter(
fileobj, fieldnames=fieldnames, extrasaction="ignore"
)
writer.writeheader()
for result in results:
r = result.as_dict(with_code=False)
r["issue_cwe"] = r["issue_cwe"]["link"]
r["more_info"] = docs_utils.get_url(r["test_id"])
writer.writerow(r)
if fileobj.name != sys.stdout.name:
LOG.info("CSV output written to file: %s", fileobj.name)

View File

@@ -0,0 +1,161 @@
#
# Copyright (c) 2017 Hewlett Packard Enterprise
#
# SPDX-License-Identifier: Apache-2.0
"""
================
Custom Formatter
================
This formatter outputs the issues in custom machine-readable format.
default template: ``{abspath}:{line}: {test_id}[bandit]: {severity}: {msg}``
:Example:
.. code-block:: none
/usr/lib/python3.6/site-packages/openlp/core/utils/__init__.py:\
405: B310[bandit]: MEDIUM: Audit url open for permitted schemes. \
Allowing use of file:/ or custom schemes is often unexpected.
.. versionadded:: 1.5.0
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
import logging
import os
import re
import string
import sys
from bandit.core import test_properties
LOG = logging.getLogger(__name__)
class SafeMapper(dict):
"""Safe mapper to handle format key errors"""
@classmethod # To prevent PEP8 warnings in the test suite
def __missing__(cls, key):
return "{%s}" % key
@test_properties.accepts_baseline
def report(manager, fileobj, sev_level, conf_level, template=None):
"""Prints issues in custom format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param template: Output template with non-terminal tags <N>
(default: '{abspath}:{line}:
{test_id}[bandit]: {severity}: {msg}')
"""
machine_output = {"results": [], "errors": []}
for fname, reason in manager.get_skipped():
machine_output["errors"].append({"filename": fname, "reason": reason})
results = manager.get_issue_list(
sev_level=sev_level, conf_level=conf_level
)
msg_template = template
if template is None:
msg_template = "{abspath}:{line}: {test_id}[bandit]: {severity}: {msg}"
# Dictionary of non-terminal tags that will be expanded
tag_mapper = {
"abspath": lambda issue: os.path.abspath(issue.fname),
"relpath": lambda issue: os.path.relpath(issue.fname),
"line": lambda issue: issue.lineno,
"col": lambda issue: issue.col_offset,
"end_col": lambda issue: issue.end_col_offset,
"test_id": lambda issue: issue.test_id,
"severity": lambda issue: issue.severity,
"msg": lambda issue: issue.text,
"confidence": lambda issue: issue.confidence,
"range": lambda issue: issue.linerange,
"cwe": lambda issue: issue.cwe,
}
# Create dictionary with tag sets to speed up search for similar tags
tag_sim_dict = {tag: set(tag) for tag, _ in tag_mapper.items()}
# Parse the format_string template and check the validity of tags
try:
parsed_template_orig = list(string.Formatter().parse(msg_template))
# of type (literal_text, field_name, fmt_spec, conversion)
# Check the format validity only, ignore keys
string.Formatter().vformat(msg_template, (), SafeMapper(line=0))
except ValueError as e:
LOG.error("Template is not in valid format: %s", e.args[0])
sys.exit(2)
tag_set = {t[1] for t in parsed_template_orig if t[1] is not None}
if not tag_set:
LOG.error("No tags were found in the template. Are you missing '{}'?")
sys.exit(2)
def get_similar_tag(tag):
similarity_list = [
(len(set(tag) & t_set), t) for t, t_set in tag_sim_dict.items()
]
return sorted(similarity_list)[-1][1]
tag_blacklist = []
for tag in tag_set:
# check if the tag is in dictionary
if tag not in tag_mapper:
similar_tag = get_similar_tag(tag)
LOG.warning(
"Tag '%s' was not recognized and will be skipped, "
"did you mean to use '%s'?",
tag,
similar_tag,
)
tag_blacklist += [tag]
# Compose the message template back with the valid values only
msg_parsed_template_list = []
for literal_text, field_name, fmt_spec, conversion in parsed_template_orig:
if literal_text:
# if there is '{' or '}', double it to prevent expansion
literal_text = re.sub("{", "{{", literal_text)
literal_text = re.sub("}", "}}", literal_text)
msg_parsed_template_list.append(literal_text)
if field_name is not None:
if field_name in tag_blacklist:
msg_parsed_template_list.append(field_name)
continue
# Append the fmt_spec part
params = [field_name, fmt_spec, conversion]
markers = ["", ":", "!"]
msg_parsed_template_list.append(
["{"]
+ [f"{m + p}" if p else "" for m, p in zip(markers, params)]
+ ["}"]
)
msg_parsed_template = (
"".join([item for lst in msg_parsed_template_list for item in lst])
+ "\n"
)
with fileobj:
for defect in results:
evaluated_tags = SafeMapper(
(k, v(defect)) for k, v in tag_mapper.items()
)
output = msg_parsed_template.format(**evaluated_tags)
fileobj.write(output)
if fileobj.name != sys.stdout.name:
LOG.info("Result written to file: %s", fileobj.name)

View File

@@ -0,0 +1,394 @@
# Copyright (c) 2015 Rackspace, Inc.
# Copyright (c) 2015 Hewlett Packard Enterprise
#
# SPDX-License-Identifier: Apache-2.0
r"""
==============
HTML formatter
==============
This formatter outputs the issues as HTML.
:Example:
.. code-block:: html
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<title>
Bandit Report
</title>
<style>
html * {
font-family: "Arial", sans-serif;
}
pre {
font-family: "Monaco", monospace;
}
.bordered-box {
border: 1px solid black;
padding-top:.5em;
padding-bottom:.5em;
padding-left:1em;
}
.metrics-box {
font-size: 1.1em;
line-height: 130%;
}
.metrics-title {
font-size: 1.5em;
font-weight: 500;
margin-bottom: .25em;
}
.issue-description {
font-size: 1.3em;
font-weight: 500;
}
.candidate-issues {
margin-left: 2em;
border-left: solid 1px; LightGray;
padding-left: 5%;
margin-top: .2em;
margin-bottom: .2em;
}
.issue-block {
border: 1px solid LightGray;
padding-left: .5em;
padding-top: .5em;
padding-bottom: .5em;
margin-bottom: .5em;
}
.issue-sev-high {
background-color: Pink;
}
.issue-sev-medium {
background-color: NavajoWhite;
}
.issue-sev-low {
background-color: LightCyan;
}
</style>
</head>
<body>
<div id="metrics">
<div class="metrics-box bordered-box">
<div class="metrics-title">
Metrics:<br>
</div>
Total lines of code: <span id="loc">9</span><br>
Total lines skipped (#nosec): <span id="nosec">0</span>
</div>
</div>
<br>
<div id="results">
<div id="issue-0">
<div class="issue-block issue-sev-medium">
<b>yaml_load: </b> Use of unsafe yaml load. Allows
instantiation of arbitrary objects. Consider yaml.safe_load().<br>
<b>Test ID:</b> B506<br>
<b>Severity: </b>MEDIUM<br>
<b>Confidence: </b>HIGH<br>
<b>CWE: </b>CWE-20 (https://cwe.mitre.org/data/definitions/20.html)<br>
<b>File: </b><a href="examples/yaml_load.py"
target="_blank">examples/yaml_load.py</a> <br>
<b>More info: </b><a href="https://bandit.readthedocs.io/en/latest/
plugins/yaml_load.html" target="_blank">
https://bandit.readthedocs.io/en/latest/plugins/yaml_load.html</a>
<br>
<div class="code">
<pre>
5 ystr = yaml.dump({'a' : 1, 'b' : 2, 'c' : 3})
6 y = yaml.load(ystr)
7 yaml.dump(y)
</pre>
</div>
</div>
</div>
</div>
</body>
</html>
.. versionadded:: 0.14.0
.. versionchanged:: 1.5.0
New field `more_info` added to output
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
import logging
import sys
from html import escape as html_escape
from bandit.core import docs_utils
from bandit.core import test_properties
from bandit.formatters import utils
LOG = logging.getLogger(__name__)
@test_properties.accepts_baseline
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Writes issues to 'fileobj' in HTML format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
header_block = """
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<title>
Bandit Report
</title>
<style>
html * {
font-family: "Arial", sans-serif;
}
pre {
font-family: "Monaco", monospace;
}
.bordered-box {
border: 1px solid black;
padding-top:.5em;
padding-bottom:.5em;
padding-left:1em;
}
.metrics-box {
font-size: 1.1em;
line-height: 130%;
}
.metrics-title {
font-size: 1.5em;
font-weight: 500;
margin-bottom: .25em;
}
.issue-description {
font-size: 1.3em;
font-weight: 500;
}
.candidate-issues {
margin-left: 2em;
border-left: solid 1px; LightGray;
padding-left: 5%;
margin-top: .2em;
margin-bottom: .2em;
}
.issue-block {
border: 1px solid LightGray;
padding-left: .5em;
padding-top: .5em;
padding-bottom: .5em;
margin-bottom: .5em;
}
.issue-sev-high {
background-color: Pink;
}
.issue-sev-medium {
background-color: NavajoWhite;
}
.issue-sev-low {
background-color: LightCyan;
}
</style>
</head>
"""
report_block = """
<body>
{metrics}
{skipped}
<br>
<div id="results">
{results}
</div>
</body>
</html>
"""
issue_block = """
<div id="issue-{issue_no}">
<div class="issue-block {issue_class}">
<b>{test_name}: </b> {test_text}<br>
<b>Test ID:</b> {test_id}<br>
<b>Severity: </b>{severity}<br>
<b>Confidence: </b>{confidence}<br>
<b>CWE: </b><a href="{cwe_link}" target="_blank">CWE-{cwe.id}</a><br>
<b>File: </b><a href="{path}" target="_blank">{path}</a><br>
<b>Line number: </b>{line_number}<br>
<b>More info: </b><a href="{url}" target="_blank">{url}</a><br>
{code}
{candidates}
</div>
</div>
"""
code_block = """
<div class="code">
<pre>
{code}
</pre>
</div>
"""
candidate_block = """
<div class="candidates">
<br>
<b>Candidates: </b>
{candidate_list}
</div>
"""
candidate_issue = """
<div class="candidate">
<div class="candidate-issues">
<pre>{code}</pre>
</div>
</div>
"""
skipped_block = """
<br>
<div id="skipped">
<div class="bordered-box">
<b>Skipped files:</b><br><br>
{files_list}
</div>
</div>
"""
metrics_block = """
<div id="metrics">
<div class="metrics-box bordered-box">
<div class="metrics-title">
Metrics:<br>
</div>
Total lines of code: <span id="loc">{loc}</span><br>
Total lines skipped (#nosec): <span id="nosec">{nosec}</span>
</div>
</div>
"""
issues = manager.get_issue_list(sev_level=sev_level, conf_level=conf_level)
baseline = not isinstance(issues, list)
# build the skipped string to insert in the report
skipped_str = "".join(
f"{fname} <b>reason:</b> {reason}<br>"
for fname, reason in manager.get_skipped()
)
if skipped_str:
skipped_text = skipped_block.format(files_list=skipped_str)
else:
skipped_text = ""
# build the results string to insert in the report
results_str = ""
for index, issue in enumerate(issues):
if not baseline or len(issues[issue]) == 1:
candidates = ""
safe_code = html_escape(
issue.get_code(lines, True).strip("\n").lstrip(" ")
)
code = code_block.format(code=safe_code)
else:
candidates_str = ""
code = ""
for candidate in issues[issue]:
candidate_code = html_escape(
candidate.get_code(lines, True).strip("\n").lstrip(" ")
)
candidates_str += candidate_issue.format(code=candidate_code)
candidates = candidate_block.format(candidate_list=candidates_str)
url = docs_utils.get_url(issue.test_id)
results_str += issue_block.format(
issue_no=index,
issue_class=f"issue-sev-{issue.severity.lower()}",
test_name=issue.test,
test_id=issue.test_id,
test_text=issue.text,
severity=issue.severity,
confidence=issue.confidence,
cwe=issue.cwe,
cwe_link=issue.cwe.link(),
path=issue.fname,
code=code,
candidates=candidates,
url=url,
line_number=issue.lineno,
)
# build the metrics string to insert in the report
metrics_summary = metrics_block.format(
loc=manager.metrics.data["_totals"]["loc"],
nosec=manager.metrics.data["_totals"]["nosec"],
)
# build the report and output it
report_contents = report_block.format(
metrics=metrics_summary, skipped=skipped_text, results=results_str
)
with fileobj:
wrapped_file = utils.wrap_file_object(fileobj)
wrapped_file.write(header_block)
wrapped_file.write(report_contents)
if fileobj.name != sys.stdout.name:
LOG.info("HTML output written to file: %s", fileobj.name)

View File

@@ -0,0 +1,155 @@
#
# SPDX-License-Identifier: Apache-2.0
r"""
==============
JSON formatter
==============
This formatter outputs the issues in JSON.
:Example:
.. code-block:: javascript
{
"errors": [],
"generated_at": "2015-12-16T22:27:34Z",
"metrics": {
"_totals": {
"CONFIDENCE.HIGH": 1,
"CONFIDENCE.LOW": 0,
"CONFIDENCE.MEDIUM": 0,
"CONFIDENCE.UNDEFINED": 0,
"SEVERITY.HIGH": 0,
"SEVERITY.LOW": 0,
"SEVERITY.MEDIUM": 1,
"SEVERITY.UNDEFINED": 0,
"loc": 5,
"nosec": 0
},
"examples/yaml_load.py": {
"CONFIDENCE.HIGH": 1,
"CONFIDENCE.LOW": 0,
"CONFIDENCE.MEDIUM": 0,
"CONFIDENCE.UNDEFINED": 0,
"SEVERITY.HIGH": 0,
"SEVERITY.LOW": 0,
"SEVERITY.MEDIUM": 1,
"SEVERITY.UNDEFINED": 0,
"loc": 5,
"nosec": 0
}
},
"results": [
{
"code": "4 ystr = yaml.dump({'a' : 1, 'b' : 2, 'c' : 3})\n5
y = yaml.load(ystr)\n6 yaml.dump(y)\n",
"filename": "examples/yaml_load.py",
"issue_confidence": "HIGH",
"issue_severity": "MEDIUM",
"issue_cwe": {
"id": 20,
"link": "https://cwe.mitre.org/data/definitions/20.html"
},
"issue_text": "Use of unsafe yaml load. Allows instantiation of
arbitrary objects. Consider yaml.safe_load().\n",
"line_number": 5,
"line_range": [
5
],
"more_info": "https://bandit.readthedocs.io/en/latest/",
"test_name": "blacklist_calls",
"test_id": "B301"
}
]
}
.. versionadded:: 0.10.0
.. versionchanged:: 1.5.0
New field `more_info` added to output
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
# Necessary so we can import the standard library json module while continuing
# to name this file json.py. (Python 2 only)
import datetime
import json
import logging
import operator
import sys
from bandit.core import docs_utils
from bandit.core import test_properties
LOG = logging.getLogger(__name__)
@test_properties.accepts_baseline
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""''Prints issues in JSON format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
machine_output = {"results": [], "errors": []}
for fname, reason in manager.get_skipped():
machine_output["errors"].append({"filename": fname, "reason": reason})
results = manager.get_issue_list(
sev_level=sev_level, conf_level=conf_level
)
baseline = not isinstance(results, list)
if baseline:
collector = []
for r in results:
d = r.as_dict(max_lines=lines)
d["more_info"] = docs_utils.get_url(d["test_id"])
if len(results[r]) > 1:
d["candidates"] = [
c.as_dict(max_lines=lines) for c in results[r]
]
collector.append(d)
else:
collector = [r.as_dict(max_lines=lines) for r in results]
for elem in collector:
elem["more_info"] = docs_utils.get_url(elem["test_id"])
itemgetter = operator.itemgetter
if manager.agg_type == "vuln":
machine_output["results"] = sorted(
collector, key=itemgetter("test_name")
)
else:
machine_output["results"] = sorted(
collector, key=itemgetter("filename")
)
machine_output["metrics"] = manager.metrics.data
# timezone agnostic format
TS_FORMAT = "%Y-%m-%dT%H:%M:%SZ"
time_string = datetime.datetime.now(datetime.timezone.utc).strftime(
TS_FORMAT
)
machine_output["generated_at"] = time_string
result = json.dumps(
machine_output, sort_keys=True, indent=2, separators=(",", ": ")
)
with fileobj:
fileobj.write(result)
if fileobj.name != sys.stdout.name:
LOG.info("JSON output written to file: %s", fileobj.name)

View File

@@ -0,0 +1,374 @@
# Copyright (c) Microsoft. All Rights Reserved.
#
# SPDX-License-Identifier: Apache-2.0
#
# Note: this code mostly incorporated from
# https://github.com/microsoft/bandit-sarif-formatter
#
r"""
===============
SARIF formatter
===============
This formatter outputs the issues in SARIF formatted JSON.
:Example:
.. code-block:: javascript
{
"runs": [
{
"tool": {
"driver": {
"name": "Bandit",
"organization": "PyCQA",
"rules": [
{
"id": "B101",
"name": "assert_used",
"properties": {
"tags": [
"security",
"external/cwe/cwe-703"
],
"precision": "high"
},
"helpUri": "https://bandit.readthedocs.io/en/1.7.8/plugins/b101_assert_used.html"
}
],
"version": "1.7.8",
"semanticVersion": "1.7.8"
}
},
"invocations": [
{
"executionSuccessful": true,
"endTimeUtc": "2024-03-05T03:28:48Z"
}
],
"properties": {
"metrics": {
"_totals": {
"loc": 1,
"nosec": 0,
"skipped_tests": 0,
"SEVERITY.UNDEFINED": 0,
"CONFIDENCE.UNDEFINED": 0,
"SEVERITY.LOW": 1,
"CONFIDENCE.LOW": 0,
"SEVERITY.MEDIUM": 0,
"CONFIDENCE.MEDIUM": 0,
"SEVERITY.HIGH": 0,
"CONFIDENCE.HIGH": 1
},
"./examples/assert.py": {
"loc": 1,
"nosec": 0,
"skipped_tests": 0,
"SEVERITY.UNDEFINED": 0,
"SEVERITY.LOW": 1,
"SEVERITY.MEDIUM": 0,
"SEVERITY.HIGH": 0,
"CONFIDENCE.UNDEFINED": 0,
"CONFIDENCE.LOW": 0,
"CONFIDENCE.MEDIUM": 0,
"CONFIDENCE.HIGH": 1
}
}
},
"results": [
{
"message": {
"text": "Use of assert detected. The enclosed code will be removed when compiling to optimised byte code."
},
"level": "note",
"locations": [
{
"physicalLocation": {
"region": {
"snippet": {
"text": "assert True\n"
},
"endColumn": 11,
"endLine": 1,
"startColumn": 0,
"startLine": 1
},
"artifactLocation": {
"uri": "examples/assert.py"
},
"contextRegion": {
"snippet": {
"text": "assert True\n"
},
"endLine": 1,
"startLine": 1
}
}
}
],
"properties": {
"issue_confidence": "HIGH",
"issue_severity": "LOW"
},
"ruleId": "B101",
"ruleIndex": 0
}
]
}
],
"version": "2.1.0",
"$schema": "https://json.schemastore.org/sarif-2.1.0.json"
}
.. versionadded:: 1.7.8
""" # noqa: E501
import datetime
import logging
import pathlib
import sys
import urllib.parse as urlparse
import sarif_om as om
from jschema_to_python.to_json import to_json
import bandit
from bandit.core import docs_utils
LOG = logging.getLogger(__name__)
SCHEMA_URI = "https://json.schemastore.org/sarif-2.1.0.json"
SCHEMA_VER = "2.1.0"
TS_FORMAT = "%Y-%m-%dT%H:%M:%SZ"
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Prints issues in SARIF format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
log = om.SarifLog(
schema_uri=SCHEMA_URI,
version=SCHEMA_VER,
runs=[
om.Run(
tool=om.Tool(
driver=om.ToolComponent(
name="Bandit",
organization=bandit.__author__,
semantic_version=bandit.__version__,
version=bandit.__version__,
)
),
invocations=[
om.Invocation(
end_time_utc=datetime.datetime.now(
datetime.timezone.utc
).strftime(TS_FORMAT),
execution_successful=True,
)
],
properties={"metrics": manager.metrics.data},
)
],
)
run = log.runs[0]
invocation = run.invocations[0]
skips = manager.get_skipped()
add_skipped_file_notifications(skips, invocation)
issues = manager.get_issue_list(sev_level=sev_level, conf_level=conf_level)
add_results(issues, run)
serializedLog = to_json(log)
with fileobj:
fileobj.write(serializedLog)
if fileobj.name != sys.stdout.name:
LOG.info("SARIF output written to file: %s", fileobj.name)
def add_skipped_file_notifications(skips, invocation):
if skips is None or len(skips) == 0:
return
if invocation.tool_configuration_notifications is None:
invocation.tool_configuration_notifications = []
for skip in skips:
(file_name, reason) = skip
notification = om.Notification(
level="error",
message=om.Message(text=reason),
locations=[
om.Location(
physical_location=om.PhysicalLocation(
artifact_location=om.ArtifactLocation(
uri=to_uri(file_name)
)
)
)
],
)
invocation.tool_configuration_notifications.append(notification)
def add_results(issues, run):
if run.results is None:
run.results = []
rules = {}
rule_indices = {}
for issue in issues:
result = create_result(issue, rules, rule_indices)
run.results.append(result)
if len(rules) > 0:
run.tool.driver.rules = list(rules.values())
def create_result(issue, rules, rule_indices):
issue_dict = issue.as_dict()
rule, rule_index = create_or_find_rule(issue_dict, rules, rule_indices)
physical_location = om.PhysicalLocation(
artifact_location=om.ArtifactLocation(
uri=to_uri(issue_dict["filename"])
)
)
add_region_and_context_region(
physical_location,
issue_dict["line_range"],
issue_dict["col_offset"],
issue_dict["end_col_offset"],
issue_dict["code"],
)
return om.Result(
rule_id=rule.id,
rule_index=rule_index,
message=om.Message(text=issue_dict["issue_text"]),
level=level_from_severity(issue_dict["issue_severity"]),
locations=[om.Location(physical_location=physical_location)],
properties={
"issue_confidence": issue_dict["issue_confidence"],
"issue_severity": issue_dict["issue_severity"],
},
)
def level_from_severity(severity):
if severity == "HIGH":
return "error"
elif severity == "MEDIUM":
return "warning"
elif severity == "LOW":
return "note"
else:
return "warning"
def add_region_and_context_region(
physical_location, line_range, col_offset, end_col_offset, code
):
if code:
first_line_number, snippet_lines = parse_code(code)
snippet_line = snippet_lines[line_range[0] - first_line_number]
snippet = om.ArtifactContent(text=snippet_line)
else:
snippet = None
physical_location.region = om.Region(
start_line=line_range[0],
end_line=line_range[1] if len(line_range) > 1 else line_range[0],
start_column=col_offset + 1,
end_column=end_col_offset + 1,
snippet=snippet,
)
if code:
physical_location.context_region = om.Region(
start_line=first_line_number,
end_line=first_line_number + len(snippet_lines) - 1,
snippet=om.ArtifactContent(text="".join(snippet_lines)),
)
def parse_code(code):
code_lines = code.split("\n")
# The last line from the split has nothing in it; it's an artifact of the
# last "real" line ending in a newline. Unless, of course, it doesn't:
last_line = code_lines[len(code_lines) - 1]
last_real_line_ends_in_newline = False
if len(last_line) == 0:
code_lines.pop()
last_real_line_ends_in_newline = True
snippet_lines = []
first_line_number = 0
first = True
for code_line in code_lines:
number_and_snippet_line = code_line.split(" ", 1)
if first:
first_line_number = int(number_and_snippet_line[0])
first = False
snippet_line = number_and_snippet_line[1] + "\n"
snippet_lines.append(snippet_line)
if not last_real_line_ends_in_newline:
last_line = snippet_lines[len(snippet_lines) - 1]
snippet_lines[len(snippet_lines) - 1] = last_line[: len(last_line) - 1]
return first_line_number, snippet_lines
def create_or_find_rule(issue_dict, rules, rule_indices):
rule_id = issue_dict["test_id"]
if rule_id in rules:
return rules[rule_id], rule_indices[rule_id]
rule = om.ReportingDescriptor(
id=rule_id,
name=issue_dict["test_name"],
help_uri=docs_utils.get_url(rule_id),
properties={
"tags": [
"security",
f"external/cwe/cwe-{issue_dict['issue_cwe'].get('id')}",
],
"precision": issue_dict["issue_confidence"].lower(),
},
)
index = len(rules)
rules[rule_id] = rule
rule_indices[rule_id] = index
return rule, index
def to_uri(file_path):
pure_path = pathlib.PurePath(file_path)
if pure_path.is_absolute():
return pure_path.as_uri()
else:
# Replace backslashes with slashes.
posix_path = pure_path.as_posix()
# %-encode special characters.
return urlparse.quote(posix_path)

View File

@@ -0,0 +1,244 @@
# Copyright (c) 2015 Hewlett Packard Enterprise
#
# SPDX-License-Identifier: Apache-2.0
r"""
================
Screen formatter
================
This formatter outputs the issues as color coded text to screen.
:Example:
.. code-block:: none
>> Issue: [B506: yaml_load] Use of unsafe yaml load. Allows
instantiation of arbitrary objects. Consider yaml.safe_load().
Severity: Medium Confidence: High
CWE: CWE-20 (https://cwe.mitre.org/data/definitions/20.html)
More Info: https://bandit.readthedocs.io/en/latest/
Location: examples/yaml_load.py:5
4 ystr = yaml.dump({'a' : 1, 'b' : 2, 'c' : 3})
5 y = yaml.load(ystr)
6 yaml.dump(y)
.. versionadded:: 0.9.0
.. versionchanged:: 1.5.0
New field `more_info` added to output
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
import datetime
import logging
import sys
from bandit.core import constants
from bandit.core import docs_utils
from bandit.core import test_properties
IS_WIN_PLATFORM = sys.platform.startswith("win32")
COLORAMA = False
# This fixes terminal colors not displaying properly on Windows systems.
# Colorama will intercept any ANSI escape codes and convert them to the
# proper Windows console API calls to change text color.
if IS_WIN_PLATFORM:
try:
import colorama
except ImportError:
pass
else:
COLORAMA = True
LOG = logging.getLogger(__name__)
COLOR = {
"DEFAULT": "\033[0m",
"HEADER": "\033[95m",
"LOW": "\033[94m",
"MEDIUM": "\033[93m",
"HIGH": "\033[91m",
}
def header(text, *args):
return f"{COLOR['HEADER']}{text % args}{COLOR['DEFAULT']}"
def get_verbose_details(manager):
bits = []
bits.append(header("Files in scope (%i):", len(manager.files_list)))
tpl = "\t%s (score: {SEVERITY: %i, CONFIDENCE: %i})"
bits.extend(
[
tpl % (item, sum(score["SEVERITY"]), sum(score["CONFIDENCE"]))
for (item, score) in zip(manager.files_list, manager.scores)
]
)
bits.append(header("Files excluded (%i):", len(manager.excluded_files)))
bits.extend([f"\t{fname}" for fname in manager.excluded_files])
return "\n".join([str(bit) for bit in bits])
def get_metrics(manager):
bits = []
bits.append(header("\nRun metrics:"))
for criteria, _ in constants.CRITERIA:
bits.append(f"\tTotal issues (by {criteria.lower()}):")
for rank in constants.RANKING:
bits.append(
"\t\t%s: %s"
% (
rank.capitalize(),
manager.metrics.data["_totals"][f"{criteria}.{rank}"],
)
)
return "\n".join([str(bit) for bit in bits])
def _output_issue_str(
issue, indent, show_lineno=True, show_code=True, lines=-1
):
# returns a list of lines that should be added to the existing lines list
bits = []
bits.append(
"%s%s>> Issue: [%s:%s] %s"
% (
indent,
COLOR[issue.severity],
issue.test_id,
issue.test,
issue.text,
)
)
bits.append(
"%s Severity: %s Confidence: %s"
% (
indent,
issue.severity.capitalize(),
issue.confidence.capitalize(),
)
)
bits.append(f"{indent} CWE: {str(issue.cwe)}")
bits.append(f"{indent} More Info: {docs_utils.get_url(issue.test_id)}")
bits.append(
"%s Location: %s:%s:%s%s"
% (
indent,
issue.fname,
issue.lineno if show_lineno else "",
issue.col_offset if show_lineno else "",
COLOR["DEFAULT"],
)
)
if show_code:
bits.extend(
[indent + line for line in issue.get_code(lines, True).split("\n")]
)
return "\n".join([bit for bit in bits])
def get_results(manager, sev_level, conf_level, lines):
bits = []
issues = manager.get_issue_list(sev_level, conf_level)
baseline = not isinstance(issues, list)
candidate_indent = " " * 10
if not len(issues):
return "\tNo issues identified."
for issue in issues:
# if not a baseline or only one candidate we know the issue
if not baseline or len(issues[issue]) == 1:
bits.append(_output_issue_str(issue, "", lines=lines))
# otherwise show the finding and the candidates
else:
bits.append(
_output_issue_str(
issue, "", show_lineno=False, show_code=False
)
)
bits.append("\n-- Candidate Issues --")
for candidate in issues[issue]:
bits.append(
_output_issue_str(candidate, candidate_indent, lines=lines)
)
bits.append("\n")
bits.append("-" * 50)
return "\n".join([bit for bit in bits])
def do_print(bits):
# needed so we can mock this stuff
print("\n".join([bit for bit in bits]))
@test_properties.accepts_baseline
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Prints discovered issues formatted for screen reading
This makes use of VT100 terminal codes for colored text.
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
if IS_WIN_PLATFORM and COLORAMA:
colorama.init()
bits = []
if not manager.quiet or manager.results_count(sev_level, conf_level):
bits.append(
header(
"Run started:%s", datetime.datetime.now(datetime.timezone.utc)
)
)
if manager.verbose:
bits.append(get_verbose_details(manager))
bits.append(header("\nTest results:"))
bits.append(get_results(manager, sev_level, conf_level, lines))
bits.append(header("\nCode scanned:"))
bits.append(
"\tTotal lines of code: %i"
% (manager.metrics.data["_totals"]["loc"])
)
bits.append(
"\tTotal lines skipped (#nosec): %i"
% (manager.metrics.data["_totals"]["nosec"])
)
bits.append(get_metrics(manager))
skipped = manager.get_skipped()
bits.append(header("Files skipped (%i):", len(skipped)))
bits.extend(["\t%s (%s)" % skip for skip in skipped])
do_print(bits)
if fileobj.name != sys.stdout.name:
LOG.info(
"Screen formatter output was not written to file: %s, "
"consider '-f txt'",
fileobj.name,
)
if IS_WIN_PLATFORM and COLORAMA:
colorama.deinit()

View File

@@ -0,0 +1,200 @@
# Copyright (c) 2015 Hewlett Packard Enterprise
#
# SPDX-License-Identifier: Apache-2.0
r"""
==============
Text Formatter
==============
This formatter outputs the issues as plain text.
:Example:
.. code-block:: none
>> Issue: [B301:blacklist_calls] Use of unsafe yaml load. Allows
instantiation of arbitrary objects. Consider yaml.safe_load().
Severity: Medium Confidence: High
CWE: CWE-20 (https://cwe.mitre.org/data/definitions/20.html)
More Info: https://bandit.readthedocs.io/en/latest/
Location: examples/yaml_load.py:5
4 ystr = yaml.dump({'a' : 1, 'b' : 2, 'c' : 3})
5 y = yaml.load(ystr)
6 yaml.dump(y)
.. versionadded:: 0.9.0
.. versionchanged:: 1.5.0
New field `more_info` added to output
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
import datetime
import logging
import sys
from bandit.core import constants
from bandit.core import docs_utils
from bandit.core import test_properties
from bandit.formatters import utils
LOG = logging.getLogger(__name__)
def get_verbose_details(manager):
bits = []
bits.append(f"Files in scope ({len(manager.files_list)}):")
tpl = "\t%s (score: {SEVERITY: %i, CONFIDENCE: %i})"
bits.extend(
[
tpl % (item, sum(score["SEVERITY"]), sum(score["CONFIDENCE"]))
for (item, score) in zip(manager.files_list, manager.scores)
]
)
bits.append(f"Files excluded ({len(manager.excluded_files)}):")
bits.extend([f"\t{fname}" for fname in manager.excluded_files])
return "\n".join([bit for bit in bits])
def get_metrics(manager):
bits = []
bits.append("\nRun metrics:")
for criteria, _ in constants.CRITERIA:
bits.append(f"\tTotal issues (by {criteria.lower()}):")
for rank in constants.RANKING:
bits.append(
"\t\t%s: %s"
% (
rank.capitalize(),
manager.metrics.data["_totals"][f"{criteria}.{rank}"],
)
)
return "\n".join([bit for bit in bits])
def _output_issue_str(
issue, indent, show_lineno=True, show_code=True, lines=-1
):
# returns a list of lines that should be added to the existing lines list
bits = []
bits.append(
f"{indent}>> Issue: [{issue.test_id}:{issue.test}] {issue.text}"
)
bits.append(
"%s Severity: %s Confidence: %s"
% (
indent,
issue.severity.capitalize(),
issue.confidence.capitalize(),
)
)
bits.append(f"{indent} CWE: {str(issue.cwe)}")
bits.append(f"{indent} More Info: {docs_utils.get_url(issue.test_id)}")
bits.append(
"%s Location: %s:%s:%s"
% (
indent,
issue.fname,
issue.lineno if show_lineno else "",
issue.col_offset if show_lineno else "",
)
)
if show_code:
bits.extend(
[indent + line for line in issue.get_code(lines, True).split("\n")]
)
return "\n".join([bit for bit in bits])
def get_results(manager, sev_level, conf_level, lines):
bits = []
issues = manager.get_issue_list(sev_level, conf_level)
baseline = not isinstance(issues, list)
candidate_indent = " " * 10
if not len(issues):
return "\tNo issues identified."
for issue in issues:
# if not a baseline or only one candidate we know the issue
if not baseline or len(issues[issue]) == 1:
bits.append(_output_issue_str(issue, "", lines=lines))
# otherwise show the finding and the candidates
else:
bits.append(
_output_issue_str(
issue, "", show_lineno=False, show_code=False
)
)
bits.append("\n-- Candidate Issues --")
for candidate in issues[issue]:
bits.append(
_output_issue_str(candidate, candidate_indent, lines=lines)
)
bits.append("\n")
bits.append("-" * 50)
return "\n".join([bit for bit in bits])
@test_properties.accepts_baseline
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Prints discovered issues in the text format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
bits = []
if not manager.quiet or manager.results_count(sev_level, conf_level):
bits.append(
f"Run started:{datetime.datetime.now(datetime.timezone.utc)}"
)
if manager.verbose:
bits.append(get_verbose_details(manager))
bits.append("\nTest results:")
bits.append(get_results(manager, sev_level, conf_level, lines))
bits.append("\nCode scanned:")
bits.append(
"\tTotal lines of code: %i"
% (manager.metrics.data["_totals"]["loc"])
)
bits.append(
"\tTotal lines skipped (#nosec): %i"
% (manager.metrics.data["_totals"]["nosec"])
)
bits.append(
"\tTotal potential issues skipped due to specifically being "
"disabled (e.g., #nosec BXXX): %i"
% (manager.metrics.data["_totals"]["skipped_tests"])
)
skipped = manager.get_skipped()
bits.append(get_metrics(manager))
bits.append(f"Files skipped ({len(skipped)}):")
bits.extend(["\t%s (%s)" % skip for skip in skipped])
result = "\n".join([bit for bit in bits]) + "\n"
with fileobj:
wrapped_file = utils.wrap_file_object(fileobj)
wrapped_file.write(result)
if fileobj.name != sys.stdout.name:
LOG.info("Text output written to file: %s", fileobj.name)

View File

@@ -0,0 +1,14 @@
# Copyright (c) 2016 Rackspace, Inc.
#
# SPDX-License-Identifier: Apache-2.0
"""Utility functions for formatting plugins for Bandit."""
import io
def wrap_file_object(fileobj):
"""If the fileobj passed in cannot handle text, use TextIOWrapper
to handle the conversion.
"""
if isinstance(fileobj, io.TextIOBase):
return fileobj
return io.TextIOWrapper(fileobj)

View File

@@ -0,0 +1,97 @@
#
# SPDX-License-Identifier: Apache-2.0
r"""
=============
XML Formatter
=============
This formatter outputs the issues as XML.
:Example:
.. code-block:: xml
<?xml version='1.0' encoding='utf-8'?>
<testsuite name="bandit" tests="1"><testcase
classname="examples/yaml_load.py" name="blacklist_calls"><error
message="Use of unsafe yaml load. Allows instantiation of arbitrary
objects. Consider yaml.safe_load().&#10;" type="MEDIUM"
more_info="https://bandit.readthedocs.io/en/latest/">Test ID: B301
Severity: MEDIUM Confidence: HIGH
CWE: CWE-20 (https://cwe.mitre.org/data/definitions/20.html) Use of unsafe
yaml load.
Allows instantiation of arbitrary objects. Consider yaml.safe_load().
Location examples/yaml_load.py:5</error></testcase></testsuite>
.. versionadded:: 0.12.0
.. versionchanged:: 1.5.0
New field `more_info` added to output
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
import logging
import sys
from xml.etree import ElementTree as ET # nosec: B405
from bandit.core import docs_utils
LOG = logging.getLogger(__name__)
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Prints issues in XML format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
issues = manager.get_issue_list(sev_level=sev_level, conf_level=conf_level)
root = ET.Element("testsuite", name="bandit", tests=str(len(issues)))
for issue in issues:
test = issue.test
testcase = ET.SubElement(
root, "testcase", classname=issue.fname, name=test
)
text = (
"Test ID: %s Severity: %s Confidence: %s\nCWE: %s\n%s\n"
"Location %s:%s"
)
text %= (
issue.test_id,
issue.severity,
issue.confidence,
issue.cwe,
issue.text,
issue.fname,
issue.lineno,
)
ET.SubElement(
testcase,
"error",
more_info=docs_utils.get_url(issue.test_id),
type=issue.severity,
message=issue.text,
).text = text
tree = ET.ElementTree(root)
if fileobj.name == sys.stdout.name:
fileobj = sys.stdout.buffer
elif fileobj.mode == "w":
fileobj.close()
fileobj = open(fileobj.name, "wb")
with fileobj:
tree.write(fileobj, encoding="utf-8", xml_declaration=True)
if fileobj.name != sys.stdout.name:
LOG.info("XML output written to file: %s", fileobj.name)

View File

@@ -0,0 +1,126 @@
# Copyright (c) 2017 VMware, Inc.
#
# SPDX-License-Identifier: Apache-2.0
r"""
==============
YAML Formatter
==============
This formatter outputs the issues in a yaml format.
:Example:
.. code-block:: none
errors: []
generated_at: '2017-03-09T22:29:30Z'
metrics:
_totals:
CONFIDENCE.HIGH: 1
CONFIDENCE.LOW: 0
CONFIDENCE.MEDIUM: 0
CONFIDENCE.UNDEFINED: 0
SEVERITY.HIGH: 0
SEVERITY.LOW: 0
SEVERITY.MEDIUM: 1
SEVERITY.UNDEFINED: 0
loc: 9
nosec: 0
examples/yaml_load.py:
CONFIDENCE.HIGH: 1
CONFIDENCE.LOW: 0
CONFIDENCE.MEDIUM: 0
CONFIDENCE.UNDEFINED: 0
SEVERITY.HIGH: 0
SEVERITY.LOW: 0
SEVERITY.MEDIUM: 1
SEVERITY.UNDEFINED: 0
loc: 9
nosec: 0
results:
- code: '5 ystr = yaml.dump({''a'' : 1, ''b'' : 2, ''c'' : 3})\n
6 y = yaml.load(ystr)\n7 yaml.dump(y)\n'
filename: examples/yaml_load.py
issue_confidence: HIGH
issue_severity: MEDIUM
issue_text: Use of unsafe yaml load. Allows instantiation of arbitrary
objects.
Consider yaml.safe_load().
line_number: 6
line_range:
- 6
more_info: https://bandit.readthedocs.io/en/latest/
test_id: B506
test_name: yaml_load
.. versionadded:: 1.5.0
.. versionchanged:: 1.7.3
New field `CWE` added to output
"""
# Necessary for this formatter to work when imported on Python 2. Importing
# the standard library's yaml module conflicts with the name of this module.
import datetime
import logging
import operator
import sys
import yaml
from bandit.core import docs_utils
LOG = logging.getLogger(__name__)
def report(manager, fileobj, sev_level, conf_level, lines=-1):
"""Prints issues in YAML format
:param manager: the bandit manager object
:param fileobj: The output file object, which may be sys.stdout
:param sev_level: Filtering severity level
:param conf_level: Filtering confidence level
:param lines: Number of lines to report, -1 for all
"""
machine_output = {"results": [], "errors": []}
for fname, reason in manager.get_skipped():
machine_output["errors"].append({"filename": fname, "reason": reason})
results = manager.get_issue_list(
sev_level=sev_level, conf_level=conf_level
)
collector = [r.as_dict(max_lines=lines) for r in results]
for elem in collector:
elem["more_info"] = docs_utils.get_url(elem["test_id"])
itemgetter = operator.itemgetter
if manager.agg_type == "vuln":
machine_output["results"] = sorted(
collector, key=itemgetter("test_name")
)
else:
machine_output["results"] = sorted(
collector, key=itemgetter("filename")
)
machine_output["metrics"] = manager.metrics.data
for result in machine_output["results"]:
if "code" in result:
code = result["code"].replace("\n", "\\n")
result["code"] = code
# timezone agnostic format
TS_FORMAT = "%Y-%m-%dT%H:%M:%SZ"
time_string = datetime.datetime.now(datetime.timezone.utc).strftime(
TS_FORMAT
)
machine_output["generated_at"] = time_string
yaml.safe_dump(machine_output, fileobj, default_flow_style=False)
if fileobj.name != sys.stdout.name:
LOG.info("YAML output written to file: %s", fileobj.name)

Some files were not shown because too many files have changed in this diff Show More