Bandit：一款Python代码安全漏洞检测工具

工具介绍

Bandit这款工具可以用来搜索Python代码中常见的安全问题，在检测过程中，Bandit会对每一份Python代码文件进行处理，并构建AST，然后针对每一个AST节点运行相应的检测插件。完成安全扫描之后，Bandit会直接给用户生成检测报告。

工具安装

Bandit使用PyPI来进行分发，建议广大用户直接使用pip来安装Bandit。

创建虚拟环境（可选）：

virtualenv bandit-env

安装Bandit：

pip install bandit
# Or if you're working with a Python 3 project
pip3 install bandit

运行Bandit：

bandit -r path/to/your/code

用户还可以使用源码文件直接安装Bandit，先从PyPI下载原tarball，然后运行下列命令：

python setup.py install

工具使用

节点树使用样例：

bandit -r ~/your_repos/project

examples/目录遍历使用样例，显示三行内容，并只报告高危问题：

bandit examples/*.py -n 3 –lll

Bandit还能够结合配置参数一起运行，运行下列命令即可使用ShellInjection来对examples目录运行安全扫描：

bandit examples/*.py -p ShellInjection

Bandit还支持使用标准输入模式来扫描指定行数的代码：

cat examples/imports.py | bandit –

使用样例：

$bandit -h
usage:bandit [-h] [-r] [-a {file,vuln}] [-n CONTEXT_LINES] [-c CONFIG_FILE]
              [-p PROFILE] [-t TESTS] [-sSKIPS] [-l] [-i]
              [-f{csv,custom,html,json,screen,txt,xml,yaml}]
              [--msg-template MSG_TEMPLATE] [-o[OUTPUT_FILE]] [-v] [-d] [-q]
              [--ignore-nosec] [-x EXCLUDED_PATHS] [-bBASELINE]
              [--ini INI_PATH] [--version]
              [targets [targets ...]]
 
Bandit- a Python source code security analyzer
 
positionalarguments:
  targets               source file(s) or directory(s)to be tested
 
optionalarguments:
  -h, --help            show this help message and exit
  -r, --recursive       find and process files in subdirectories
  -a {file,vuln}, --aggregate {file,vuln}
                        aggregate output byvulnerability (default) or by
                        filename
  -n CONTEXT_LINES, --number CONTEXT_LINES
                        maximum number of codelines to output for each issue
  -c CONFIG_FILE, --configfile CONFIG_FILE
                        optional config file touse for selecting plugins and
                        overriding defaults
  -p PROFILE, --profile PROFILE
                        profile to use(defaults to executing all tests)
  -t TESTS, --tests TESTS
                        comma-separated list oftest IDs to run

-s SKIPS, --skip SKIPS

                        comma-separated list oftest IDs to skip

  -l, --level           report only issues of a givenseverity level or higher

                        (-l for LOW, -ll for MEDIUM, -lll forHIGH)

  -i, --confidence      report only issues of a given confidencelevel or

                        higher (-i for LOW, -iifor MEDIUM, -iii for HIGH)

  -f{csv,custom,html,json,screen,txt,xml,yaml}, --format{csv,custom,html,json,screen,txt,xml,yaml}

                        specify output format

  --msg-template MSG_TEMPLATE

                        specify output messagetemplate (only usable with

                        --format custom), seeCUSTOM FORMAT section for list

                        of available values

  -o [OUTPUT_FILE], --output [OUTPUT_FILE]

                        write report tofilename

  -v, --verbose         output extra information like excludedand included

                        files

  -d, --debug           turn on debug mode

  -q, --quiet, --silent

                        only show output in thecase of an error

  --ignore-nosec        do not skip lines with # nosec comments

  -x EXCLUDED_PATHS, --exclude EXCLUDED_PATHS

                        comma-separated list ofpaths (glob patterns supported)

                        to exclude from scan(note that these are in addition

                        to the excluded pathsprovided in the config file)

  -b BASELINE, --baseline BASELINE

                        path of a baselinereport to compare against (only

                        JSON-formatted filesare accepted)

  --ini INI_PATH        path to a .bandit file that suppliescommand line

                        arguments

  --version             show program's version number andexit

 

CUSTOMFORMATTING

-----------------

 

Availabletags:

 

    {abspath}, {relpath}, {line},  {test_id},

    {severity}, {msg}, {confidence}, {range}

 

Exampleusage:

 

    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 moreinformation about formatting style:

    https://docs.python.org/3.4/library/string.html

 

Thefollowing tests were discovered and loaded:

-----------------------------------------------

 

  B101 assert_used

  B102 exec_used

  B103 set_bad_file_permissions

  B104 hardcoded_bind_all_interfaces

  B105 hardcoded_password_string

  B106 hardcoded_password_funcarg

  B107 hardcoded_password_default

  B108 hardcoded_tmp_directory

  B110 try_except_pass

  B112 try_except_continue

  B201 flask_debug_true

  B301 pickle

  B302 marshal

  B303 md5

  B304 ciphers

  B305 cipher_modes

  B306 mktemp_q

  B307 eval

  B308 mark_safe

  B309 httpsconnection

  B310 urllib_urlopen

  B311 random

  B312 telnetlib

  B313 xml_bad_cElementTree

  B314 xml_bad_ElementTree

  B315 xml_bad_expatreader

  B316 xml_bad_expatbuilder

  B317 xml_bad_sax

  B318 xml_bad_minidom

  B319 xml_bad_pulldom

  B320 xml_bad_etree

  B321 ftplib

  B322 input

  B323 unverified_context

  B324 hashlib_new_insecure_functions

  B325 tempnam

  B401 import_telnetlib

  B402 import_ftplib

  B403 import_pickle

  B404 import_subprocess

  B405 import_xml_etree

  B406 import_xml_sax

  B407 import_xml_expat

  B408 import_xml_minidom

  B409 import_xml_pulldom

  B410 import_lxml

  B411 import_xmlrpclib

  B412 import_httpoxy

  B413 import_pycrypto

  B501 request_with_no_cert_validation

  B502 ssl_with_bad_version

  B503 ssl_with_bad_defaults

  B504 ssl_with_no_version

  B505 weak_cryptographic_key

  B506 yaml_load

  B507 ssh_no_host_key_verification

  B601 paramiko_calls

  B602 subprocess_popen_with_shell_equals_true

  B603 subprocess_without_shell_equals_true

  B604 any_other_function_with_shell_equals_true

  B605 start_process_with_a_shell

  B606 start_process_with_no_shell

  B607 start_process_with_partial_path

  B608 hardcoded_sql_expressions

  B609 linux_commands_wildcard_injection

  B610 django_extra_used

  B611 django_rawsql_used

  B701 jinja2_autoescape_false

  B702 use_of_mako_templates

  B703 django_mark_safe

基准线

Bandit允许用户指定需要进行比对的基线报告路径：

bandit -b BASELINE

这样可以帮助大家忽略某些已知问题，或者是那些你不认为是问题的“问题”。大家可以使用下列命令生成基线报告：

bandit -f json -o PATH_TO_OUTPUT_FILE

版本控制整合

安装并使用pre-commit，将下列内容添加至代码库的.pre-commit-config.yaml文件中：

repos:

-   repo: https://github.com/PyCQA/bandit

    rev: '' # Update me!

    hooks:

- id: bandit

然后运行pre-commit即可。

扩展Bandit

Bandit允许用户编写和注册扩展以实现自定义检测或格式化（Formatter）功能。Bandit可以从下列两个节点加载插件：

bandit.formatters
bandit.plugins

Formatter需要接收下列四种输入参数：

result_store:一个bandit.core.BanditResultStore实例
file_list:需要扫描检测的文件列表
scores:每个文件的扫描评分
excluded_files:列表中不需要扫描的文件

利用bandit.checks来对特定类型的AST节点进行检测扫描：

@bandit.checks('Call')
defprohibit_unsafe_deserialization(context):
    if 'unsafe_load' incontext.call_function_name_qual:
        return bandit.Issue(
            severity=bandit.HIGH,
            confidence=bandit.HIGH,
            text="Unsafe deserializationdetected."
        )

注册插件时Bandit给用户提供了两个选项：

1、 如果你直接使用了安装工具（setuptools），我们需要在setup调用中添加下列信息：

# Ifyou have an imaginary bson formatter in the bandit_bson module
# anda function called `formatter`.
entry_points={'bandit.formatters':['bson = bandit_bson:formatter']}
# Ora check for using mako templates in bandit_mako that
entry_points={'bandit.plugins':['mako = bandit_mako']}

2、 如果你使用的是pbr，你需要在setup.cfg文件中添加下列信息：

[entry_points]
bandit.formatters=
    bson= bandit_bson:formatter
bandit.plugins=
    mako = bandit_mako

项目地址

参考文档：【 最新版本 】

Bandit：【 GitHub传送门 】

漏洞提交：【 传送门 】

许可证协议

本项目遵循Apache开源许可证协议。

* 参考来源： bandit ，FB小编Alpha_h4ck编译，转载请注明来自FreeBuf.COM