Vanguard

Description

Vanguard is a comprehensive web penetration testing tool written in Perl that identifies vulnerabilities in web applications.

Features

Main application features:

• Fully Configurable
• WebCrawlers crawl all open HTTP and HTTPS ports output from nmap
• LibWhisker2 For HTTP IDS Evasion (Same options as nikto)
• Tests via GET,POST, and COOKIE

Web penetration tests:

• SQL injection (This test is signature free!)
• LDAP Injection
• XSS
• File inclusion
• Command Injection

Limitations

A private, more featureful version does exist, but is being ported to ruby in order to solve a threading concurrency issue we had while programming the user interface. More information on this at a later date.

• Vanguard's public release can identify but not exploit vulnerabilities.
• This release does NOT crawl AJAX pages. If you're looking for something full featured, this public release isn't for you.
• The SQL injection test depends on checksums. This can be changed if the optimization method isn't working due to other forms of dynamic content (e.g. sentence spinners)

Usage

 perl scan.pl -h [hostname] -e [evasion option]

Installation

Application Dependencies

Notice: You must run this application as root.

You must have nmap from http://nmap.org installed to run this application correctly.

Protip: You can undo the root requirement by removing the check for root and modifying the nmap configuration.

Perl Dependencies

LibWhisker2 requires Net::SSLeay. You may need to get this from cpan, compile it in, or install it from your distribution's package manager.

• YAML
• Clone
  Notice: You can install these libraries with cpan.

Configuration

Main Configuration

This is the configuration in config.yml.

Vanguard has a very simple set of configuration options. --- rewrite: 0 use_whitelist: 1 module_whitelist: - WEBAPPS - SHELL - NMAP - CRAWL - XSS_GET - SQL_GET - LFI_GET - RCI_GET rewrite: Specifies whether or not to use the expiremental mod_rewrite data tampering engine. 1 for enabled, 0 for disabled. use_whitelist: Specifies whether or not to use the module_whitelist settings. 1 for enabled, 0 for disabled. When disabled, vanguard will attempt to load every module in the /modules/* directories. module_whitelist: The module_whitelist allows you to specify by directory name in the modules/recon, modules/api and modules/test directories.

WebCrawler

This configuration is located in /modules/recon/CRAWL/conf.yml

The only option for the webcrawler is the crawl depth. --- depth: 20 depth: The number of links to follow recursively from each page. A higher or lower setting will yield a slower or faster scan, however more or less thorough, respectively.

Nmap Module

This configuration is located in /modules/recon/NMAP/conf.yml

This code is currently only used to specify the flags used on nmap at runtime. Read the module's code for more information. --- flags: "-P0 --defeat-rst-ratelimit -sSV -F" flags: The command line flag arguments

Notice: See the nmap manual for additional information.

Protip: The S in -sSV is the reason this scan requires root.

Local File Inclusion

You can find this configuration in /modules/test/LFI_*/conf.yml.

The file inclusion test is relatively simple. --- lfi_test: '%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/%2e%2e/etc/passwd' lfi_match: "root\:.\:0\:0" lfi_exits: - '%00' - lfi_test: This is a local filename to look for on the remote host. Most linux hosts will allow access to /etc/passwd, but the user can specify anything here. lfi_match: Contents inside of the file (in regular expression format) to confirm file inclusion. lfi_exits: Sometimes this test does not require an exit, other times it does (like a null byte). These strings are appended to the end of the filename during testing.

Protip: It can be a good idea to use file extensions or language codes (e.g. %00en, %00php) as exits in this configuration file.

LDAP

These configurations are located in /modules/test/LDAP_*/conf.yml.

The LDAP test is similar to the SQL test. --- ldap_true: - ")(&" - ")(&)(" ldap_false: - ")(|" - ")(|)(" ldap_true: This ldap should return either the same result or all of the results, you can mess around with this by replacing & with * and changing around the code in the module. ldap_false: This should be valid ldap that returns no results.

Remote File Inclusion

You can find this configuration in /modules/test/RFI_*/conf.yml.

This can be set to any site specified. --- rfi_test: http://asdf.com/ rfi_match: 89asdf.gi rfi_test: A remote file, page, or site to include. rfi_match: A string inside of the test file used for confirmation during testing.

Protip: Randomize these options to evade signature based heuristics.

Command Injection

This configuration is located in /modules/test/RCI_*/conf.yml.

The escape strings used to inject commands are the only configuration options for this module. --- entries: - '|' - ';' - '&&' entries: Each of these comes before an attempt to inject a command.

Protip: Sometimes you may want a single or double quote (%27 or %22) before the escape string to escape any quote trickery.

SQL injection

You can find these configuration files & options in modules/test/SQL_GET/conf.yml and modules/test/SQL_POST/conf.yml.

This file defines several variables for automated SQL injection testing. --- sql_spacers: - '%20' - '+' sql_entries: - "" - '%27' - '%bf%27' sql_exits: - -- - /* - )-- - )/* sql_spacers: Different database backends parse spaces differently. Microsoft Access, for example, prefers '+' to be used as a "space" character, however for most linux based database solutions, a simple uri encoded space (%20) will suffice. sql_entries: Because SQL injection utilizes an escape string, entries are used to define what escape string is necessary. Some injections are mis-handled integers and do not require this, hence we have an empty entry. The next entry is a url encoded single quote, and the final entry is an escape string affecting non-utf8 character encodings to bypass php's addslashes() function. sql_exits: Different SQL databases use a different syntax for commenting code. Some engines begin comments with -- while others begin comments with /*. This is used to cancel the rest of the query during the truth/false test. Because some query inputs are nested in parenthesis, the last two exits are listed as fallbacks.

Download

The end-user is liable for his-or her own actions with the use of this software. Running this against a system you do not own without written authorization is a criminal act.

Other Tools

• Lfi_autopwn.pl
• GScrape
• MySql 5 Enumeration
• Wordpress Fingerprinting

---

See Also

• Web Exploitation