Arachni - Web Application Security Scanner Framework

Version	1.0.4
Homepage	http://arachni-scanner.com
Blog	http://arachni-scanner.com/blog
Github	http://github.com/Arachni/arachni
Documentation	https://github.com/Arachni/arachni/wiki
Code Documentation	http://rubydoc.info/github/Arachni/arachni
Support	http://support.arachni-scanner.com
Author	Tasos Laskos (@Zap0tek)
Twitter	@ArachniScanner
Copyright	2010-2014 Tasos Laskos
License	Dual-licensed (Apache License v2.0/Commercial) - (see LICENSE file)

Synopsis

Arachni is an Open Source, feature-full, modular, high-performance Ruby framework aimed towards helping penetration testers and administrators evaluate the security of web applications.

It is smart, it trains itself by monitoring and learning from the web application's behavior during the scan process and is able to perform meta-analysis using a number of factors in order to correctly assess the trustworthiness of results and intelligently identify (or avoid) false-positives.

Unlike other scanners, it takes into account the dynamic nature of web applications, can detect changes caused while travelling through the paths of a web application’s cyclomatic complexity and is able to adjust itself accordingly. This way, attack/input vectors that would otherwise be undetectable by non-humans can be handled seamlessly.

Moreover, due to its integrated browser environment, it can also audit and inspect client-side code, as well as support highly complicated web applications which make heavy use of technologies such as JavaScript, HTML5, DOM manipulation and AJAX.

Finally, it is versatile enough to cover a great deal of use cases, ranging from a simple command line scanner utility, to a global high performance grid of scanners, to a Ruby library allowing for scripted audits, to a multi-user multi-scan web collaboration platform.

Note: Despite the fact that Arachni is mostly targeted towards web application security, it can easily be used for general purpose scraping, data-mining, etc. with the addition of custom components.

Arachni offers:

A stable, efficient, high-performance framework

Check, report and plugin developers are allowed to easily and quickly create and deploy their components with the minimum amount of restrictions imposed upon them, while provided with the necessary infrastructure to accomplish their goals.

Furthermore, they are encouraged to take full advantage of the Ruby language under a unified framework that will increase their productivity without stifling them or complicating their tasks.

Moreover, that same framework can be utilized as any other Ruby library and lead to the development of brand new scanners or help you create highly customized scan/audit scenarios and/or scripted scans.

Simplicity

Although some parts of the Framework are fairly complex you will never have to deal them directly. From a user’s or a component developer’s point of view everything appears simple and straight-forward all the while providing power, performance and flexibility.

From the simple command-line utility scanner to the intuitive and user-friendly Web interface and collaboration platform, Arachni follows the principle of least surprise and provides you with plenty of feedback and guidance.

In simple terms

Arachni is designed to automatically detect security issues in web applications. All it expects is the URL of the target website and after a while it will present you with its findings.

Features

General

• Cookie-jar/cookie-string support.
• Custom header support.
• SSL support.
• User Agent spoofing.
• Proxy support for SOCKS4, SOCKS4A, SOCKS5, HTTP/1.1 and HTTP/1.0.
• Proxy authentication.
• Site authentication (Automated form-based, Cookie-Jar, Basic-Digest, NTLMv1 and others).
• Automatic log-out detection and re-login during the scan (when the initial login was performed via the autologin or proxy plugins).
• Custom 404 page detection.
• UI abstraction:
  • Command-line Interface.
  • Web User Interface.
• Pause/resume functionality.
• Hibernation support -- Suspend to and restore from disk.
• High performance asynchronous HTTP requests.
  • With adjustable concurrency.
  • With the ability to auto-detect server health and adjust its concurrency automatically.
• Support for custom default input values, using pairs of patterns (to be matched against input names) and values to be used to fill in matching inputs.

Integrated browser environment

Arachni includes an integrated, real browser environment in order to provide sufficient coverage to modern web applications which make use of technologies such as HTML5, JavaScript, DOM manipulation, AJAX, etc.

In addition to the monitoring of the vanilla DOM and JavaScript environments, Arachni's browsers also hook into popular frameworks to make the logged data easier to digest:

• JQuery
• AngularJS
• More to come...

In essence, this turns Arachni into a DOM and JavaScript debugger, allowing it to monitor DOM events and JavaScript data and execution flows. As a result, not only can the system trigger and identify DOM-based issues, but it will accompany them with a great deal of information regarding the state of the page at the time.

Relevant information include:

• Page DOM, as HTML code.
  • With a list of DOM transitions required to restore the state of the page to the one at the time it was logged.
• Original DOM (i.e. prior to the action that caused the page to be logged), as HTML code.
  • With a list of DOM transitions.
• Data-flow sinks -- Each sink is a JS method which received a tainted argument.
  • Parent object of the method (ex.: DOMWindow).
  • Method signature (ex.: decodeURIComponent()).
  • Arguments list.
    • With the identified taint located recursively in the included objects.
  • Method source code.
  • JS stacktrace.
• Execution flow sinks -- Each sink is a successfully executed JS payload, as injected by the security checks.
  • Includes a JS stacktrace.
• JavaScript stack-traces include:
  • Method names.
  • Method locations.
  • Method source codes.
  • Argument lists.

In essence, you have access to roughly the same information that your favorite debugger (for example, FireBug) would provide, as if you had set a breakpoint to take place at the right time for identifying an issue.

Browser-cluster

The browser-cluster is what coordinates the browser analysis of resources and allows the system to perform operations which would normally be quite time consuming in a high-performance fashion.

Configuration options include:

• Adjustable pool-size, i.e. the amount of browser workers to utilize.
• Timeout for each job.
• Worker TTL counted in jobs -- Workers which exceed the TTL have their browser process respawned.
• Ability to disable loading images.
• Adjustable screen width and height.
  • Can be used to analyze responsive and mobile applications.

Coverage

The system can provide great coverage to modern web applications due to its integrated browser environment. This allows it to interact with complex applications that make heavy use of client-side code (like JavaScript) just like a human would.

In addition to that, it also knows about which browser state changes the application has been programmed to handle and is able to trigger them programatically in order to provide coverage for a full set of possible scenarios.

By inspecting all possible pages and their states (when using client-side code) Arachni is able to extract and audit the following elements and their inputs:

• Forms
  • Along with ones that require interaction with a real browser due to DOM events.
• Links
  • Along with ones that have client-side parameters in their fragment, i.e.: http://example.com/#/?param=val&param2=val2
  • With support for rewrite rules.
• LinkTemplates -- Allowing for extraction of arbitrary inputs from generic paths, based on user-supplied templates -- useful when rewrite rules are not available.
  • Along with ones that have client-side parameters in their URL fragments, i.e.: http://example.com/#/param/val/param2/val2
• Cookies
• Headers
• Generic client-side elements like inputs which have associated DOM events.
• AJAX-request parameters.

Open distributed architecture

• High-performance/low-bandwidth communication protocol.
  • MessagePack serialization for performance, efficiency and ease of integration with 3rd party systems.
• Remote monitoring and management of Dispatchers and Instances.
• Parallel scans -- Each scan is compartmentalized to its own OS process to take advantage of:
  • Multi-core/SMP architectures.
  • OS-level scheduling/restrictions.
  • Sandboxed failure propagation.
• Multi-Instance scans for parallelization of individual scans using multiple Instances to:
  • Take advantage of multi-core/SMP architectures.
  • Greatly diminish scan-times.
• Dispatcher Grid:
  • Self-healing.
  • Scale up/down by hot-plugging/hot-unplugging nodes.
    • Can scale up infinitely by adding nodes to increase scan capacity.
  • (Always-on) Load-balancing -- All Instances are automatically provided by the least burdened Grid member.
    • With optional per-scan opt-out/override.
  • (Optional) High-Performance mode -- Combines the resources of multiple nodes to perform multi-Instance scans.
    • Enabled on a per-scan basis.
• SSL encryption (with optional peer authentication).

Scope configuration

• Filters for redundant pages like galleries, catalogs, etc. based on regular expressions and counters.
  • Can optionally detect and ignore redundant pages automatically.
• URL exclusion filters using regular expressions.
• Page exclusion filters based on content, using regular expressions.
• URL inclusion filters using regular expressions.
• Can be forced to only follow HTTPS paths and not downgrade to HTTP.
• Can optionally follow subdomains.
• Adjustable page count limit.
• Adjustable redirect limit.
• Adjustable directory depth limit.
• Adjustable DOM depth limit.
• Adjustment using URL-rewrite rules.
• Can read paths from multiple user supplied files (to both restrict and extend the scope).

Audit

• Can audit:
  • Forms
    • Can automatically refresh nonce tokens.
    • Can submit them via the integrated browser environment.
  • Links
    • Can load them via the integrated browser environment.
  • LinkTemplates
    • Can load them via the integrated browser environment.
  • Cookies
    • Can load them via the integrated browser environment.
  • Headers
  • Generic client-side DOM elements like inputs.
• Can ignore binary/non-text pages.
• Can optionally audit elements using both GET and POST HTTP methods.
• Can optionally submit all links and forms of the page along with the cookie permutations to provide extensive cookie-audit coverage.
• Can exclude specific input vectors by name.
• Can include specific input vectors by name.

Components

Arachni is a highly modular system, employing several components of distinct types to perform its duties.

In addition to enabling or disabling the bundled components so as to adjust the system's behavior and features as needed, functionality can be extended via the addition of user-created components to suit almost every need.

Platform fingerprinters

In order to make efficient use of the available bandwidth, Arachni performs rudimentary platform fingerprinting and tailors the audit process to the server-side deployed technologies by only using applicable payloads.

Currently, the following platforms can be identified:

• Operating systems
  • BSD
  • Linux
  • Unix
  • Windows
  • Solaris
• Web servers
  • Apache
  • IIS
  • Nginx
  • Tomcat
  • Jetty
• Programming languages
  • PHP
  • ASP
  • ASPX
  • JSP
  • Python
  • Ruby
• Frameworks
  • Rack

The user also has the option of specifying extra platforms (like a DB server) in order to help the system be as efficient as possible. Alternatively, fingerprinting can be disabled altogether.

Finally, Arachni will always err on the side of caution and send all available payloads when it fails to identify specific platforms.

Checks

Checks are system components which perform security checks and log issues.

Active

Active checks engage the web application via its inputs.

• SQL injection (sql_injection) -- Error based detection.
  • Oracle
  • ColdFusion
  • InterBase
  • PostgreSQL
  • MySQL
  • MSSQL
  • EMC
  • SQLite
  • DB2
  • Informix
  • Firebird
  • SaP Max DB
  • Sybase
  • Frontbase
  • Ingres
  • HSQLDB
  • MS Access
• Blind SQL injection using differential analysis (sql_injection_differential).
• Blind SQL injection using timing attacks (sql_injection_timing).
  • MySQL
  • PostgreSQL
  • MSSQL
• NoSQL injection (no_sql_injection) -- Error based vulnerability detection.
  • MongoDB
• Blind NoSQL injection using differential analysis (no_sql_injection_differential).
• CSRF detection (csrf).
• Code injection (code_injection).
  • PHP
  • Ruby
  • Python
  • JSP
  • ASP.NET
• Blind code injection using timing attacks (code_injection_timing).
  • PHP
  • Ruby
  • Python
  • JSP
  • ASP.NET
• LDAP injection (ldap_injection).
• Path traversal (path_traversal).
  • *nix
  • Windows
  • Tomcat
• File inclusion (file_inclusion).
  • *nix
  • Windows
  • Tomcat
  • PHP
  • Perl
• Response splitting (response_splitting).
• OS command injection (os_cmd_injection).
  • *nix
  • *BSD
  • IBM AIX
  • Windows
• Blind OS command injection using timing attacks (os_cmd_injection_timing).
  • Linux
  • *BSD
  • Solaris
  • Windows
• Remote file inclusion (rfi).
• Unvalidated redirects (unvalidated_redirect).
• XPath injection (xpath_injection).
  • Generic
  • PHP
  • Java
  • dotNET
  • libXML2
• XSS (xss).
• Path XSS (xss_path).
• XSS in event attributes of HTML elements (xss_event).
• XSS in HTML tags (xss_tag).
• XSS in "script" context (xss_script_context).
• Source code disclosure (source_code_disclosure)

Passive

Passive checks look for the existence of files, folders and signatures.

• Allowed HTTP methods (allowed_methods).
• Back-up files (backup_files).
• Backup directories (backup_directories)
• Common directories (common_directories).
• Common files (common_files).
• HTTP PUT (http_put).
• Insufficient Transport Layer Protection for password forms (unencrypted_password_form).
• WebDAV detection (webdav).
• HTTP TRACE detection (xst).
• Credit Card number disclosure (credit_card).
• CVS/SVN user disclosure (cvs_svn_users).
• Private IP address disclosure (private_ip).
• Common backdoors (backdoors).
• .htaccess LIMIT misconfiguration (htaccess_limit).
• Interesting responses (interesting_responses).
• HTML object grepper (html_objects).
• E-mail address disclosure (emails).
• US Social Security Number disclosure (ssn).
• Forceful directory listing (directory_listing).
• Mixed Resource/Scripting (mixed_resource).
• Insecure cookies (insecure_cookies).
• HttpOnly cookies (http_only_cookies).
• Auto-complete for password form fields (password_autocomplete).
• Origin Spoof Access Restriction Bypass (origin_spoof_access_restriction_bypass)
• Form-based upload (form_upload)
• localstart.asp (localstart_asp)
• Cookie set for parent domain (cookie_set_for_parent_domain)
• Missing Strict-Transport-Security headers for HTTPS sites (hsts).

Reporters

• Standard output
• HTML (zip) (html).
• XML (xml).
• Text (text).
• JSON (json)
• Marshal (marshal)
• YAML (yaml)
• AFR (afr)
  • The default Arachni Framework Report format.

Plugins

Plugins add extra functionality to the system in a modular fashion, this way the core remains lean and makes it easy for anyone to add arbitrary functionality.

• Passive Proxy (proxy) -- Analyzes requests and responses between the web app and the browser assisting in AJAX audits, logging-in and/or restricting the scope of the audit.
• Form based AutoLogin (autologin).
• Dictionary attacker for HTTP Auth (http_dicattack).
• Dictionary attacker for form based authentication (form_dicattack).
• Cookie collector (cookie_collector) -- Keeps track of cookies while establishing a timeline of changes.
• WAF (Web Application Firewall) Detector (waf_detector) -- Establishes a baseline of normal behavior and uses rDiff analysis to determine if malicious inputs cause any behavioral changes.
• BeepNotify (beep_notify) -- Beeps when the scan finishes.
• EmailNotify (email_notify) -- Sends a notification (and optionally a report) over SMTP at the end of the scan.
• VectorFeed (vector_feed) -- Reads in vector data from which it creates elements to be audited. Can be used to perform extremely specialized/narrow audits on a per vector/element basis. Useful for unit-testing or a gazillion other things.
• Script (script) -- Loads and runs an external Ruby script under the scope of a plugin, used for debugging and general hackery.
• Uncommon headers (uncommon_headers) -- Logs uncommon headers.
• Content-types (content_types) -- Logs content-types of server responses aiding in the identification of interesting (possibly leaked) files.

Defaults

Default plugins will run for every scan and are placed under /plugins/defaults/.

• AutoThrottle (autothrottle) -- Dynamically adjusts HTTP throughput during the scan for maximum bandwidth utilization.
• Healthmap (healthmap) -- Generates sitemap showing the health of each crawled/audited URL
• Resolver (resolver) -- Resolves vulnerable hostnames to IP addresses.

Meta

Plugins under /plugins/defaults/meta/ perform analysis on the scan results to determine trustworthiness or just add context information or general insights.

• TimingAttacks (timing_attacks) -- Provides a notice for issues uncovered by timing attacks when the affected audited pages returned unusually high response times to begin with. It also points out the danger of DoS attacks against pages that perform heavy-duty processing.
• Discovery (discovery) -- Performs anomaly detection on issues logged by discovery checks and warns of the possibility of false positives where applicable.
• Uniformity (uniformity) -- Reports inputs that are uniformly vulnerable across a number of pages hinting to the lack of a central point of input sanitization.

Trainer subsystem

The Trainer is what enables Arachni to learn from the scan it performs and incorporate that knowledge, on the fly, for the duration of the audit.

Checks have the ability to individually force the Framework to learn from the HTTP responses they are going to induce.

However, this is usually not required since Arachni is aware of which requests are more likely to uncover new elements or attack vectors and will adapt itself accordingly.

Still, this can be an invaluable asset to Fuzzer checks.

Installation

Usage

Running the specs

You can run rake spec to run all specs or you can run them selectively using the following:

rake spec:core            # for the core libraries
rake spec:checks          # for the checks
rake spec:plugins         # for the plugins
rake spec:reports         # for the reports
rake spec:path_extractors # for the path extractors

Please be warned, the core specs will require a beast of a machine due to the necessity to test the Grid/multi-Instance features of the system.

Note: The check specs will take about 90 minutes due to the timing-attack tests.

Bug reports/Feature requests

Submit bugs using GitHub Issues and get support via the Support Portal.

Contributing

(Before starting any work, please read the instructions for working with the source code.)

We're happy to accept help from fellow code-monkeys and these are the steps you need to follow in order to contribute code:

• Fork the project.
• Start a feature branch based on the experimental branch (git checkout -b <feature-name> experimental).
• Add specs for your code.
• Run the spec suite to make sure you didn't break anything (rake spec:core for the core libs or rake spec for everything).
• Commit and push your changes.
• Issue a pull request and wait for your code to be reviewed.

License

Dual-licensed (Apache License v2.0/Commercial) -- please see the LICENSE file for more information.

Disclaimer

This is free software and you are allowed to use it as you see fit. However, neither the development team nor any of our contributors can be held responsible for your actions nor for any damage caused by the use of this software.