Unleash the Power of Dockerized Analyzers in 5 Minutes

One of the big improvements you’ll notice in Cortex 3 is the support for dockerized analyzers. And amongst some of their benefits, the installation process has been significantly simplified. So let’s assume you do not want to bang your head against Python, or other library dependencies. Then read one for a way to set up analyzers and run them quickly.

The following instructions have been tested on Ubuntu 18.04. If you already have a Cortex instance up and running, you can jump directly to the docker installation section below.

Install System Packages

Prerequisites

  • Ensure your system contains the required packages:
sudo apt-get install apt-transport-https ca-certificates curl software-properties-common
  • Oh, and Cortex should have access to the internet 😉

Cortex

Configure your system to install packages from TheHive-Project repositories.

echo 'deb https://dl.bintray.com/thehive-project/debian-stable any main' | sudo tee -a /etc/apt/sources.list.d/thehive-project.list
sudo apt-key adv --keyserver hkp://pgp.mit.edu --recv-key 562CBC1C

Install Cortex:

sudo apt-get update
sudo apt-get install cortex

The Cortex files will be put in /opt/cortex. The Cortex binary will be executed using the cortex account.

The configuration file is /etc/cortex/application.conf.

The Cortex log file is /var/log/cortex/application.log.

Docker

We recommend reading the Docker documentation to install the software.

For Ubuntu 18.04, you can run the following commands:

wget -O- https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu bionic stable"
sudo apt-get update
sudo apt-get install docker-ce

Use Docker with Cortex

Start by giving the cortex account the permission to run Docker.

usermod -a -G docker cortex

Next, in the Cortex configuration file (/etc/cortex/application.conf) ensure jobs will be executed using Docker:

job {
runner = [docker]
}

Follow the Standard Analyzer Catalog

If you are a complete Cortex noob, please read the Quick Start Guide.

Configure Cortex to get the catalog with up-to-date versions of analyzers. Edit /etc/cortex/application.conf to set the catalog URLs in the analyzer configuration section as shown below:

## ANALYZERS
#
analyzer {
  urls = ["https://dl.bintray.com/thehive-project/cortexneurons/analyzers.json"]
[..]
}

Once Cortex is configured, restart the service with the following command, wait a few seconds and you should be able to connect to Cortex on https://<yourCortexHost&gt;:9001 et voilà!

sudo service restart cortex

Important Note: The catalog analyzers.json contains information regarding versions of analyzers we consider stable and that are updated with bug fixes. This is typically synchronised with our master branch on Github. When you are using this catalog, you are de facto benefiting from the latest analyzer updates without needing to refresh anything in Cortex or setup again the configuration to get the latest version.

We also provide two additional catalogs:

  • analyzers-stable.json which strictly follows versions of analyzers if you do not want any uncontrolled updates. What does that mean in practice? You will have to click on the Refresh button in Cortex to update your analyzers, disable old ones and enable new versions. Moreover, you will also have to setup again their configuration. Typically, if you installed and setup Cortex with this catalog and the current version of FileInfo analyzers is 6.0, you won’t benefit from the next version, let’s say 6.1, unless you refresh Cortex.
  • analyzers-devel.jsonwhich contains information about new analyzers or version of analyzers that contains code that has been reviewed but not tested enough (or even not tested at all at times) to be deemed ready for production environments. This is typically synchronized with the develop branch of our Github repository.

Same goes for responders. All available catalogs for Cortex are published on bintray so you can choose the one that better fits your needs (or your risk/gambling profile :p).

Searching for an Elastic? Here, Take 6!

As we announced on May 14, 2019, we have been working very hard to add Elasticsearch 6 support to TheHive and Cortex as Elasticsearch 5.x went the way of the dodo when Elastic plugged life support off this venerable version. We also took this occasion to upgrade AngularJS and its sub projects to 1.7.8, the latest 1.x version as of this writing. Additionally, Grunt build dependencies have also been updated to their latest compatible versions.

It took us more time than initially foreseen but hey, we all love deadlines. We all love the whooshing noise they make as they go by.

TheHive 3.4.0-RC1 and Cortex 3.0.0-RC3 are now available on every Internet pipe near you and before you take them for a spin to help us identify any issues to make the stable releases rock-solid, let us walk you through some important information. Relax and grab a drink (and send good wine our way, we can always use some!).

Source: https://dilbert.com/strip/1995-11-10

TheHive 3.4.0-RC1

In addition to ES5 and 6 support and the update of AngularJS, this version corrects a few bugs that were identified in the latest stable version (3.3.1) and adds a few features. The most important one in our opinion is the ability to import a file from a Cortex report. This requires Cortex 3.0.0-RC3. The full list of changes is available at the following location.

Prior to migrating to 3.4.0-RC1, please read the migration guide.

Cortex 3.0.0-RC3

ES5 and ES6 support, AngularJS et cetera et cetera. Well you know the song right? Not quite as Cortex 3.0.0 significantly facilitates analyzer and responder installation and updates, thanks to Docker as we touched upon in a blog post earlier this year.

As detailed in the Cortex migration guide, which we recommend you read thoroughly, you can migrate from Cortex 2 and keep using analyzers and responders the same way (using processes), use the new Docker-based analyzers and responders or mix and match between running processes and docker containers (but then, you gotta pay extra attention to configure properly which analyzer/responder runs in which fashion).

Moreover, if you use the new dockerised analyzers and responders, you will be able to choose if you want to have them autoupdated (that’s the default behaviour) and if so, pick the bleeding edge, potentially buggy versions, the minor releases or, if you are risk-averse, stick with stable ones.

Cortex 3.0.0-RC3 also adds the ability to retrieve files resulting from analyzer jobs and last but not least, corrects an information disclosure bug that allowed non-admin users to retrieve the details of other users through the API. The vulnerability was reported by Adam Maris so kudos to him!

Warning: Regressions Ahead!

As outlined in our previous post about these new versions:

  • TheHive 3.4.0-RC1 and Cortex 3.0.0-RC3 use HTTP transport (9200/tcp by default) to connect to Elasticsearch instead of its native binary protocol (9300/tcp by default).
  • SSL/TLS, including when using a client certificate, can be configured to connect securely to ES. However this has not been tested yet.
  • Support of X-Pack and Search Guard is discontinued for anything but basic and SSL client authentication, which would still work.

Caution: Performance May Take a Hit!

The parent-child relationships we use behind the scene in Elasticsearch could make queries significantly slower with ES 6 and in our limited testing, we had the impression that performance took a hit. So please be cautious there and we’d be grateful if you could report any sluggishness you notice during your tests of the new versions with ES6.

UnshortenLink, SSRF and Cortex-Analyzers 1.15.2

We urge our fellow fighters of cybercrime and funny and not-so-funny animals-in-disguise, to update their Cortex analyzers to the latest 1.15.2 version which plugs a serious vulnerability in the Unshortenlink analyzer and fixes several bugs.

Unshortenlink Vulnerability

On Jan 24, Alexandre Basquin discovered a SSRF vulnerability in the Unshortenlink 1.0 analyzer. By exploiting it, an attacker which has access to a Cortex instance with an analyze role can scan the ports of localhost and possibly of all the hosts reachable by Cortex. This could be automated through Cortex4py by making repetitive calls to the API and thus scanning multiple ports & hosts. In essence, the attacker can perform reconnaissance thanks to Unshortenlink and gain knowledge on which ports are open and which aren’t.

Version 1.1 of Unshortenlink, included in Cortex-Analyzers 1.15.2, plugs this vulnerability by disabling submissions containing IPs & ports. Only URLs are now accepted.

Bug Fixes

Additionally, Cortex-Analyzers 1.15.2 corrects the following bugs:

  • #416: the Mailer responder now supports UTF-8 encoding.
  • #410: remove wrapping of the results produced by crt.sh as a list which ends up breaking the output of the report template. Contributed by Thomas Kastner.
  • #409: enum is not required for Python 3.4+ to make the MISP Search analyzer work.
  • #408: FileInfo’s Manalyze plugin did not work after Manalyze renamed plugin_btcaddress to plugin_cryptoaddress.
  • #406: fix a broken link in the Cymon_Check_IP report, submitted by Manabu Niseki.
  • #313: fix wrong file handling in OTXQuery.

How to Update your Analyzers?

To update your analyzers to version 1.15.2, please read the Updating section of the Cortex installation guide.

Since the version numbers of few analyzers have changed, you’ll need to connect as an orgadmin to Cortex to refresh the analyzer list, disabling old ones and enabling the new versions.

TheHive Report Templates

If you are using TheHive, you must import the new report templates in your instance as follows:

  • download the updated package
  • log in TheHive using an administrator account
  • go to Admin > Report templates menu
  • click on Import templates button and select the downloaded package

Wait a Minute! Where’s the Blog Post about 1.15.1?

Good catch! There was no blog post about Cortex-Analyzers 1.15.1. Call us lazy but there was nothing Earth-shattering, pole-shifting in that release. Nonetheless, here are a list of fixes included in that release and from which you’ll inherit automatically if you update 1.15.0 to 1.15.2 directly:

  • #402: Malwares analyzer code relied on functionality that’s only available in Python 3.7+. It has been fixed to work with 3.4 and up thanks to the work of Arcuri Davide.
  • #404: fixes a bad folder renaming in the HIBP (Have I Been Pwned) analyzer.
  • #398: MISP Search analyzer wouldn’t run without the enum dependency. Contributed by Mars Huang. Later improved in 1.15.2 (see above).

No Luck Luke?

If something does not work as expected, keep calm and ask on our user forum, reach out to fellow users on Gitter, or send us an email at support@thehive-project.org.

TheHive 3.3-RC2, Hot out of the Oven

TheHive Project’s code Chefs, dressed in their outright haute cuisine outfit, including the traditional toque blanche, have been quite busy lately, working on dockerizing all the Cortex analyzers (more on this later in an upcoming post), and doing tedious work to prepare the replacement of Elasticsearch by a GraphDB which will help us finally release much-awaited features such as multi-tenancy, delayed for way too many months (yeah, yeah, don’t chastise them but feel free to help them). In the meantime, they found enough bandwidth to release a new major version of TheHive.

Version 3.3, currently a release candidate includes several bug fixes and many new features as outlined below. Please note that TheHive 3.3-RC2 is beta software. As all our other release candidates, you can grab it from the pre-release, beta repositories. As usual, we would truly appreciate your help making it a great stable release by testing it as thoroughly as possible and reporting back any bugs or issues you encounter so we can address them before the final release.

Check TheHive Installation guide for further details.

Wait! Where’s RC1?

TheHive 3.3-RC1 was very short-lived. Few hours after its release, and thanks to Chris (a.k.a. crackytsi on GitHub), Thomas Franco, our back-end mastermind, discovered an issue with the Debian 8 and Debian 9 packages.

New Features

  • #836: add a new exportCaseTags parameter to the MISP configuration section. If set to true, all the tags associated with a case will be exported along with it to MISP.
  • #861: add support for Java higher than 8, such as OpenJDK 11.
  • #271: bulk merge alerts into a case. Select multiple alerts at once and create a single case out of them or merge them into an existing case using its ID.
  • #824: add ability to sort alerts by reference, status, type, source…
  • #826: when previewing an alert, there are sometimes no overlap with an existing case. However, an analyst might already know, thanks to HI (Human Intelligence), that the alert should be merged into a specific case. This is now possible thanks to a new button.
  • #769: improve case template selection for case creation. If you have defined a large set of case templates, you will be able to sort/filter to find the case you want to use when creating a New Case.
New case template selector
  • #657: add observable tags auto-completion. Contributed by Tyler Chong (Thanks!).
Observable tag auto-completion

Fixed Bugs

  • #864: do not return a session cookie when making an API call.
  • #856: there was a bug where after a followed alert PATCH, if the alert has already been promoted to a case, the case is not updated. Now, if the alert has follow=true, if it gets updated, its status is set to Updated and the related case is updated too.
  • #845: assigned but unstarted tasks were not showing up in My Tasks.
  • #844: enable user account locking through the Delete API endpoint.

Stuck?

Something does not work as expected? You have troubles installing or upgrading? No worries, please join our user forum, contact us on Gitter, or send us an email at support@thehive-project.org. We are here to help.

TheHive 3.2.0-RC1: The MISP Love Edition

Guess what? Our integration with MISP, the de facto standard for threat sharing, has just gotten better with our latest beta release: TheHive 3.2.0-RC1.

While you could synchronize TheHive with one or multiple MISP instances in earlier versions and select events using filters like their age, the number of attributes they contain or exclude those which are created by specific organisations or contain one or several black-listed tags, 3.2.0-RC1 adds the ability to whitelist tags, thus limiting the events that would show up in TheHive’s Alerts pane to only those which have been tagged with labels your SOC/CSIRT/CERT needs to act on. This can be very useful for example if your Cyber Threat Intelligence analysts pre-select or create events in MISP and tag for SOC consumption those that need to be acted on.

The Hive - Logo - Schéma - V1_Plan de travail 1.png
TheHive, Cortex, MISP: The Power DFIR & CTI Trio

To use this feature, use the whitelist.tags parameter in the MISP section of TheHive’s application.conf as described in the documentation.

This new version also adds the ability to create dashboards out of responder actions, log responder operations, and offers a confirmation dialog before running a responder to avoid noob over-clicks and errors made by seasoned incident handlers running low on caffeine.

TheHive 3.2.0-RC1 will also show you the description of an observable if any while hovering over one in the Observables tab. You can also see observable tags when previewing an alert in the Alerts pane.

Last but not least, some users reported severe problems when they enabled TLS/SSL directly on TheHive without resorting to a reverse proxy such as NGINX. Blame that on the crappy TLS support in Play framework ;-). So we highly recommend using a reverse proxy for that purpose, and delegate authentication to it if you are relying on X.509 authentication, as TheHive 3.2.0-RC1 allows you to. Please check the Single Sign-On on TheHive with X.509 Certificates guide for further information.

For additional details on this release, please check the full changelog.

Warning Capt’n Robinson!

The RC in 3.2.0-RC1 stands for Release Candidate. Please help us make a great stable release out of it by testing it as thoroughly as possible and reporting back any bugs or issues you encounter so we can address them before the final release. You’ll find this release candidate in the pre-release, beta repositories.

Please check TheHive Installation guide for further details.

You got a problem?

Something does not work as expected? You have troubles installing or upgrading? Spotted new bugs? No worries, please open issues on GitHub or comment on existing ones, join our user forum, contact us on Gitter, or send us an email at support@thehive-project.org. We are here to help.

Cortex 101: Dissecting Observables a Hundred Ways

TheHive Project’s Code Chefs are glad to announce that, thanks to the precious contributions of the user community, Cortex has broken the one hundred analyzer mark.

Cortex-Analyzers version 1.14.0 is out and includes new analyzers, some improvements and some bug fixes.

New Analyzers

New and enhanced analyzers, described below, are:

  • Cisco Investigate by Cisco Umbrella Research @opendns
  • Proofpoint Forensics Lookup by Emmanuel Torquato
  • Proofpoint Threat Insight Forensics Analyzer by Emmanuel Torquato
  • RecordedFuture by jojoob
  • urlscan.io search by ninoseki
  • Google DNS over HTTP by 0xswitch
  • RTF files support in FileInfo by Nils Kuhnert
  • Datascan and Inetnum flavors in Onyphe analyzer by Pierre Baudry and Adrien Barchapt

Again, huge thanks for the awesome work that has been performed by all our contributors!

Investigate

Cisco Umbrella Investigate provides threat intelligence about domains and IP addresses accross the Internet. The analyzer can be used to query the Cisco Umbrella (formerly OpenDNS) API and get information about an IP or a domain name. An API key is required to use this analyzer.

Results are displayed in TheHive in the following manner:

sc-investigate-short

sc-investigate-long
Cisco Investigate: short and long reports

Proofpoint Forensics Lookup

According to Proofpoint’s website, the Forensics API allows insight in detailed forensic evidences about individual threats or compaigns. The analyzer can be used to check observables against given indicators of compromise stored in the ProofPoint service.

Unfortunately, there are currently no sample report screenshots available, because TheHive’s Core Team does not have access to Proofpoint services. Also, due to the same reason, this analyzer could not be tested by us. If you have access to the service and can test the analyzer and/or provide report screenshots, please let us know.

RecordedFuture

This analyzer lets you get the latest risk data from RecordedFuture for a hash, domain or an IP address. It can be used to query the API and get information. An API key is required to use this analyzer.

Results are displayed in TheHive in the following manner:

sc-recordedfuture-short

sc-recordedfuture-long
RecordedFuture: short and long reports

Urlscan.io search

Urlscan.io is a service that scans and analyzes websites. Submitted pages will be browsed like a regular user would do and every activity gets recorded. The analyzer submitted by ninoseki queries urlscan without initiating a scan which would be publicly visible on the website. Accepted datatypes for this analyzer are URL, domain, hash and IP.

The templates which display the results of the analyzer look like the following screenshots:

Screenshot from 2018-10-25 09-55-34

Screenshot from 2018-10-25 09-48-58
Urlscan.io: short and long reports

Google DNS over HTTP

This analyzer provides DNS information for an IP, a domain or a FQDN by making calls to Google DNS-over-HTTP (DoH). No API key is required.

Results are displayed in TheHive in the following manner:sc-googleDNS-short

sc-googleDNS-long
Google DNS: short and long reports

RTF files support in FileInfo

The FileInfo meta analyzer has been improved and now leverages the rtfobj tool provided in the Oletools suite by Decalage.

Results are displayed in TheHive in the following manner:

sc-fileinfo_RTF-short

sc-rtfobj-long
FileInfo with rtfobj: short and long reports

Datascan and Inetnum flavors in Onyphe analyzer

The Onyphe analyzer has been enhanced with two new flavors. Datascan provides information about known open ports on a specific IP, and Inetnum enumerates all known network information about the analyzed IP address.

An API key is required to use the analyzer and can be obtained by creating an account on the Onyphe website.

Results are displayed in TheHive in the following manner:

sc-onyphe-inetnum-short

sc-onyphe_inetnum-long
Onyphe  Inetnum: short and long reports

sc-onyphe_datascan-short

sc-onyphe_datascan-long
Onyphe Datascan: long and short reports

Bug fixes and enhancements

  • #248: Improve error msg when VT Get Report does not have an entry for
  • #323: Fix an issue with HybridAnalysis analyzer filenames handler
  • #329: Enhance PassiveTotal Passive DNS report
  • #348: Fix a typo in URLhaus’s long.html
  • #356 : Force the use of Python3 in MISP analyzer
  • #362: Fix file not found issue and empty result set in CERT.at passive DNS analyzer

Get It While Supply Lasts!

Each analyzer comes with its own, pip compatible requirements.txt file. To update your Cortex analyzers to 1.14.0, run the following commands:

cd path/to/Cortex-Analyzers
git pull
for I in analyzers/*/requirements.txt; do sudo -H pip2 install -U -r $I || true; done && \
for I in analyzers/*/requirements.txt; do sudo -H pip3 install -U -r $I || true; done

Once done, do not forget to login to Cortex as an orgadmin and click on the Refresh Analyzers button. Refer to the online Cortex documentation for further details.

Update TheHive Report Templates

If you are using TheHive, get the latest version of  the report templates and import them into TheHive.

Running Into Trouble?

Shall you encounter any difficulty, please join our  user forum, contact us on Gitter, or send us an email at support@thehive-project.org. We will be more than happy to help!

TheHive 3.1.2 & Cortex 2.1.2 Released

We could not leave for the week-end without issuing a minor release or two so here we go.

TheHive 3.1.2

Starting from TheHive 3.0.1, an administrator has the ability to configure Cortex job polling by defining the time between two polls thanks to the cortex.refreshDelay parameter as well as the number of consecutive failures before giving up (via cortex.MaxRetryOnError). However, these settings prevent the service from starting correctly. TheHive 3.1.2 corrects this issue.

Cortex 2.1.2

When running a job in Cortex with the exact same details, the function findSimilarJob is called. It should return results from any previous jobs, but in the latest versions (2.1.0, 2.1.1) it does not because of a change that went past our QA.

In a similar fashion, the GUI search function was broken. Cortex 2.1.2 fixes both issues.

Excuse my French but I Need Help

Keep calm. We speak French. So if you encounter any difficulty to update TheHive or Cortex, please join our  user forum, contact us on Gitter, or send us an email at support@thehive-project.org. We are always ready to help as does our user community.