Automated incident response in Office 365 Advanced Threat Protection now generally available

Security teams responsible for investigating and responding to incidents often deal with a massive number of signals from widely disparate sources. As a result, rapid and efficient incident response continues to be the biggest challenge facing security teams today. The sheer volume of these signals, combined with an ever-growing digital estate of organizations, means that a lot of critical alerts miss getting the timely attention they deserve. Security teams need help to scale better, be more efficient, focus on the right issues, and deal with incidents in a timely manner.

This is why I’m excited to announce the general availability of Automated Incident Response in Office 365 Advanced Threat Protection (ATP). Applying these powerful automation capabilities to investigation and response workflows can dramatically improve the effectiveness and efficiency of your organization’s security teams.

A day in the life of a security analyst

To give you an idea of the complexity that security teams deal with in the absence of automation, consider the following typical workflow that these teams go through when investigating alerts:

Infographic showing these steps: Alert, Analyze, Investigate, Assess impact, Contain, and Respond.

And as they go through this flow for every single alert—potentially hundreds in a week—it can quickly become overwhelming. In addition, the analysis and investigation often require correlating signals across multiple different systems. This can make effective and timely response very difficult and costly. There are just too many alerts to investigate and signals to correlate for today’s lean security teams.

To address these challenges, earlier this year we announced the preview of powerful automation capabilities to help improve the efficiency of security teams significantly. The security playbooks we introduced address some of the most common threats that security teams investigate in their day-to-day jobs and are modeled on their typical workflows.

This story from Ithaca College reflects some of the feedback we received from customers of the preview of these capabilities, including:

“The incident detection and response capabilities we get with Office 365 ATP give us far more coverage than we’ve had before. This is a really big deal for us.”
—Jason Youngers, Director and Information Security Officer, Ithaca College

Two categories of automation now generally available

Today, we’re announcing the general availability of two categories of automation—automatic and manually triggered investigations:

  1. Automatic investigations that are triggered when alerts are raisedAlerts and related playbooks for the following scenarios are now available:
    • User-reported phishing emails—When a user reports what they believe to be a phishing email, an alert is raised triggering an automatic investigation.
    • User clicks a malicious link with changed verdict—An alert is raised when a user clicks a URL, which is wrapped by Office 365 ATP Safe Links, and is determined to be malicious through detonation (change in verdict). Or if the user clicks through the Office 365 ATP Safe Links warning pages an alert is also raised. In both cases, the automated investigation kicks in as soon as the alert is raised.
    • Malware detected post-delivery (Malware Zero-Hour Auto Purge (ZAP))—When Office 365 ATP detects and/or ZAPs an email with malware, an alert triggers an automatic investigation.
    • Phish detected post-delivery (Phish ZAP)—When Office 365 ATP detects and/or ZAPs a phishing email previously delivered to a user’s mailbox, an alert triggers an automatic investigation.
  1. Manually triggered investigations that follow an automated playbook—Security teams can trigger automated investigations from within the Threat Explorer at any time for any email and related content (attachment or URLs).

Rich security playbooks

In each of the above cases, the automation follows rich security playbooks. These playbooks are essentially a series of carefully logged steps to comprehensively investigate an alert and offer a set of recommended actions for containment and mitigation. They correlate similar emails sent or received within the organization and any suspicious activities for relevant users. Flagged activities for users might include mail forwarding, mail delegation, Office 365 Data Loss Prevention (DLP) violations, or suspicious email sending patterns.

In addition, aligned with our Microsoft Threat Protection promise, these playbooks also integrate with signals and detections from Microsoft Cloud App Security and Microsoft Defender ATP. For instance, anomalies detected by Microsoft Cloud App Security are ingested as part of these playbooks. And the playbooks also trigger device investigations with Microsoft Defender ATP (for malware playbooks) where appropriate.

Let’s look at each of these automation scenarios in detail:

User reports a phishing email—This represents one of the most common flows investigated today. The alert is raised when a user reports a phish email using the Report message add-in in Outlook or Outlook on the web and triggers an automatic investigation using the User Reported Message playbook.

Screenshot of a phishing email being investigated.

User clicks on a malicious linkA very common vector used by attackers is to weaponize a link after delivery of an email. With Office 365 ATP Safe Links protection, we can detect such attacks when links are detonated at time-of-click. A user clicking such links and/or overriding the Safe Links warning pages is at risk of compromise. The alert raised when a malicious URL is clicked triggers an automatic investigation using the URL verdict change playbook to correlate any similar emails and any suspicious activities for the relevant users across Office 365.

Image of a clicked URL being assigned as malicious.

Email messages containing malware removed after delivery—One of the critical pillars of protection in Office 365 Exchange Online Protection (EOP) and Office 365 ATP is our capability to ZAP malicious emails. Email messages containing malware removed after delivery alert trigger an investigation into similar emails and related user actions in Office 365 for the period when the emails were present in a user’s inbox. In addition, the playbook also triggers an investigation into the relevant devices for the users by leveraging the native integration with Microsoft Defender ATP.

Screenshot showing malware being zapped.

Email messages containing phish removed after deliveryWith the rise in phishing attack vectors, Office 365 EOP and Office 365 ATP’s ability to ZAP malicious emails detected after delivery is a critical protection feature. The alert raised triggers an investigation into similar emails and related user actions in Office 365 for the period when the emails were present in a user’s inbox and also evaluates if the user clicked any of the links.

Screenshot of a phish URL being zapped.

Automated investigation triggered from within the Threat Explorer—As part of existing hunting or security operations workflows, Security teams can also trigger automated investigations on emails (and related URLs and attachments) from within the Threat Explorer. This provides Security Operations (SecOps) a powerful mechanism to gain insights into any threats and related mitigations or containment recommendations from Office 365.

Screenshot of an action being taken in the Office 365 Security and Compliance dash. An email is being investigated.

Try out these capabilities

Based on feedback from our public preview of these automation capabilities, we extended the Office 365 ATP events and alerts available in the Office 365 Management API to include links to these automated investigations and related artifacts. This helps security teams integrate these automation capabilities into existing security workflow solutions, such as SIEMs.

These capabilities are available as part of the following offerings. We hope you’ll give it a try.

Bringing SecOps efficiency by connecting the dots between disparate threat signals is a key promise of Microsoft Threat Protection. The integration across Microsoft Threat Protection helps bring broad and valuable insights that are critical to the incident response process. Get started with a Microsoft Threat Protection trial if you want to experience the comprehensive and integrated protection that Microsoft Threat Protection provides.


Hacker movies that have an echo of truth

Films about hacks and cyberattacks have been popular for decades. These movies helped create the image of the hacker genius — just think of Stanley Jobson in “Swordfish.”

There is “Hackers,” in which a group of high schoolers access the mainframe of an oil company and discover evidence of embezzlement and “The Net,” about a woman (Sandra Bullock) whose identity is stolen.

You may think Hollywood depictions of hacking bear no resemblance to real life, but in each of the films below, there is an echo of truth in the fiction.

[Subscribe to Microsoft on the Issues for more on the topics that matter most.]

 “The Italian Job” (1969)

The Italian Job

A classic caper on a list of hacker movies. This story of British bank robbers undertaking a job in Turin, Italy, offers a surprising nod of things to come.

How do Michael Caine and his team plan to escape from this city? By hacking its traffic light system and causing widespread gridlock. This leads to the famous Mini Cooper getaway scene.

From Saudi Arabia to South Africa, billions of dollars are being invested in smart city projects. Some researchers estimate that spending on smart cities will reach $27.5 billion by 2023. It makes protecting those cities reliant on technology from disruption ever more crucial.

“WarGames” (1983)

At the height of the Cold War, a young hacker (Matthew Broderick) breaks into a US military supercomputer and comes close to starting a nuclear war. He thinks he’s playing a game based on a simulation, but this is not a drill.

According to a 2016 study by ISACA and RSA Conference, 74 percent of the world’s businesses expect to be hacked each year. And the economic loss due to cybercrime is estimated to reach $3 trillion by 2020.  It is one of the reasons that Microsoft has called for a Digital Geneva Convention to help protect cyberspace in times of peace.

“Sneakers” (1992)


This tech thriller, which spans from the late 1960s to the more computer-literate 1990s, boasts a heavyweight cast that includes Robert Redford and Sidney Poitier. A team of security specialists is approached by the NSA and commissioned to locate a mysterious black box. This team, propelled into a world of espionage, is soon hunted by rogue agents. The box turns out to be the key to cracking all known encryption and is, the team realizes, too powerful to fall into the wrong hands.

In 1992, the idea that something could break all known encryption sounded scary and a little implausible. Today, the existence of such technology is more likely thanks to quantum computing.

According to Martin Giles of the MIT Technology Review, quantum computers are “a security threat that we’re still totally unprepared for,” and it could be 20 years before cybersecurity catches up. Working closely with the United States National Institute for Standards and Technology, Microsoft is engaged with the development of post-quantum cryptography that will be able to withstand quantum computer capabilities, while still working with existing protocols.

“Hackers” (1995)


Starring Angelina Jolie and Jonny Lee Miller, “Hackers” is the story of a group of high school technology enthusiasts with codenames and complicated backstories. They hack into the mainframe of an oil company and discover evidence of embezzlement — but their activities are soon detected.

Robust cybersecurity is important for businesses and to the future of national economies, and it has become a priority for governments around the world. The Cybersecurity Tech Accord, announced by Microsoft in April 2018, is a public commitment among more than 100 global companies to protect and empower civilians online and help defend them against threats.

“The Net” (1995)

Sandra Bullock plays the lead role in this thriller that foreshadows one worry of the modern security landscape: identity theft.

Admittedly, in 1995, many important records were still paper-based, after all. Still, this is one of the central themes in the plot: Can Angela Bennett (Bullock) overcome a series of interconnected threats and regain her identity?

Today, secure passwords are a must. There is a wealth of personal data stored digitally that can all too easily compromise the security of your identity.

Cyberspace has become a battlefield and powerful cyberweapons are being used against civilians. Tools and Weapons, by Microsoft President Brad Smith and Carol Ann Browne, looks at how the world can respond. To read more and pre-order the book, visit Tools and Weapons. And follow @MSFTIssues on Twitter.


Deep learning rises: new methods for detecting malicious PowerShell scripts

Scientific and technological advancements in deep learning, a category of algorithms within the larger framework of machine learning, provide new opportunities for development of state-of-the art protection technologies. Deep learning methods are impressively outperforming traditional methods on such tasks as image and text classification. With these developments, there’s great potential for building novel threat detection methods using deep learning.

Machine learning algorithms work with numbers, so objects like images, documents, or emails are converted into numerical form through a step called feature engineering, which, in traditional machine learning methods, requires a significant amount of human effort. With deep learning, algorithms can operate on relatively raw data and extract features without human intervention.

At Microsoft, we make significant investments in pioneering machine learning that inform our security solutions with actionable knowledge through data, helping deliver intelligent, accurate, and real-time protection against a wide range of threats. In this blog, we present an example of a deep learning technique that was initially developed for natural language processing (NLP) and now adopted and applied to expand our coverage of detecting malicious PowerShell scripts, which continue to be a critical attack vector. These deep learning-based detections add to the industry-leading endpoint detection and response capabilities in Microsoft Defender Advanced Threat Protection (Microsoft Defender ATP).

Word embedding in natural language processing

Keeping in mind that our goal is to classify PowerShell scripts, we briefly look at how text classification is approached in the domain of natural language processing. An important step is to convert words to vectors (tuples of numbers) that can be consumed by machine learning algorithms. A basic approach, known as one-hot encoding, first assigns a unique integer to each word in the vocabulary, then represents each word as a vector of 0s, with 1 at the integer index corresponding to that word. Although useful in many cases, the one-hot encoding has significant flaws. A major issue is that all words are equidistant from each other, and semantic relations between words are not reflected in geometric relations between the corresponding vectors.

Contextual embedding is a more recent approach that overcomes these limitations by learning compact representations of words from data under the assumption that words that frequently appear in similar context tend to bear similar meaning. The embedding is trained on large textual datasets like Wikipedia. The Word2vec algorithm, an implementation of this technique, is famous not only for translating semantic similarity of words to geometric similarity of vectors, but also for preserving polarity relations between words. For example, in Word2vec representation:

Madrid – Spain + Italy ≈ Rome

Embedding of PowerShell scripts

Since training a good embedding requires a significant amount of data, we used a large and diverse corpus of 386K distinct unlabeled PowerShell scripts. The Word2vec algorithm, which is typically used with human languages, provides similarly meaningful results when applied to PowerShell language. To accomplish this, we split the PowerShell scripts into tokens, which then allowed us to use the Word2vec algorithm to assign a vectorial representation to each token .

Figure 1 shows a 2-dimensional visualization of the vector representations of 5,000 randomly selected tokens, with some tokens of interest highlighted. Note how semantically similar tokens are placed near each other. For example, the vectors representing -eq, -ne and -gt, which in PowerShell are aliases for “equal”, “not-equal” and “greater-than”, respectively, are clustered together. Similarly, the vectors representing the allSigned, remoteSigned, bypass, and unrestricted tokens, all of which are valid values for the execution policy setting in PowerShell, are clustered together.

Figure 1. 2D visualization of 5,000 tokens using Word2vec

Examining the vector representations of the tokens, we found a few additional interesting relationships.

Token similarity: Using the Word2vec representation of tokens, we can identify commands in PowerShell that have an alias. In many cases, the token closest to a given command is its alias. For example, the representations of the token Invoke-Expression and its alias IEX are closest to each other. Two additional examples of this phenomenon are the Invoke-WebRequest and its alias IWR, and the Get-ChildItem command and its alias GCI.

We also measured distances within sets of several tokens. Consider, for example, the four tokens $i, $j, $k and $true (see the right side of Figure 2). The first three are usually used to represent a numeric variable and the last naturally represents a Boolean constant. As expected, the $true token mismatched the others – it was the farthest (using the Euclidean distance) from the center of mass of the group.

More specific to the semantics of PowerShell in cybersecurity, we checked the representations of the tokens: bypass, normal, minimized, maximized, and hidden (see the left side of Figure 2). While the first token is a legal value for the ExecutionPolicy flag in PowerShell, the rest are legal values for the WindowStyle flag. As expected, the vector representation of bypass was the farthest from the center of mass of the vectors representing all other four tokens.

Figure 2. 3D visualization of selected tokens

Linear Relationships: Since Word2vec preserves linear relationships, computing linear combinations of the vectorial representations results in semantically meaningful results. Below are a few interesting relationships we found:

high – $false + $true ≈’ low
‘-eq’ – $false + $true ‘≈ ‘-neq’
DownloadFile – $destfile + $str ≈’ DownloadString ‘
Export-CSV’ – $csv + $html ‘≈ ‘ConvertTo-html’
‘Get-Process’-$processes+$services ‘≈ ‘Get-Service’

In each of the above expressions, the sign ≈ signifies that the vector on the right side is the closest (among all the vectors representing tokens in the vocabulary) to the vector that is the result of the computation on the left side.

Detection of malicious PowerShell scripts with deep learning

We used the Word2vec embedding of the PowerShell language presented in the previous section to train deep learning models capable of detecting malicious PowerShell scripts. The classification model is trained and validated using a large dataset of PowerShell scripts that are labeled “clean” or “malicious,” while the embeddings are trained on unlabeled data. The flow is presented in Figure 3.

Figure 3 High-level overview of our model generation process

Using GPU computing in Microsoft Azure, we experimented with a variety of deep learning and traditional ML models. The best performing deep learning model increases the coverage (for a fixed low FP rate of 0.1%) by 22 percentage points compared to traditional ML models. This model, presented in Figure 4, combines several deep learning building blocks such as Convolutional Neural Networks (CNNs) and Long Short-Term Memory Recurrent Neural Networks (LSTM-RNN). Neural networks are ML algorithms inspired by biological neural systems like the human brain. In addition to the pretrained embedding described here, the model is provided with character-level embedding of the script.

Figure 4 Network architecture of the best performing model

Real-world application of deep learning to detecting malicious PowerShell

The best performing deep learning model is applied at scale using Microsoft ML.Net technology and ONNX format for deep neural networks to the PowerShell scripts observed by Microsoft Defender ATP through the AMSI interface. This model augments the suite of ML models and heuristics used by Microsoft Defender ATP to protect against malicious usage of scripting languages.

Since its first deployment, this deep learning model detected with high precision many cases of malicious and red team PowerShell activities, some undiscovered by other methods. The signal obtained through PowerShell is combined with a wide range of ML models and signals of Microsoft Defender ATP to detect cyberattacks.

The following are examples of malicious PowerShell scripts that deep learning can confidently detect but can be challenging for other detection methods:

Figure 5. Heavily obfuscated malicious script

Figure 6. Obfuscated script that downloads and runs payload

Figure 7. Script that decrypts and executes malicious code

Enhancing Microsoft Defender ATP with deep learning

Deep learning methods significantly improve detection of threats. In this blog, we discussed a concrete application of deep learning to a particularly evasive class of threats: malicious PowerShell scripts. We have and will continue to develop deep learning-based protections across multiple capabilities in Microsoft Defender ATP.

Development and productization of deep learning systems for cyber defense require large volumes of data, computations, resources, and engineering effort. Microsoft Defender ATP combines data collected from millions of endpoints with Microsoft computational resources and algorithms to provide industry-leading protection against attacks.

Stronger detection of malicious PowerShell scripts and other threats on endpoints using deep learning mean richer and better-informed security through Microsoft Threat Protection, which provides comprehensive security for identities, endpoints, email and data, apps, and infrastructure.

Shay Kels and Amir Rubin
Microsoft Defender ATP team

Additional references:

One simple action you can take to prevent 99.9 percent of attacks on your accounts

There are over 300 million fraudulent sign-in attempts to our cloud services every day. Cyberattacks aren’t slowing down, and it’s worth noting that many attacks have been successful without the use of advanced technology. All it takes is one compromised credential or one legacy application to cause a data breach. This underscores how critical it is to ensure password security and strong authentication. Read on to learn about common vulnerabilities and the single action you can take to protect your accounts from attacks.

Animated image showing the number of malware attacks and data breaches organizations face every day. 4,000 daily ransomware attacks. 300,000,000 fraudulent sign-in attempts. 167,000,000 daily malware attacks. 81% of breaches are caused by credential theft. 73% of passwords are duplicates. 50% of employees use apps that aren't approved by the enterprise. 99.9% of attacks can be blocked with multi-factor authentication.

Common vulnerabilities

In a recent paper from the SANS Software Security Institute, the most common vulnerabilities include:

  • Business email compromise, where an attacker gains access to a corporate email account, such as through phishing or spoofing, and uses it to exploit the system and steal money. Accounts that are protected with only a password are easy targets.
  • Legacy protocols can create a major vulnerability because applications that use basic protocols, such as SMTP, were not designed to manage Multi-Factor Authentication (MFA). So even if you require MFA for most use cases, attackers will search for opportunities to use outdated browsers or email applications to force the use of less secure protocols.
  • Password reuse, where password spray and credential stuffing attacks come into play. Common passwords and credentials compromised by attackers in public breaches are used against corporate accounts to try to gain access. Considering that up to 73 percent of passwords are duplicates, this has been a successful strategy for many attackers and it’s easy to do.

What you can do to protect your company

You can help prevent some of these attacks by banning the use of bad passwords, blocking legacy authentication, and training employees on phishing. However, one of the best things you can do is to just turn on MFA. By providing an extra barrier and layer of security that makes it incredibly difficult for attackers to get past, MFA can block over 99.9 percent of account compromise attacks. With MFA, knowing or cracking the password won’t be enough to gain access. To learn more, read Your Pa$$word doesn’t matter.

MFA is easier than you think

According to the SANS Software Security Institute there are two primary obstacles to adopting MFA implementations today:

  1. Misconception that MFA requires external hardware devices.
  2. Concern about potential user disruption or concern over what may break.

Matt Bromiley, SANS Digital Forensics and Incident Response instructor, says, “It doesn’t have to be an all-or-nothing approach. There are different approaches your organization could use to limit the disruption while moving to a more advanced state of authentication.” These include a role-based or by application approach—starting with a small group and expanding from there. Bret Arsenault shares his advice on transitioning to a passwordless model in Preparing your enterprise to eliminate passwords.

Take a leap and go passwordless

Industry protocols such as WebAuthn and CTAP2, ratified in 2018, have made it possible to remove passwords from the equation altogether. These standards, collectively known as the FIDO2 standard, ensure that user credentials are protected end-to-end and strengthen the entire security chain. The use of biometrics has become more mainstream, popularized on mobile devices and laptops, so it’s a familiar technology for many users and one that is often preferred to passwords anyway. Passwordless authentication technologies are not only more convenient for people but are extremely difficult and costly for hackers to compromise. Learn more about Microsoft passwordless authentication solutions in a variety of form factors to meet user needs.

Convince your boss

Download the SANS white paper Bye Bye Passwords: New Ways to Authenticate to read more on guidance for companies ready to take the next step to better protect their environments from password risk. Remember, talk is easy, action gets results!


Protect yourself against ‘wormable’ BlueKeep vulnerability

Worms are the cause of many cyber headaches. They can easily replicate themselves to spread malicious malware to other computers in your network. As the field responders providing Microsoft enterprise customers with onsite assistance to serious cybersecurity threats, our Detection and Response Team (DART) has seen quite a few worms. If you’ve met the DART Team, then you know your worms are our concern and that’s why we keep an eye out for BlueKeep.

Protect against BlueKeep

This summer, the DART team has been preparing for CVE-2019-0708, colloquially known as BlueKeep, and has some advice on how you can protect your network. The BlueKeep vulnerability is “wormable,” meaning it creates the risk of a large-scale outbreak due to its ability to replicate and propagate, similar to Conficker and WannaCry. Conficker has been widely estimated to have impacted 10- to 12-million computer systems worldwide. WannaCry was responsible for approximately $300 million in damages at just one global enterprise.

To protect against BlueKeep, we strongly recommend you apply the Windows Update, which includes a patch for the vulnerability. If you use Remote Desktop in your environment, it’s very important to apply all the updates. If you have Remote Desktop Protocol (RDP) listening on the internet, we also strongly encourage you to move the RDP listener behind some type of second factor authentication, such as VPN, SSL Tunnel, or RDP gateway.

You also want to enable Network Level Authentication (NLA), which is a mitigation to prevent un-authenticated access to the RDP tunnel. NLA forces users to authenticate before connecting to remote systems, which dramatically decreases the chance of success for RDP-based worms. The DART team highly recommends you enable NLA regardless of this patch, as it mitigates a whole slew of other attacks against RDP.

If you’re already aware of the BlueKeep remediation methods, but are thinking about testing it before going live, we recommend that you deploy the patch. It’s important to note that the exploit code is now publicly and widely available to everyone, including malicious actors. By exploiting a vulnerable RDP system, attackers will also have access to all user credentials used on the RDP system.

Why the urgency?

Via open source telemetry, we see more than 400,000 endpoints lacking any form of network level authentication, which puts each of these systems potentially at risk from a worm-based weaponization of the BlueKeep vulnerability.

The timeline between patch release and the appearance of a worm outbreak is difficult to predict and varies from case to case. As always, the DART team is ready for the worst-case scenario. We also want to help our customers be prepared, so we’re sharing a few previous worms and the timeline from patch to attack. Hopefully, this will encourage everyone to patch immediately.

Chart showing vulnerability, patch release, and outbreak. Vulnerability: MS08-067; Patch release: October 23, 2008; Outbreak: late December 2008. Vulnerability: MS17-010; Patch release: March 14, 2017; Outbreak: May 12, 2017. Vulnerability: CVE-2019-0708; Patch release: May 13, 2019; Outbreak column shows three question marks.

Learn more

To learn more about DART, our engagements, and how they are delivered by experienced cybersecurity professionals who devote 100 percent of their time to providing cybersecurity solutions to customers worldwide, please contact your account executive. Bookmark the Security blog to keep up with our expert coverage on security matters. Also, follow us at @MSFTSecurity for the latest news and updates on cybersecurity.

This document is for informational purposes only and Microsoft makes no warranties, express or implied, in this blog.


General availability for the Azure Security Center for IoT announced

As organizations pursue digital transformation by connecting vital equipment or creating new connected products, IoT deployments will get bigger and more common. In fact, IDC forecasts that IoT will continue to grow at double digit rates until IoT spending surpasses $1 trillion in 2022. As these IoT deployments come online, newly connected devices will expand the attack surface available to attackers, creating opportunities to target the valuable data generated by IoT.

Organizations understand the risks and are rightly worried about IoT. Bain’s research shows that security concerns are the top reason organizations have slowed or paused IoT rollouts*. Because IoT requires integrating many different technologies (heterogenous devices must be linked to IoT cloud services that connect to analytics services and business applications), organizations face the challenge of securing both the pieces of their IoT solution and the connections between those pieces. Attackers target weak spots; even one weak device configuration, cloud service, or admin account can provide a way into your solution. Your organization must monitor for threats and misconfigurations across all parts of your IoT solution: devices, cloud services, the supporting infrastructure, and the admin accounts who access them.

To give your organization IoT threat protection and security posture management across your entire IoT solution, we’re announcing the general availability of Azure Security Center for IoT. Azure Security Center allows you to protect your end-to-end IoT deployment by identifying and responding to emerging threats, as well as finding issues in your configurations before attackers can use them to compromise your deployment. As organizations use Azure Security Center for IoT to manage their security roadblocks, they remove the barriers keeping them from business transformation:

“With Azure Security Center for IoT, we can both address very real IoT threat models with the velocity of Azure and gain management control over the fastest scaling part of our business, which allows me to focus on delivering outcomes rather than hot fixing devices.” – Alex Kreilein, CISO RapidDeploy

Building secure IoT solutions with Azure Security Center

Securing IoT is challenging for many reasons: IoT deployments are complicated, creating opportunity for integration errors that attackers can exploit; IoT devices are heterogenous and often lack proper security measures; organizations may not have the skillsets or SecOps headcount to take on a new IoT security workload; and IoT deployments are difficult to monitor using traditional IT security tools. When organizations choose Microsoft for their IoT deployments, however, they get secure-by-design devices and services such as Azure Sphere and IoT Hub, end-to-end integration and monitoring from device to cloud, and the expertise from Microsoft and our partners to build a secure solution that meets their exact use case.

Azure Security Center for IoT builds on Microsoft’s secure-by-design IoT services with threat protection and security posture management designed for securing entire IoT deployments, including Microsoft and 3rd party devices. Azure Security Center is the first IoT security service from a major cloud provider that enables organizations to prevent, detect, and help remediate potential attacks on all the different components that make up an IoT deployment: from small sensors, to edge computing devices and gateways, to Azure IoT Hub, and on to the compute, storage, databases, and AI/ML workloads that organizations connect to their IoT deployments. This end-to-end protection is vital to secure IoT deployments. Although devices may be a common target for attackers, the services that store your data and the admins who manage your IoT solution are also valuable targets.

An image showing the Overview tab in Azure Security Center.

As IoT threats evolve due to creative attackers analyzing the new devices, use cases, and applications the industry creates, Microsoft’s unique threat intelligence, sourced from the more than 6 trillion signals that Microsoft collects every day, keeps your organization ahead of attackers. Azure Security Center creates a list of potential threats, ranked by importance, so security pros and IoT admins can remediate problems across devices, IoT services, connected Azure services, and the admins who use them.

Azure Security Center also creates ranked lists of possible misconfigurations and insecure settings, allowing IoT admins and security pros to fix the most important issues in their IoT security posture first. To create these security posture suggestions, Azure Security Center draws from Microsoft’s unique threat intelligence, as well as the industry standards. Customers can also port their data into SIEMs such as Azure Sentinel, allowing security pros to combine IoT security data with data from across the organization for artificial intelligence or advanced analysis.

Organizations can monitor their entire IoT solution, stay ahead of evolving threats, and fix configuration issues before they become threats. When combined with Microsoft’s secure-by-design devices, services, and the expertise we share with you and your partners, Azure Security Center for IoT provides an important way to reduce the risk of IoT while achieving your business goals. 

Next steps

*Used with permission from Bain & Company

Command line quick tips: More about permissions

A previous article covered some basics about file permissions on your Fedora system. This installment shows you additional ways to use permissions to manage file access and sharing. It also builds on the knowledge and examples in the previous article, so if you haven’t read that one, do check it out.

Symbolic and octal

In the previous article you saw how there are three distinct permission sets for a file. The user that owns the file has a set, members of the group that owns the file has a set, and then a final set is for everyone else. These permissions are expressed on screen in a long listing (ls -l) using symbolic mode.

Each set has r, w, and x entries for whether a particular user (owner, group member, or other) can read, write, or execute that file. But there’s another way to express these permissions: in octal mode.

You’re used to the decimal numbering system, which has ten distinct values (0 through 9). The octal system, on the other hand, has eight distinct values (0 through 7). In the case of permissions, octal is used as a shorthand to show the value of the r, w, and x fields. Think of each field as having a value:

  • r = 4
  • w = 2
  • x = 1

Now you can express any combination with a single octal value. For instance, read and write permission, but no execute permission, would have a value of 6. Read and execute permission only would have a value of 5. A file’s rwxr-xr-x symbolic permission has an octal value of 755.

You can use octal values to set file permissions with the chmod command similarly to symbolic values. The following two commands set the same permissions on a file:

chmod u=rw,g=r,o=r myfile1
chmod 644 myfile1

Special permission bits

There are several special permission bits also available on a file. These are called setuid (or suid), setgid (or sgid), and the sticky bit (or delete inhibit). Think of this as yet another set of octal values:

  • setuid = 4
  • setgid = 2
  • sticky = 1

The setuid bit is ignored unless the file is executable. If that’s the case, the file (presumably an app or a script) runs as if it were launched by the user who owns the file. A good example of setuid is the /bin/passwd utility, which allows a user to set or change passwords. This utility must be able to write to files no user should be allowed to change. Therefore it is carefully written, owned by the root user, and has a setuid bit so it can alter the password related files.

The setgid bit works similarly for executable files. The file will run with the permissions of the group that owns it. However, setgid also has an additional use for directories. If a file is created in a directory with setgid permission, the group owner for the file will be set to the group owner of the directory.

Finally, the sticky bit, while ignored for files, is useful for directories. The sticky bit set on a directory will prevent a user from deleting files in that directory owned by other users.

The way to set these bits with chmod in octal mode is to add a value prefix, such as 4755 to add setuid to an executable file. In symbolic mode, the u and g can be used to set or remove setuid and setgid, such as u+s,g+s. The sticky bit is set using o+t. (Other combinations, like o+s or u+t, are meaningless and ignored.)

Sharing and special permissions

Recall the example from the previous article concerning a finance team that needs to share files. As you can imagine, the special permission bits help to solve their problem even more effectively. The original solution simply made a directory the whole group could write to:

drwxrwx---. 2 root finance 4096 Jul 6 15:35 finance

One problem with this directory is that users dwayne and jill, who are both members of the finance group, can delete each other’s files. That’s not optimal for a shared space. It might be useful in some situations, but probably not when dealing with financial records!

Another problem is that files in this directory may not be truly shared, because they will be owned by the default groups of dwayne and jill — most likely the user private groups also named dwayne and jill.

A better way to solve this is to set both setgid and the sticky bit on the folder. This will do two things — cause files created in the folder to be owned by the finance group automatically, and prevent dwayne and jill from deleting each other’s files. Either of these commands will work:

sudo chmod 3770 finance
sudo chmod u+rwx,g+rwxs,o+t finance

The long listing for the file now shows the new special permissions applied. The sticky bit appears as T and not t because the folder is not searchable for users outside the finance group.

drwxrws--T. 2 root finance 4096 Jul 6 15:35 finance


New machine learning model sifts through the good to unearth the bad in evasive malware

We continuously harden machine learning protections against evasion and adversarial attacks. One of the latest innovations in our protection technology is the addition of a class of hardened malware detection machine learning models called monotonic models to Microsoft Defender ATP‘s Antivirus.

Historically, detection evasion has followed a common pattern: attackers would build new versions of their malware and test them offline against antivirus solutions. They’d keep making adjustments until the malware can evade antivirus products. Attackers then carry out their campaign knowing that the malware won’t initially be blocked by AV solutions, which are then forced to catch up by adding detections for the malware. In the cybercriminal underground, antivirus evasion services are available to make this process easier for attackers.

Microsoft Defender ATP’s Antivirus has significantly advanced in becoming resistant to attacker tactics like this. A sizeable portion of the protection we deliver are powered by machine learning models hosted in the cloud. The cloud protection service breaks attackers’ ability to test and adapt to our defenses in an offline environment, because attackers must either forgo testing, or test against our defenses in the cloud, where we can observe them and react even before they begin.

Hardening our defenses against adversarial attacks doesn’t end there. In this blog we’ll discuss a new class of cloud-based ML models that further harden our protections against detection evasion.

Most machine learning models are trained on a mix of malicious and clean features. Attackers routinely try to throw these models off balance by stuffing clean features into malware.

Monotonic models are resistant against adversarial attacks because they are trained differently: they only look for malicious features. The magic is this: Attackers can’t evade a monotonic model by adding clean features. To evade a monotonic model, an attacker would have to remove malicious features.

Monotonic models explained

Last summer, researchers from UC Berkeley (Incer, Inigo, et al, “Adversarially robust malware detection using monotonic classification”, Proceedings of the Fourth ACM International Workshop on Security and Privacy Analytics, ACM, 2018) proposed applying a technique of adding monotonic constraints to malware detection machine learning models to make models robust against adversaries. Simply put, the said technique only allows the machine learning model to leverage malicious features when considering a file – it’s not allowed to use any clean features.

Figure 1. Features used by a baseline versus a monotonic constrained logistic regression classifier. The monotonic classifier does not use cleanly-weighted features so that it’s more robust to adversaries.

Inspired by the academic research, we deployed our first monotonic logistic regression models to Microsoft Defender ATP cloud protection service in late 2018. Since then, they’ve played an important part in protecting against attacks.

Figure 2 below illustrates the production performance of the monotonic classifiers versus the baseline unconstrained model. Monotonic-constrained models expectedly have lower outcome in detecting malware overall compared to classic models. However, they can detect malware attacks that otherwise would have been missed because of clean features.

Figure 2. Malware detection machine learning classifiers comparing the unconstrained baseline classifier versus the monotonic constrained classifier in customer protection.

The monotonic classifiers don’t replace baseline classifiers; they run in addition to the baseline and add additional protection. We combine all our classifiers using stacked classifier ensembles–monotonic classifiers add significant value because of the unique classification they provide.

How Microsoft Defender ATP uses monotonic models to stop adversarial attacks

One common way for attackers to add clean features to malware is to digitally code-sign malware with trusted certificates. Malware families like ShadowHammer, Kovter, and Balamid are known to abuse certificates to evade detection. In many of these cases, the attackers impersonate legitimate registered businesses to defraud certificate authorities into issuing them trusted code-signing certificates.

LockerGoga, a strain of ransomware that’s known for being used in targeted attacks, is another example of malware that uses digital certificates. LockerGoga emerged in early 2019 and has been used by attackers in high-profile campaigns that targeted organizations in the industrial sector. Once attackers are able breach a target network, they use LockerGoga to encrypt enterprise data en masse and demand ransom.

Figure 3. LockerGoga variant digitally code-signed with a trusted CA

When Microsoft Defender ATP encounters a new threat like LockerGoga, the client sends a featurized description of the file to the cloud protection service for real-time classification. An array of machine learning classifiers processes the features describing the content, including whether attackers had digitally code-signed the malware with a trusted code-signing certificate that chains to a trusted CA. By ignoring certificates and other clean features, monotonic models in Microsoft Defender ATP can correctly identify attacks that otherwise would have slipped through defenses.

Very recently, researchers demonstrated an adversarial attack that appends a large volume of clean strings from a computer game executable to several well-known malware and credential dumping tools – essentially adding clean features to the malicious files – to evade detection. The researchers showed how this technique can successfully impact machine learning prediction scores so that the malware files are not classified as malware. The monotonic model hardening that we’ve deployed in Microsoft Defender ATP is key to preventing this type of attack, because, for a monotonic classifier, adding features to a file can only increase the malicious score.

Given how they significantly harden defenses, monotonic models are now standard components of machine learning protections in Microsoft Defender ATP‘s Antivirus. One of our monotonic models uniquely blocks malware on an average of 200,000 distinct devices every month. We now have three different monotonic classifiers deployed, protecting against different attack scenarios.

Monotonic models are just the latest enhancements to Microsoft Defender ATP’s Antivirus. We continue to evolve machine learning-based protections to be more resilient to adversarial attacks. More effective protections against malware and other threats on endpoints increases defense across the entire Microsoft Threat Protection. By unifying and enabling signal-sharing across Microsoft’s security services, Microsoft Threat Protection secures identities, endpoints, email and data, apps, and infrastructure.

Geoff McDonald (@glmcdona),Microsoft Defender ATP Research team
with Taylor Spangler, Windows Data Science team

Talk to us

Questions, concerns, or insights on this story? Join discussions at the Microsoft Defender ATP community.

Follow us on Twitter @MsftSecIntel.

Manage your passwords with Bitwarden and Podman

You might have encountered a few advertisements the past year trying to sell you a password manager. Some examples are LastPass, 1Password, or Dashlane. A password manager removes the burden of remembering the passwords for all your websites. No longer do you need to re-use passwords or use easy-to-remember passwords. Instead, you only need to remember one single password that can unlock all your other passwords for you.

This can make you more secure by having one strong password instead of many weak passwords. You can also sync your passwords across devices if you have a cloud-based password manager like LastPass, 1Password, or Dashlane. Unfortunately, none of these products are open source. Luckily there are open source alternatives available.

Open source password managers

These alternatives include Bitwarden, LessPass, or KeePass. Bitwarden is an open source password manager that stores all your passwords encrypted on the server, which works the same way as LastPass, 1Password, or Dashlane. LessPass is a bit different as it focuses on being a stateless password manager. This means it derives passwords based on a master password, the website, and your username rather than storing the passwords encrypted. On the other side of the spectrum there’s KeePass, a file-based password manager with a lot of flexibility with its plugins and applications.

Each of these three apps has its own downsides. Bitwarden stores everything in one place and is exposed to the web through its API and website interface. LessPass can’t store custom passwords since it’s stateless, so you need to use their derived passwords. KeePass, a file-based password manager, can’t easily sync between devices. You can utilize a cloud-storage provider together with WebDAV to get around this, but a lot of clients do not support it and you might get file conflicts if devices do not sync correctly.

This article focuses on Bitwarden.

Running an unofficial Bitwarden implementation

There is a community implementation of the server and its API called bitwarden_rs. This implementation is fully open source as it can use SQLite or MariaDB/MySQL, instead of the proprietary Microsoft SQL Server that the official server uses.

It’s important to recognize some differences exist between the official and the unofficial version. For instance, the official server has been audited by a third-party, whereas the unofficial one hasn’t. When it comes to implementations, the unofficial version lacks email confirmation and support for two-factor authentication using Duo or email codes.

Let’s get started running the server with SELinux in mind. Following the documentation for bitwarden_rs you can construct a Podman command as follows:

$ podman run -d \ 
--userns=keep-id \
--name bitwarden \
-e ROCKET_PORT=8080 \
-v /home/egustavs/Bitwarden/bw-data/:/data/:Z \
-p 8080:8080 \

This downloads the bitwarden_rs image and runs it in a user container under the user’s namespace. It uses a port above 1024 so that non-root users can bind to it. It also changes the volume’s SELinux context with :Z to prevent permission issues with read-write on /data.

If you host this under a domain, it’s recommended to put this server under a reverse proxy with Apache or Nginx. That way you can use port 80 and 443 which points to the container’s 8080 port without running the container as root.

Running under systemd

With Bitwarden now running, you probably want to keep it that way. Next, create a unit file that keeps the container running, automatically restarts if it doesn’t respond, and starts running after a system restart. Create this file as /etc/systemd/system/bitwarden.service:

Description=Bitwarden Podman container

ExecStart=/usr/bin/podman run 'bitwarden'
ExecStop=-/usr/bin/podman stop -t 10 'bitwarden'


Now, enable and start it using sudo:

$ sudo systemctl enable bitwarden.service && sudo systemctl start bitwarden.service
$ systemctl status bitwarden.service
bitwarden.service - Bitwarden Podman container
Loaded: loaded (/etc/systemd/system/bitwarden.service; enabled; vendor preset: disabled)
Active: active (running) since Tue 2019-07-09 20:23:16 UTC; 1 day 14h ago
Main PID: 14861 (podman)
Tasks: 44 (limit: 4696)
Memory: 463.4M

Success! Bitwarden is now running under system and will keep running.

Adding LetsEncrypt

It’s strongly recommended to run your Bitwarden instance through an encrypted channel with something like LetsEncrypt if you have a domain. Certbot is a bot that creates LetsEncrypt certificates for us, and they have a guide for doing this through Fedora.

After you generate a certificate, you can follow the bitwarden_rs guide about HTTPS. Just remember to append :Z to the LetsEncrypt volume to handle permissions while not changing the port.

Photo by CMDR Shane on Unsplash.


Microsoft’s Threat & Vulnerability Management solution now generally available

I’m excited to announce that Microsoft’s Threat & Vulnerability Management solution is generally available as of June 30! We have been working closely with customers for more than a year to incorporate their real needs and feedback to better address vulnerability management. Our goal is to empower defenders with the tools they need to better protect against evolving threats, and we believe this solution will help provide that additional visibility and agility they need.

Threat & Vulnerability Management (TVM) is a built-in capability in Microsoft Defender Advanced Threat Protection (ATP) that uses a risk-based approach to discover, prioritize, and remediate endpoint vulnerabilities and misconfigurations. With Microsoft Defender ATP’s Threat & Vulnerability Management, customers benefit from:

  • Continuous discovery of vulnerabilities and misconfigurations
  • Prioritization based on business context and dynamic threat landscape
  • Correlation of vulnerabilities with endpoint detection and response (EDR) alerts to expose breach insights
  • Machine-level vulnerability context during incident investigations
  • Built-in remediation processes through unique integration with Microsoft Intune and Microsoft System Center Configuration Manager

Traditional vulnerability scanning only happens periodically, leaving organizations with security blind spots between scans. The one-size-fits-all approach that these traditional solutions use ignores critical business-specific context, as well as the dynamic threat landscape. This is coupled with the fact that mitigation of vulnerabilities is a manual process, often across teams, that can take days, weeks, or months to complete. This leaves a window of opportunity for attackers and puts our defenders in a tough spot.

To address these challenges Microsoft partnered with a dozen enterprise customers on the design and creation of this new Threat & Vulnerability Management solution. One of them is Telit, a global leader in IoT enablement offering end-to-end IoT solutions, including enterprise-grade hardware, connectivity, platform, and consulting services. Telit already had a well-defined vulnerability management program in place, but said they were missing several critical capabilities, including visibility, prioritization, and remediation.

Our design partners play a key role throughout the entire process, from planning and building to operationalizing and maturing the product so we can deliver the best experience. Many of our customers have existing vulnerability management programs, so we knew that to have them switch to Microsoft we would need a disruptive approach to vulnerability management. From private preview to general availability and beyond, our key goals were to bridge the gap between Security and IT roles in threat protection, to reduce time to threat resolution while enabling real-time prioritization and risk reduction based on the evolving threat landscape and business context. The team continues to incorporate feedback from customers and partners, adding these new capabilities on a monthly basis.

“Telit’s previous threat and vulnerability solutions were limited to on-premises connected endpoints. Moving to Microsoft’s TVM cloud-based solution provides us much better visibility into roaming endpoints with a continuous assessment, especially when our endpoints are connected to untrusted networks.”
— Itzik Menashe, VP of IT & Information Security, Telit

Working together with Telit, we quickly understood that the current prioritization norm is not enough to properly reduce risk in an organization. We consulted with our partners on a new risk-based approach, which is focused on continuous discovery of vulnerabilities and misconfigurations and correlated those insights with context specific to their business and the dynamic threat landscape.

Microsoft’s built-in, end-to-end remediation process helps Telit bridge the gap between their security and operations teams. The unique integration with Microsoft Intune allows their security team to create remediation requests with a click of a button, and the operations team receives the requests automatically with all relevant information and can start the remediation process right away. The security team can then watch their exposure score drop in real time as remediation progresses.

“Microsoft’s TVM provides Telit with an easy-to-use solution that incorporates strong discovery capabilities, a risk-based approach to prioritization, and an effective remediation process. With this solution we are able to cover a large number of endpoints using a very small team of security engineers.”
— Mor Asher, Global IT and Information Security Manager, Telit

The product experience and ease of implementation was a big driver for Telit and thousands of other active customers to start using Microsoft Defender ATP Threat & Vulnerability Management. Telit had Microsoft Defender ATP’s TVM up and running within seconds.

To learn more about threat and vulnerability management watch our video that walks you through the experience.

If you already have Microsoft Defender ATP, the TVM solution is now available within your ATP portal. If you would like to sign up for a trial of Microsoft Defender ATP including TVM, sign up here.

We’re excited for our customers to evaluate this new solution and are looking forward to continued feedback.