Inside the Huntress Tragic Quadrant: Cyber Threats on the Rise
The Huntress Tragic Quadrant organizes threats by the business outcomes they enable, rather than by malware family or endpoint. The report draws on Huntress incident-response data from the first two quarters of 2026 and describes the framework as a way to map techniques to outcomes. Huntress's Kyle Schreur says the goal is to show “all the techniques that we're seeing” on a quadrant organized around business impact. [1]
The framework highlights how different techniques can lead to similar outcomes, including data theft and encryption. It also gives defenders a way to compare the severity and adoption trajectory of the threats covered in the report. [1]
Reading the Quadrant
The Tragic Quadrant plots threats on two axes, as described in the Huntress report: [1]
- Y-axis — business outcome severity: threats associated with data exfiltration or encryption appear toward the top; those associated with staging, access, or discovery appear lower.
- X-axis — adoption trajectory: whether reported use is rising or falling over the measurement period.
The upper-right “tragic” quadrant contains threats the report characterizes as both critical and growing. Huntress places Microsoft device-code authentication abuse, malicious browser extensions, and ClickFix-style social engineering in this cluster. The framework also shows that other tracked categories follow different trajectories. [1]
The Critical-and-Growing Cluster
Microsoft device-code phishing is notable because the authentication flow can be abused without malware running on the victim’s computer. In the attack pattern described by Huntress, a user is directed to Microsoft’s legitimate device-code page and enters a code supplied by an attacker, potentially authorizing access to the attacker’s session. The report discusses access to services such as email, OneDrive, and SharePoint, and the role of refresh tokens in maintaining access. Huntress calls the technique “the perfect persistence mechanism,” emphasizing that the resulting activity can resemble normal user behavior. [1]
Malicious browser extensions also appear in the critical-and-growing cluster. Huntress describes extensions that can steal or alter web traffic, modify pages, redirect searches or advertising clicks, and route browsing through compromised machines. These capabilities can expose browser-based accounts and sessions to abuse. [1]
ClickFix social engineering uses deceptive instructions to persuade a person to run or paste attacker-provided commands, for example through a fake verification prompt. Huntress reports investigating cases in which ClickFix activity led to infostealers, remote-access tools, persistence, and ransomware — a delivery pattern echoed in our breakdown of how threat actors lure users to fake custom ChatGPTs that deploy RATs via ClickFix. The report’s examples include BlackCat/ALPHV and Akira activity. [1]
High-Volume Initial Access: Phishing Is Still the Front Door
The report discusses phishing and credential theft among the initial-access techniques seen in Huntress investigations. Huntress notes that convincing phishing messages can make older advice focused on spotting poor grammar less reliable. The article attributes improvements in phishing quality in part to AI, but does not establish that AI-generated lures account for the overall volume of phishing. [1]
Two phishing approaches receive particular attention. Vendor phishing can involve an attacker studying a company and inserting a fraudulent message into a supplier conversation, such as a request to change payment details. QR-code phishing places a link in an image; Huntress describes malicious QR campaigns reaching Microsoft 365 inboxes where email defenses did not detect them, a tactic also seen in campaigns that weaponized Microsoft Sway for large-scale quishing. [1]
Huntress also says stolen credentials are the most common initial-access vector in its investigations, a reminder of why the human layer remains the weak link. The report cautions that this finding may reflect the email-security coverage of its customer base: phishing that is filtered before it results in a credential issue may not appear in the same way in incident-response data. [1]
Living off the Land: RMM and DLL Side-Loading
Huntress describes legitimate remote monitoring and management (RMM) tools and DLL side-loading as techniques used in post-compromise activity and defense evasion, respectively. [1]
Legitimate RMM tools can provide attackers with remote access and persistence while resembling software used by IT providers. Huntress identifies RMM software as a prominent tool in post-compromise activity and discusses the difficulty of distinguishing malicious use from legitimate administration. [1]
DLL side-loading takes advantage of how a legitimate executable may load a nearby DLL. Huntress identifies it as a leading defense-evasion technique and describes cases involving compromised open-source projects, including a GitLab CI/CD incident in which a runner’s SSH key was abused to push malicious commits. [1]
AI-Related Techniques in the Report
The report includes several AI-related items on its defense-evasion axis, but does not characterize them as the quadrant’s critical-and-growing cluster. [1]
- Malicious use of AI: Huntress discusses AI-assisted code obfuscation and AI-generated malware as activity with moderate severity and a relatively small, growing footprint.
- AI platform abuse: The report describes threat actors using infrastructure associated with AI services, including domains hosted through providers’ platforms. This can complicate blocking because defenders may not want to block legitimate services wholesale.
- AI-assisted code signing: Huntress describes an observed case in which a malware executable was signed in under seven minutes using a signing platform’s process. The report presents this as a concerning example of how attackers may exploit frictionless signing workflows. [1]
These examples point to practical questions about access controls and review in software-signing and automation workflows. They do not, by themselves, establish that autonomous AI agents are a leading threat in the data covered by the report. [1]
The Data-Theft Endgame and the Shrinking Threats
Huntress describes a gradual rise in data-theft extortion and reports data destruction as the fastest-rising impact category in its observations. The report includes cases in which data was deliberately damaged, as well as incidents where attackers used compromised access to delete virtual machines. [1]
The report also describes some categories as declining among Huntress customers. It says ransomware activity is down in its data, while noting that the threat continues to evolve. It reports declines in business email compromise-style invoice fraud and infostealer activity as well. These are findings about the report’s observed data, not proof that the threats have disappeared or are declining everywhere. [1]
Defenses That Match the Quadrant
The quadrant’s outcome-based approach can help teams connect observed techniques to defensive priorities. The following measures respond to techniques discussed in the Huntress report; they are practical recommendations, not a checklist attributed to the report. [1]
- Govern device-code authentication and tokens. Monitor device-code sign-ins, restrict the flow where feasible, and protect refresh tokens as credentials.
- Inventory and constrain browser extensions. Use managed-extension policies and monitor for unauthorized installations.
- Prepare users for ClickFix tactics. Teach users not to paste unsolicited commands into Run, terminals, or other execution prompts.
- Control RMM access. Inventory approved tools, restrict installations, and monitor remote-access activity that does not match normal administration.
- Review software allow-lists. Assess whether trusted executables could load untrusted DLLs from writable locations.
- Add controls to signing and automation workflows. Require appropriate authentication and human review for sensitive actions.
- Improve email defenses. Use sender-authentication measures such as DMARC, DKIM, and SPF, and ensure QR codes in messages are inspected where possible.
- Protect and segment backups. Use access-controlled, resilient backups to reduce the impact of destructive incidents.
The Tragic Quadrant’s value is its focus on the outcomes and adoption trends reported by Huntress. It directs attention to techniques the report places in its critical-and-growing cluster, while keeping other findings—such as AI-related activity—in their reported context. [1]
Sources
[1] BleepingComputer, sponsored by Huntress Labs, “The Huntress Tragic Quadrant: Ranking Threats Hitting Businesses Hardest,” October 27, 2026. The article summarizes the Huntress Tragic Quadrant report, based on Huntress incident-response data from Q1–Q2 2026. https://bleepingcomputer.com/news/security/the-huntress-tragic-quadrant-ranking-threats-hitting-businesses-hardest
[2] Huntress, “Tragic Quadrant” report summary, cited via the BleepingComputer article in [1].
the tragic quadrant reframes ai cybersecurity threats