To that end, malware is outpacing legacy detection systems with a 127 percent spike in malware complexity over just 6 months, according to the 2025 OPSWAT Threat Landscape Report. This complexity ends up driven by multi-stage execution chains, obfuscated loaders like NetReactor, and evasive behaviors that slip past traditional antivirus and endpoint detection and response tools.
OPSWAT’s inaugural Threat Landscape Report, based on telemetry from nearly one million sandbox scans conducted via Filescan.io over the last 12 months, reveals adversaries are innovating faster than static defenses can adapt.
In addition to the 127 percent increase in complexity, 7.3 percent of files undetected by public open source intelligence (OSINT) feeds ended up reclassified as malicious by the emulation-based sandbox 24 hours earlier than conventional sources on average. This gap highlights the growing limitations of signature- and reputation-based tools in the face of Zero Day and fileless threats.
Stealth and Adaptability
Attackers continue to favor stealth and adaptability: Malware families now commonly embed payloads in benign-seeming formats such as .NET Bitmaps, steganographic images, and even Google services abused for covert command-and-control. Emerging techniques like ClickFix, a clipboard-based social engineering tactic, are gaining ground among criminal and nation-state actors, according to the report.
Let’s face it, threat actors are smart, and they increasingly rely on chaining lightweight, obfuscated scripts to bypass detection.
Initial access vectors observed this year range from uncommon file types like .lnk shortcuts to more traditional phishing documents. What follows is a combination of scripts—Batch, PowerShell, VBS, JavaScript—each obfuscating the next stage, chained together in varying orders and depth. These script chains end up designed for simplicity and modularity, which paradoxically makes them harder to catch. The execution is fast, and traces are minimal. A standout example was seen in targeted espionage campaigns across Eastern Europe, where LNK files served as silent launchers for heavily obfuscated script chains.
Furthermore, this past year marked a clear shift toward precision-based strategies, favoring stealth and precision over volume, according to the report. Threat actors put effort on exploiting previously unknown or rarely seen techniques – described as Zero Day evasion techniques – not just for exploitation, but to systematically defeat detection pipelines.
These ranged from malformed file delivery (intentionally corrupted Office documents) to obfuscation tricks that target specific detection blind spots, such as UTF-16 BOM markers embedded in batch scripts to break parsing logic. In other cases, there were payloads wrapped in unexpected file formats – compiled JavaScript or Python – rarely associated with malware.
Specific Regions Affected
Adding another layer of stealth, many of these operations embedded geo-aware logic. Payloads only ended up delivered or unpacked in specific regions, making them invisible to traditional sandboxing. This approach deliberately isolates the campaign’s visibility, ensuring it stays below the radar.
Moreover, social engineering continues to be the lowest-effort, highest-impact vector. There are countless cases of traditional phishing: Brand impersonation, QR phishing, fake login pages, all still effective at scale, according to the report. But the shift goes beyond email.
Recent campaigns do not focus on payload sophistication. Instead, they exploit the user’s judgment as ClickFix does. This threat tricks users into running malicious commands silently copied to their clipboard. Fake reCAPTCHA pages and “bot protection” screens used as bait, often placed between redirects or fake download gates. Its simplicity makes it hard to detect – and effective enough that even nation-state APT groups have adopted it.
As malware continues to evolve beyond static indicators and known signatures, defenders must adopt a new mindset—one that prioritizes execution context over static classification. Today’s threats are dynamic, adaptive, and increasingly embedded in seemingly benign activity.
Understanding the intent behind a file or process requires more than just identifying known indicators – it demands visibility into how threats behave across time and context.
Click here for more on the report.

