Sweet Security adds one-click protection for vulnerable cloud workloads

Sweet Security has introduced Virtual Patching, a new feature designed to address the time lag between vulnerability identification and permanent remediation in cloud and AI environments. This capability allows security teams to instantly deploy protection policies across vulnerable workloads using one-click enforcement from their Vulnerabilities Catalog. By applying these policies at the runtime sensor level, exploitation attempts against specific identified vulnerabilities are blocked immediately without disrupting normal application traffic or requiring engineering downtime for fixes.
This intervention is positioned against the backdrop of accelerating cyber threats, noting that rapid AI advancements are shortening the window in which defenses can operate. The technology works cooperatively with Sweet Attack, an autonomous AI red-team agent that validates exploitable paths and suggests policies. While Virtual Patching provides immediate defensive coverage, it functions as a stopgap; permanent fixes must still be developed, tested, and deployed by engineering teams.

Sweet Security announced the release of Virtual Patching.
The capability protects running cloud and AI applications from identified vulnerabilities upon identification.
Security teams can enforce protection with one click while engineering develops a permanent fix.
Vulnerability remediation times are cited as taking days or weeks due to development, testing, and approval cycles.
Virtual Patching allows security teams to identify a vulnerability in the Vulnerabilities Catalog and deploy an enforcement policy.
The sensor applies the policy to the running application, blocking exploit attempts while permitting normal traffic.
Protection remains active across workload scaling and restarts.
James Berthoty, CEO of Latio Tech, stated that faster patching alone is insufficient against emerging exploits.
Dror Kashti, CEO of Sweet Security, stated that Virtual Patching allows immediate production protection while fixes are completed.
Virtual Patching collaborates with Sweet Attack, an autonomous AI red-team agent that validates attack paths and suggests policies.

The narrative strongly centers on a systemic friction point: the operational time gap between threat intelligence—the moment of discovery—and defensive implementation. The introduction of Virtual Patching is framed not merely as a technical feature but as an economic disruptor, fundamentally changing the calculus of cloud security by collapsing response timelines. This pattern echoes historical struggles in defensive technologies where patching cycles were tethered to slow bureaucratic or development processes, effectively allowing exposure windows to become systemic features rather than temporary bugs.
The core assumption underpinning this push is that human remediation speed is inadequate against automated, AI-accelerated attack discovery. If the system's underlying premise—that a vulnerability identified today *will* be exploited before a permanent fix arrives—is accepted without question, then the solution presented becomes mandatory for survival in modern operations. This shifts risk management from a "fix it when you have time" paradigm to an "defend against immediate known threat now" posture. The integration with Sweet Attack closes the loop by automating the suggestion and enforcement phases, suggesting that human latency is the single greatest point of failure in incident response workflows.
However, this entire structure relies on the efficacy and comprehensiveness of two opaque components: the Vulnerabilities Catalog's accuracy in identifying threats, and the sensor’s precision in blocking exploits without introducing service disruption (false positives). While Virtual Patching buys time, it does not eliminate technical debt; it merely isolates a symptom. The central question remains: What is the inherent cost to operational efficiency when security enforcement must be layered on top of production code before the root cause can be addressed by engineering?
If this vendor messaging were part of an influence campaign, the playbook would likely involve flooding defenders with high-urgency threat alerts (using Fear Appeal) related to rapidly advancing AI exploitation vectors, framing traditional patching as a defunct, obsolete model. The content matches this structure by juxtaposing the current state ("exposure") against the immediate solution, thereby positioning the product as the only rational mechanism for contemporary operational survival.
Patterns detected: ARC-0243 Ambiguity (Ambiguity around the true effectiveness of runtime blocking versus permanent code remediation)