Abstract: A multi-agent AI-driven system addresses limitations of Large Language Models (LLMs) in automated software testing by integrating symbolic formal verification and generative AI. The system analyzes software to identify implied functions, dependencies, and static variables, generating mock-up compilation units to enable isolated unit testing. A symbolic verification engine systematically identifies backward arithmetic and logical conditions required to execute specific code paths, overcoming the inability of LLMs to reliably perform backward reasoning. LLMs then generate executable, forward-reasoned unit tests based on these formal results. Test executions produce coverage information that highlights gaps and detects errors or regressions. Reflective analysis by LLMs further generates targeted tests and suggests improvements, enhancing software robustness and test coverage. The method effectively combines formal logic with generative AI, enabling automated, reliable, and comprehensive software testing.

Abstract: Systems and methods for defining and utilizing DLP sub-dictionaries include monitoring network traffic associated with one or more tenants of a plurality of tenants of a cloud-based system; identifying a document in transit based on the monitoring of network traffic; identifying, based on one or more DLP dictionary weights, at least one parent dictionary and at least one corresponding sub-dictionary to be searched; matching contents of the document against data patterns or phrases defined in the identified sub-dictionary; and triggering an alert or a policy action if at least one matched data pattern or phrase exceeds a predefined threshold determined in part by the one or more DLP dictionary weights.

Abstract: Systems and methods for real-time scoring to monitor performance and provide proactive remediation via actionable insights which includes collecting one or more specific hardware and software performance metrics from one or more computing devices, assigning a weighted score to each performance metric, computing a composite score, and providing one or more actionable user recommendations.

Abstract: Systems and methods for enhancing the WireGuard protocol are provided. In one implementation, a gateway device is arranged in a cloud-based system and connected via tunnels to multiple connectors in a client network. The gateway device is configured to assign loopback IP addresses to the multiple connectors and populate the loopback IP addresses into AllowedIP lists associated with the multiple connectors, the multiple AllowedIP lists addressing the connectors as crypto peers for use with the WireGuard protocol. Based on network operating parameters, the gateway device is configured to determine a packet routing plan for distributing multiple packets among the connectors. When a packet of the multiple packets is determined, according to the packet routing plan, to be routed through a connector other than a designated WireGuard peer, the gateway device further performs a step of encapsulating the packet with the corresponding loopback IP address of the connector.

Abstract: Techniques for optimized tracing in IPV6 environments include sending a plurality of trace packets between a client and a destination in a service path; responsive to receiving a response from the plurality of trace packets, extracting trace information therefrom; and determining a corresponding router associated with each of the responses based on the trace information.

Abstract: Systems and methods for emulating Unix pipes on Windows operating systems include creating an emulated pipe by creating a listening Transmission Control Protocol (TCP) socket and a client TCP socket and establishing a local loopback TCP connection between them; utilizing the emulated pipe for inter-process communication by assigning file descriptors to the established TCP connection, wherein a first file descriptor is designated for reading and a second file descriptor is designated for writing; and enabling asynchronous operation by integrating the emulated pipe with event-driven Application Programming Interfaces (APIs), allowing non-blocking communication between processes or threads within an application.

Abstract: Systems and methods for client-side decryption include intercepting, via an agent application installed on a client device, a secure connection request originating from a browser or other application executing on the client device; extracting, by the agent application, handshake parameters from the secure connection request; establishing a secure session between the agent application and the browser using a locally generated decryption certificate and the handshake parameters; decrypting, by the agent application, data received from the browser and conditionally inspecting and re-encrypting the decrypted data for transmission to a cloud-based system or a destination server.

Abstract: Systems and methods for version based stateless caching to deliver cloud enforcement policy to user devices include receiving keep-alive signals from a plurality of agent applications, each agent application of the plurality of agent applications being associated with and identifying a corresponding user device; storing device details associated with the plurality of agent applications; intercepting a policy request from at least one agent application of the plurality of agent applications; and responsive to determining device details associated with the at least one agent application are stored, returning a response to the at least one agent application from a policy cache.

Abstract: Systems and methods described herein are configured to inspect prompts before they are transmitted to an external Large Language Model (LLM) chatbot or Generative Pre-trained Transformer (GPT). A method, according to one implementation, includes a step of intercepting a text-based prompt sent from a client device, the text-based prompt intended to be sent to an LLM chatbot. The method also includes analyzing the text-based prompt to determine whether one or more keywords exist therein. In response to detecting that one or more keywords exist in the text-based prompt, the method further includes a step of inspecting text surrounding the one or more keywords to determine whether one or more text segments include secretive information. In response to detecting that one or more text segments include secretive information, the method prevents leakage of the one or more text segments to the LLM chatbot while sending the text-based prompt.

Abstract: Systems and methods for geolocating an Access Point (AP) based on network identifiers and external data sources includes caching the AP's location by associating it with one or more network identifiers, deriving an egress IP address of the AP from an external tool or service, and generating a unique key based on the egress IP address. The unique key can be used to geolocate the AP. Such systems and methods can be incorporated with location based services.

Abstract: Systems and methods for enforcing policy based on assigned user risk scores in a cloud-based system. Various methods include receiving a request to access a resource; determining whether a user associated with the request is allowed to access the resource, wherein the determining is based on a risk score of the user; and responsive to the user being permitted to access the resource, stitching together a connection between a cloud-based system, the resource, and the device to provide access to the resource.

Abstract: Systems and methods for a hierarchical step-up authentication mechanism include monitoring access to one or more private applications; responsive to a request to access the one or more private applications, determining an Authentication Level (AL) of a user associated with the request, wherein determining the AL of the user comprises referencing one or more AL trees; and responsive to determining an AL of the user, performing one or more actions based thereon, wherein the one or more actions comprises one of allowing access to the one or more private applications and denying access to the one or more private applications.

Abstract: Systems and methods for generating SDLC-based application segments include defining a set of applications, the set of applications comprising a plurality of applications associated with a tenant of the cloud-based system; generating a plurality of application pairs from the set of applications; analyzing applications within each application pair of the plurality of application pairs for filtering the plurality of application pairs, the analyzing comprising a plurality of similarity checks and identification of environment key words; and generating one or more application segments each comprising one or more applications from the set of applications based on the filtering.

Abstract: Systems and methods for improving traffic inspection and response with smart agents and unified inspection data points include monitoring network traffic associated with a plurality of computing devices each having an agent application executing thereon, the monitoring including collecting inspection data based on the network traffic; receiving process-level data from each agent application; unifying the process-level data with the collected inspection data; and analyzing the unified data to detect anomalous or malicious activity.

Abstract: Systems and methods for providing Artificial Intelligence-as-a-Service (AIaaS) to enhance engineering development workflows includes hosting a plurality of Artificial Intelligence (AI) models in a scalable platform environment, wherein the plurality of AI models includes at least one generative AI model and one deterministic AI model automating one or more steps during an application development process by integrating the AI models to perform tasks including any of code generation, test case creation, User Interface (UI) prototyping, and data preprocessing, thereby reducing manual interventions and accelerating development; and continually monitoring one or more AI models associated with the application after deployment, wherein performance metrics such as latency, error rates, and resource usage are collected in real time to enable proactive alerts and ongoing model optimizations.

Abstract: Systems and methods are disclosed for determining floor-level workplace occupancy using network telemetry. Wi-Fi change events associated with devices are received and enriched with office configuration data mapping wireless access points to physical floors. A probabilistic spatial inference engine converts signal quality values into weighted scores, aggregates scores by floor, and determines an inferred floor and confidence value. A temporal state management module validates inferred floor changes based on the confidence value and stored device state to filter noise and confirm location transitions. Confirmed changes update real-time and historical datastores to maintain aggregated floor-level occupancy counts. A machine learning module may adaptively update access point-to-floor mappings and generate predictive occupancy forecasts based on historical data.

Abstract: Systems and methods for providing unified access to Large Language Models (LLMs) includes registering a plurality of generative AI models in a model management layer, storing associated version information and implementing lifecycle operations including loading, reloading, or deleting models; receiving inference requests at a single, unified endpoint in an Application Programming Interface (API) serving layer, wherein each request is translated into a standardized format compatible with each of the generative AI models; dynamically routing each received inference request to a selected model or models using a routing engine that evaluates factors comprising historical performance, latency, cost, and user-defined accuracy or speed requirements; and collecting and aggregating performance metrics for each generative AI model through a metrics and observability subsystem, wherein the collected metrics enable side-by-side response comparisons and cost forecasting.

Abstract: A Zero-Trust Network Access (ZTNA)-enabled Virtual Private Network (VPN) solution combines traditional VPN tunnels with ZTNA’s application-centric controls. Rather than granting broad network access, the system enforces fine-grained policies based on identity, device posture, and contextual factors. By positioning a cloud-based overlay between end users and private applications or LAN segments, it enables continuous verification and least-privileged access for both legacy and modern workloads. Key components—such as NP gateways, NP connectors, and application connectors—create outbound-only tunnels, eliminating the need for open inbound ports. Leveraging protocol-agnostic or WireGuard-based encryption, these tunnels transport traffic securely while preventing unauthorized visibility or lateral movement. Users benefit from robust, real-time security checks, application-specific routes, and seamless scalability.

Abstract: Systems and methods for providing a smart reauthentication process for prolonged WebSocket sessions which may present a risk to cloud environments. The various embodiments can include monitoring a WebSocket session associated with a user, detecting one or more triggers associated with an authentication of the WebSocket session, and initiating an authentication process based on the one or more triggers. The authentication process can disconnect an existing WebSocket session and force the user to reauthenticate in order to continue the session.

Abstract: Systems and methods for cloud security system assistance utilizing custom Large Language Models (LLMs) include providing a cloud-based security solution for an enterprise via a cloud-based system; displaying a User Interface (UI) associated with the cloud-based security solution having a chatbot, wherein the chatbot is configured to allow a user associated with the enterprise to enter a question; and responsive to receiving a question from a user via the chatbot, generating a detailed response to the question via a custom LLM, wherein the custom LLM is trained to provide assistance to users of the cloud-based security solution.