# US 2026/0093818 A1 · Automatically detecting multi-layer toxic combinations of application security risks in cloud environments > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2026/0093818 A1 - Document ID: US20260093818A1 - Status: Published application - Kind: A1 - Application number: 18/901,870 - Filed: 2024-09-30 - Published: 2026-04-02 - Earliest claimed: 2024-09-30 - Assignees: Operant AI, Inc. - Applicants: Operant AI, Inc. - Inventors: Priyanka Mohan Tembey; Glenn McDonald; Rotem Wolfovich; Vrajesh Rajesh Bhavsar; Ashley Michelle Chapman Roof - Web page: https://patentgrep.com/patent/US20260093818A1 ## CPC - G06F 21/577: {Assessing vulnerabilities and evaluating computer system security} - H04L 63/1433: {Vulnerability analysis} - G06F 21/55: Detecting local intrusion or implementing counter-measures - H04L 63/1408: {by monitoring network traffic (monitoring network traffic per se H04L43/00)} - G06F 2221/034: Test or assess a computer or a system - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract In an embodiment, one or more non-transitory computer-readable media storing sequences of instructions which, when executed using processors, cause the processors to execute, using a runtime security engine deployed within a computing environment: obtaining telemetry data comprising security-related data from microservices executing in the computing environment; accessing risk categories codifying insecure behavior across multiple layers in the computing environment; accessing toxic combination patterns across the layers, the toxic combination patterns being generated based on the risk categories, and each of the toxic combination pattern indicating a high-severity security impact in case of an attack in the computing environment; evaluating the telemetry data against the risk categories and toxic combination patterns, the evaluating producing a customized set of toxic combination patterns specific to the application within the computing environment; and displaying the customized set of toxic combination patterns in real time in a graphical user interface of a computer display device. ## Claims ### Claim 1 One or more non-transitory computer-readable media storing one or more sequences of instructions which, when executed using one or more processors, cause the one or more processors to execute, using a runtime security engine deployed within a computing environment: obtaining, from a local controller service deployed within the computing environment, telemetry data from microservices executing in the computing environment, and the telemetry data comprising security-related data; accessing a plurality of risk categories codifying insecure behavior across a plurality of layers in the computing environment; accessing a plurality of toxic combination patterns across the plurality of layers, the plurality of toxic combination patterns being generated based on the plurality of risk categories, and each of the toxic combination pattern indicating a high-severity security impact in case of an attack in the computing environment; evaluating the telemetry data against the plurality of risk categories and the plurality of toxic combination patterns, the evaluating producing a customized set of toxic combination patterns specific to the microservices within the computing environment; and displaying the customized set of toxic combination patterns in real time in a graphical user interface of a computer display device. ### Claim 2 (depends on 1) The one or more computer-readable media of claim 1, wherein the plurality of layers comprises a service layer, wherein the plurality of risk categories for the service layer comprises one or more of: lateral movement to critical services or over-permissioned services. ### Claim 3 (depends on 1) The one or more computer-readable media of claim 1, wherein the plurality of layers comprises an application programming interface (API) layer, wherein the plurality of risk categories for the API layer comprises one or more of: unauthenticated API access or unauthorized API access. ### Claim 4 (depends on 1) The one or more computer-readable media of claim 1, wherein the plurality of layers comprises a data layer, wherein the plurality of risk categories for the data layer comprises one or more of: data exfiltration to untrusted application programming interfaces (APIs) or weak database authorization. ### Claim 5 (depends on 1) The one or more computer-readable media of claim 1, wherein the plurality of layers comprises a container layer, wherein the plurality of risk categories for the container layer comprise one or more of: over-privileged containers or unlimited resources. ### Claim 6 (depends on 1) The one or more computer-readable media of claim 1, wherein the plurality of layers comprise a process layer, wherein the plurality of risk categories for the process layer comprises one or more of: unknown process execution or vulnerable packages. ### Claim 7 (depends on 1) The one or more computer-readable media of claim 1, wherein the plurality of layers comprise an artificial intelligence (AI) layer, wherein the plurality of risk categories for the AI layer comprises one or more of: insecure prompts or untrusted AI plugins. ### Claim 8 (depends on 1) The one or more computer-readable media of claim 1, wherein each of the toxic combination patterns correlates two or more risk categories across two or more layers of the plurality of layers. ### Claim 9 (depends on 1) The one or more computer-readable media of claim 1, further comprising sequence of instructions which when executed by the one or more processors cause: using a rule processor, comparing the telemetry data with each of the risk categories and each of the toxic combination patterns, the comparing producing one or more matched results, and storing the one or more matched results in a database as risk detections. ### Claim 10 (depends on 1) The one or more computer-readable media of claim 1, wherein evaluating the telemetry data against the plurality of risk categories and the plurality of toxic combination patterns is based on a set frequency. ### Claim 11 (depends on 1) The one or more computer-readable media of claim 1, further comprising sequence of instructions which when executed by the one or more processors cause: determining user preferences on prioritizing one or more toxic combination patterns of the customized set of toxic combination patterns; updating one or more risk categories of the plurality of risk categories based on preference data corresponding to previously received input signals specifying the preferences; and updating the customized set of toxic combination patterns based on the preference data. ### Claim 12 (depends on 1) The one or more computer-readable media of claim 1, further comprising sequence of instructions which when executed by the one or more processors cause: detecting, based on the telemetry data, one or more risk features associated with an application associated with the microservices; updating one or more risk categories of the plurality of risk categories based on the one or more risk features; and updating the customized set of toxic combination patterns based on the one or more risk features. ### Claim 13 (depends on 1) The one or more computer-readable media of claim 1, wherein the telemetry data includes at least one of: a request source namespace, a request destination namespace, a request application programming interface (API) endpoint, a request API method, a request authentication header, or identifying information about entities accessing any one of the microservices. ### Claim 14 (depends on 1) The one or more computer-readable media of claim 1, wherein the security-related data includes at least one of: a number of application programming interface (API) requests that are unauthenticated, security-related vulnerabilities within containers and virtual machines hosting the microservices, or security-related gaps in data storage with which the microservices interact. ### Claim 15 (depends on 1) The one or more computer-readable media of claim 1, wherein the local controller service comprises a plurality of sidecar processes, each sidecar process among the plurality of sidecar processes being generated for and corresponding to each microservice among the microservices; wherein each sidecar process among the plurality of sidecar processes is configured to have access to data flowing in and out of a corresponding microservice of the microservices. ### Claim 16 (depends on 1) The one or more computer-readable media of claim 1, wherein the local controller service comprising an application container that a container platform application programming interface (API) brings up in the computing environment, wherein the local controller service is further configured to query a metadata service of the container platform API to obtain additional security-related data including: whether data stored in a database accessed by a microservice is encrypted; whether a connection to a data service is encrypted; and how open is an access to a specific data resource based on its network configuration. ### Claim 17 (depends on 1) The one or more computer-readable media of claim 1, wherein the local controller service is programmed to automatically detect an open port of a microservice from the microservices, and output associated telemetry data based on a protocol associated with the open port. ### Claim 18 A computer system, comprising: one or more central processing units; one or more network interfaces that are configured to communicatively couple the one or more central processing units to a data communication network; and electronic digital random-access memory storing one or more sequences of stored program instructions configured for managing security of an application and which, when executed using the one or more central processing units, cause the one or more central processing units to execute, use a runtime security engine deployed within a computing environment: obtaining, from a local controller service deployed within the computing environment, telemetry data from microservices executing in the computing environment, and the telemetry data comprising security-related data; accessing a plurality of risk categories codifying insecure behavior across a plurality of layers in the computing environment; accessing a plurality of toxic combination patterns across the plurality of layers, the plurality of toxic combination patterns being generated based on the plurality of risk categories, and each of the toxic combination pattern indicating a high-severity security impact in case of an attack in the computing environment; evaluating the telemetry data against the plurality of risk categories and the plurality of toxic combination patterns, the evaluating producing a customized set of toxic combination patterns specific to the application within the computing environment; and displaying the customized set of toxic combination patterns in real time in a graphical user interface of a computer display device. ### Claim 19 (depends on 18) The computer system of claim 18, wherein the plurality of layers comprises a service layer, wherein the plurality of risk categories for the service layer comprises one or more of: lateral movement to critical services or over-permissioned services. ### Claim 20 (depends on 18) The computer system of claim 18, wherein the plurality of layers comprises an application programming interface (API) layer, wherein the plurality of risk categories for the API layer comprises one or more of: unauthenticated API access or unauthorized API access. ### Claim 21 (depends on 18) The computer system of claim 18, wherein the plurality of layers comprises a data layer, wherein the plurality of risk categories for the data layer comprises one or more of: data exfiltration to untrusted application programming interfaces (APIs) or weak database authorization. ### Claim 22 (depends on 18) The computer system of claim 18, wherein the plurality of layers comprises a container layer, wherein the plurality of risk categories for the container layer comprise one or more of: over-privileged containers or unlimited resources. ### Claim 23 (depends on 18) The computer system of claim 18, wherein the plurality of layers comprise a process layer, wherein the plurality of risk categories for the process layer comprises one or more of: unknown process execution or vulnerable packages. ### Claim 24 (depends on 18) The computer system of claim 18, wherein the plurality of layers comprise an artificial intelligence (AI) layer, wherein the plurality of risk categories for the AI layer comprises one or more of: insecure prompts or untrusted AI plugins. ### Claim 25 (depends on 18) The computer system of claim 18, wherein each of the toxic combination patterns correlates two or more risk categories across two or more layers of the plurality of layers. ### Claim 26 (depends on 18) The computer system of claim 18, further comprising sequences of instructions which, when executed using the one or more central processing units, cause the one or more central processing units to execute: using a rule processor, comparing the telemetry data with each of the risk categories and each of the toxic combination patterns, the comparing producing one or more matched results, and storing the one or more matched results in a database as risk detections. ### Claim 27 (depends on 18) The computer system of claim 18, wherein evaluating the telemetry data against the plurality of risk categories and the plurality of toxic combination patterns is based on a set frequency. ### Claim 28 (depends on 18) The computer system of claim 18, further comprising sequences of instructions which, when executed using the one or more central processing units, cause the one or more central processing units to execute: determining user preferences on prioritizing one or more toxic combination patterns of the customized set of toxic combination patterns; updating one or more risk categories of the plurality of risk categories based on preference data corresponding to previously received input signals specifying the preferences; and updating the customized set of toxic combination patterns based on the preference data. ### Claim 29 (depends on 18) The computer system of claim 18, further comprising sequences of instructions which, when executed using the one or more central processing units, cause the one or more central processing units to execute: detecting, based on the telemetry data, one or more risk features associated with an application associated with the microservices; updating one or more risk categories of the plurality of risk categories based on the one or more risk features; and updating the customized set of toxic combination patterns based on the one or more risk features. ### Claim 30 (depends on 18) The computer system of claim 18, wherein the telemetry data includes at least one of: a request source namespace, a request destination namespace, a request application programming interface (API) endpoint, a request API method, a request authentication header, or identifying information about entities accessing any one of the microservices. ### Claim 31 (depends on 18) The computer system of claim 18, wherein the security-related data includes at least one of: a number of application programming interface (API) requests that are unauthenticated, security-related vulnerabilities within containers and virtual machines hosting the microservices, or security-related gaps in data storage with which the microservices interact. ### Claim 32 (depends on 18) The computer system of claim 18, wherein the local controller service comprises a plurality of sidecar processes, each sidecar process among the plurality of sidecar processes being generated for and corresponding to each microservice among the microservices; wherein each sidecar process among the plurality of sidecar processes is configured to have access to data flowing in and out of a corresponding microservice of the microservices. ### Claim 33 (depends on 18) The computer system of claim 18, wherein the local controller service comprising an application container that a container platform application programming interface (API) brings up in the computing environment, wherein the local controller service is further configured to query a metadata service of the container platform API to obtain additional security-related data including: whether data stored in a database accessed by a microservice is encrypted; whether a connection to a data service is encrypted; and how open is an access to a specific data resource based on its network configuration. ### Claim 34 (depends on 18) The computer system of claim 18, wherein the local controller service is programmed to automatically detect an open port of a microservice from the microservices, and output associated telemetry data based on a protocol associated with the open port.