User service requirement processing method, apparatus, device, and medium

Abstract

The application discloses a user service demand processing method and device, equipment and medium, relates to the technical field of information security, and comprises the following steps: determining the user service demand of a client by a first intelligent agent, analyzing the user service demand by a second intelligent agent, determining the initial execution intention corresponding to the user service demand by the first intelligent agent by using the obtained analysis result, and transmitting the initial execution intention to a trusted execution environment; in the trusted execution environment, the initial execution intention is verified for compliance, and the initial execution intention that passes the verification is determined as the target execution intention; a corresponding execution path is generated based on the target execution intention, an execution authorization credential is generated based on the execution path, a corresponding trusted execution operation is triggered based on the execution path and the execution authorization credential, a corresponding execution result is obtained, and the execution result is transmitted to the client. The application can improve the security and trust level of the operation process of the multi-intelligent agent system.

Description

Technical Field

[0001] This invention relates to the field of information security technology, and in particular to methods, apparatus, equipment and media for processing user business requirements. Background Technology

[0002] With the development of artificial intelligence technology, multi-agent systems have been widely used in many fields due to their distributed collaboration capabilities, and the integration of large models has further improved the intelligence level of the system's automatic decision-making and scheduling.

[0003] In existing multi-agent systems, the introduction of large models allows agents to directly trigger data access and computational execution. However, key execution processes generally rely on ordinary computing environments, lacking hardware-level trust constraint mechanisms. Ordinary computing environments are vulnerable to security threats such as malicious attacks and data tampering. Furthermore, the execution paths of agents lack effective traceability and verification methods, failing to ensure compliance and integrity. Simultaneously, because the computation process is not within a trusted protection boundary, the authenticity and accuracy of the execution results are difficult to verify with authoritative credibility. In the event of data leakage or execution deviations, it is impossible to accurately pinpoint the responsible party, posing significant security risks to system operation. Therefore, the execution control phase of existing multi-agent systems after the introduction of large models suffers from core defects such as a lack of trust constraints, unverifiable execution paths, and insufficient result reliability, making it difficult to meet the high security and trustworthiness requirements of critical domains.

[0004] Therefore, improving the security and reliability of multi-agent systems is a pressing technical problem that needs to be solved. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for processing user business requirements, which can improve the security and reliability of multi-agent systems during operation. The specific solution is as follows:

[0006] Firstly, this application provides a user business requirement processing method based on a trusted execution environment and intelligent agents, applied to a multi-agent system built on a trusted execution environment, including:

[0007] The first intelligent agent determines the user's business needs of the client, the second intelligent agent analyzes the user's business needs, the first intelligent agent uses the analysis results to determine the initial execution intent corresponding to the user's business needs, and the initial execution intent is transmitted to the trusted execution environment;

[0008] In the trusted execution environment, the initial execution intent is verified for compliance using preset execution conditions, and the initial execution intent that passes the verification is determined as the target execution intent;

[0009] In the trusted execution environment, a corresponding execution path is generated based on the target execution intent, and an execution authorization credential for the target execution intent is generated based on the execution path. Based on the execution path and the execution authorization credential, and using the local business resource system, a corresponding trusted execution operation is triggered to obtain the corresponding execution result. The execution result is then transmitted to the first intelligent agent, so that the first intelligent agent can transmit the execution result to the client.

[0010] Optionally, the step of determining the user's business needs through a first intelligent agent, analyzing the user's business needs through a second intelligent agent, and determining the initial execution intent corresponding to the user's business needs through the analysis results obtained by the first intelligent agent includes:

[0011] The first intelligent agent obtains the natural language instructions sent by the client, determines the user service requirements of the client based on the natural language instructions, performs preliminary analysis on the user service requirements, and obtains the corresponding analysis results.

[0012] The second intelligent agent obtains the user service requirements sent by the first intelligent agent, analyzes the user service requirements, and transmits the analysis results back to the first intelligent agent.

[0013] The first intelligent agent generates an initial execution intent corresponding to the user's business requirements based on the parsing results and the analysis results.

[0014] Optionally, before performing compliance verification on the initial execution intent using preset execution conditions, the method further includes:

[0015] Obtain the initial execution intent sent by the first intelligent agent;

[0016] The data structure, field integrity, and version information of the initial execution intent are verified to complete the preliminary verification operation of the initial execution intent.

[0017] Optionally, the compliance verification of the initial execution intent using preset execution conditions includes:

[0018] Perform semantic consistency verification on the initial execution intent to obtain the first verification result;

[0019] The initial execution intent that passes the first verification result is determined as the first execution intent;

[0020] Based on the intent type, resource domain, and interface domain of the first execution intent, the corresponding operation rules, authorization scope, and execution constraints are determined from the preset execution conditions;

[0021] The first execution intent is verified using the operation rules, the authorization scope, and the execution constraints to obtain a second verification result. The first execution intent that passes the verification according to the second verification result is determined as the second execution intent, and an executable constraint set is generated based on the second execution intent. The executable constraint set is a set containing constraints for restricting trusted execution operations.

[0022] Optionally, generating the corresponding execution path based on the target execution intent includes:

[0023] The target execution intent is mapped to execution step nodes, and the resource call requirements and interface call requirements corresponding to the execution step nodes are determined.

[0024] Based on the input and output variables between the execution step nodes, the data dependencies between the execution step nodes are derived, and the execution order between the execution step nodes is determined using the data dependencies;

[0025] Based on the resource call requirements, the interface call requirements, the execution step nodes, and the execution order, a corresponding execution plan is generated.

[0026] The executable constraint set is used as constraint terms, and the constraint terms are injected into the execution plan to obtain the execution path.

[0027] Optionally, transmitting the execution result to the first intelligent agent, so that the first intelligent agent transmits the execution result to the client, includes:

[0028] Generate verification information for verifying the trusted execution operation, and bind the verification information, the measurement information of the trusted execution environment, and the execution result to obtain the target result;

[0029] The target result is transmitted to the first intelligent agent so that the first intelligent agent can perform a reliable verification of the target result and transmit the execution result of the verified target result to the client.

[0030] Optionally, the step of triggering a corresponding trusted execution operation based on the execution path and the execution authorization credential, and utilizing the local business resource system, includes:

[0031] Based on the execution path and the execution authorization credential, a corresponding access request is generated and transmitted to the local business resource system so that the business resource system can perform trusted verification of the access request and allow the access request to call the corresponding business resources and business execution interfaces from the business resource system after the access request is verified.

[0032] The execution path is used to trigger corresponding trusted execution operations based on the business resources and the business execution interface.

[0033] Secondly, this application provides a user service requirement processing device based on a trusted execution environment and intelligent agents, applied to a multi-agent system built on a trusted execution environment, including:

[0034] The intent transmission module is used to determine the user's business needs of the client through a first intelligent agent, analyze the user's business needs through a second intelligent agent, determine the initial execution intent corresponding to the user's business needs through the analysis results obtained by the first intelligent agent, and transmit the initial execution intent to the trusted execution environment.

[0035] The intent determination module is used to perform compliance verification on the initial execution intent using preset execution conditions in the trusted execution environment, and to determine the verified initial execution intent as the target execution intent.

[0036] The result transmission module is configured to generate a corresponding execution path based on the target execution intent in the trusted execution environment, generate an execution authorization credential for the target execution intent based on the execution path, trigger a corresponding trusted execution operation based on the execution path and the execution authorization credential and using the local business resource system to obtain the corresponding execution result, and transmit the execution result to the first intelligent agent so that the first intelligent agent can transmit the execution result to the client.

[0037] Thirdly, this application provides an electronic device, comprising:

[0038] Memory, used to store computer programs;

[0039] A processor is used to execute the computer program to implement the aforementioned user business requirement processing method based on a trusted execution environment and intelligent agents.

[0040] Fourthly, this application provides a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned user business requirement processing method based on a trusted execution environment and an intelligent agent.

[0041] In this application, a first intelligent agent determines the user's business requirements from the client, a second intelligent agent analyzes these requirements, and the first intelligent agent uses the analysis results to determine the initial execution intent corresponding to the user's business requirements. This initial execution intent is then transmitted to the trusted execution environment. In the trusted execution environment, the initial execution intent is validated for compliance using preset execution conditions, and the validated initial execution intent is determined as the target execution intent. In the trusted execution environment, a corresponding execution path is generated based on the target execution intent, and an execution authorization credential for the target execution intent is generated based on the execution path. Based on the execution path and the execution authorization credential, and using the local business resource system, a corresponding trusted execution operation is triggered to obtain the corresponding execution result. This execution result is then transmitted to the first intelligent agent, which then transmits the execution result to the client. As can be seen, in this application, the first intelligent agent obtains the user's business requirements from the client, which are then analyzed by the second intelligent agent. Based on the aforementioned analysis results, the first intelligent agent determines the initial execution intent matching the user's business requirements and sends this initial execution intent to the trusted execution environment. Within a trusted execution environment, the initial execution intent is verified for compliance based on preset execution conditions, and the verified initial execution intent is defined as the target execution intent. Subsequently, within this trusted execution environment, a corresponding execution path is generated based on the target execution intent, and an execution authorization credential corresponding to the target execution intent is generated based on this execution path. Combining the execution path and the execution authorization credential, the local business resource system is invoked to trigger the corresponding trusted execution operation. After obtaining the corresponding execution result, it is fed back to the first intelligent agent, which then pushes the execution result to the client. In this way, this application can improve the security and trustworthiness of the multi-agent system operation process. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0043] Figure 1 This application discloses a flowchart of a user business requirement processing method based on a trusted execution environment and an intelligent agent.

[0044] Figure 2 This is a system architecture diagram of a user business requirement processing method based on a trusted execution environment and intelligent agent disclosed in this application;

[0045] Figure 3 This application discloses a processing flowchart for a trusted execution environment.

[0046] Figure 4 This is a schematic diagram of a user business demand processing device based on a trusted execution environment and an intelligent agent disclosed in this application;

[0047] Figure 5 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Currently, in existing multi-agent systems, after introducing large models, agents can directly trigger data access and computational execution operations. However, their key execution processes generally rely on ordinary computing environments, lacking hardware-level trust constraint mechanisms. Ordinary computing environments are vulnerable to security threats such as malicious attacks and data tampering. The execution paths of agents lack effective traceability and verification methods, failing to ensure the compliance and integrity of the execution process. Furthermore, because the computation process is not within a trusted protection boundary, the authenticity and accuracy of the execution results are difficult to verify with authoritative credibility. In the event of data leakage or execution deviations, it is impossible to accurately pinpoint the responsible party, posing significant security risks to system operation. Therefore, the execution control phase of existing multi-agent systems after introducing large models suffers from core defects such as a lack of trust constraints, unverifiable execution paths, and insufficient result credibility, making it difficult to meet the high security and trustworthiness requirements of critical domains. To address this, this application provides user business requirement processing methods, devices, equipment, and media that can improve the security and trustworthiness of multi-agent system operation.

[0050] See Figure 1 As shown, this embodiment of the invention discloses a user business requirement processing method based on a trusted execution environment and intelligent agents, applied to a multi-agent system built on a trusted execution environment, including:

[0051] Step S11: Determine the user's business requirements of the client through the first intelligent agent, analyze the user's business requirements through the second intelligent agent, determine the initial execution intent corresponding to the user's business requirements through the analysis results obtained by the first intelligent agent, and transmit the initial execution intent to the trusted execution environment.

[0052] In this embodiment, the first intelligent agent serves as the interface for direct interaction with the client, responsible for receiving and processing service requests initiated by the client. Therefore, the first intelligent agent obtains the natural language instructions sent by the client, determines the client's user service requirements based on these instructions, performs preliminary parsing of the user service requirements, and obtains corresponding parsing results. Subsequently, the second intelligent agent obtains the user service requirements sent by the first intelligent agent, analyzes these requirements, and transmits the analysis results back to the first intelligent agent.

[0053] After receiving the analysis results returned by the second agent, the first agent generates an initial execution intent corresponding to the user's business requirements based on the parsing and analysis results. After generating the initial execution intent, the first agent securely transmits it to a trusted execution environment, laying the foundation for subsequent intent verification, optimization, and reliable execution within a protected environment.

[0054] Step S12: In the trusted execution environment, the initial execution intent is verified for compliance using preset execution conditions, and the verified initial execution intent is determined as the target execution intent.

[0055] In this embodiment, after the trusted execution environment receives the initial execution intent transmitted by the first agent, it obtains the initial execution intent sent by the first agent; it verifies the data structure, field integrity, and version information of the initial execution intent to complete the preliminary verification operation of the initial execution intent, thereby ensuring the format validity and basic reliability of the objects to be processed subsequently.

[0056] After initial verification, semantic consistency verification is performed on the initial execution intent to obtain a first verification result. The initial execution intent that passes the first verification result is determined as the first execution intent. Based on the intent type, resource domain, and interface domain of the first execution intent, corresponding operation rules, authorization scope, and execution constraints are determined from the preset execution conditions. The first execution intent is verified using the operation rules, authorization scope, and execution constraints to obtain a second verification result. The first execution intent that passes the second verification result is determined as the second execution intent, and an executable constraint set is generated based on the second execution intent. The executable constraint set is a set containing restrictions used to constrain trusted execution operations. Only when the initial execution intent passes compliance checks at all levels, i.e., is determined to be safe, legal, and compliant with the policy, will the trusted execution environment ultimately determine it as the target execution intent, providing verified and reliable input for subsequent execution stages.

[0057] Step S13: In the trusted execution environment, a corresponding execution path is generated based on the target execution intent, and an execution authorization credential for the target execution intent is generated based on the execution path. Based on the execution path and the execution authorization credential, and using the local business resource system, a corresponding trusted execution operation is triggered to obtain the corresponding execution result. The execution result is then transmitted to the first intelligent agent so that the first intelligent agent can transmit the execution result to the client.

[0058] In this embodiment, after obtaining the verified target execution intent, the trusted execution environment maps the target execution intent to execution step nodes and determines the resource call requirements and interface call requirements corresponding to the execution step nodes. Based on the input and output variables between the execution step nodes, it derives the data dependencies between the execution step nodes and uses the data dependencies to determine the execution order between the execution step nodes. Based on the resource call requirements, the interface call requirements, the execution step nodes, and the execution order, it generates a corresponding execution plan. The executable constraint set is used as constraint terms, and the constraint terms are injected into the execution plan to obtain an execution path. This path may include the operation nodes and data flow required from the start to the completion of the target intent.

[0059] Furthermore, based on the determined execution path, the trusted execution environment generates an execution authorization credential uniquely bound to it. Subsequently, based on the execution path and the execution authorization credential, a corresponding access request is generated and transmitted to the local business resource system. The business resource system then performs trusted verification on the access request and, upon successful verification, allows the access request to invoke the corresponding business resources and business execution interfaces from the business resource system. Utilizing the execution path and based on the business resources and business execution interfaces, corresponding trusted execution operations are triggered to complete the entire processing flow corresponding to the user's business requirements.

[0060] To ensure the auditability and integrity of the execution process and results, the trusted execution environment generates verification information to verify the trusted execution operation after the operation is completed. The verification information, the measurement information of the trusted execution environment, and the execution result are bound together to obtain the target result. The target result is then transmitted to the first intelligent agent so that the first intelligent agent can perform trusted verification on the target result. The execution result in the verified target result is then transmitted to the client, thereby completing the secure, trusted, and closed-loop processing of user business requirements.

[0061] Furthermore, it should be emphasized that although steps S12 and S13 are described in this embodiment based on a trusted execution environment, the compliance verification and authorized execution operations they embody are universal, meaning that a trusted execution environment can be considered an enhancement option. In practical implementation, these steps can also be adapted to other systems or architectures with sufficient security guarantees. As long as the system can provide equivalent data isolation, integrity protection, and trusted measurement capabilities, similar secure execution effects can be achieved.

[0062] As can be seen from the above, in this application, the first intelligent agent obtains the user's business requirements from the client and hands them over to the second intelligent agent for analysis. Based on the analysis results, the first intelligent agent determines the initial execution intent that matches the user's business requirements and sends the initial execution intent to the trusted execution environment. Within the trusted execution environment, the initial execution intent undergoes compliance verification according to preset execution conditions, and the verified initial execution intent is defined as the target execution intent. Subsequently, within the trusted execution environment, a corresponding execution path is generated based on the target execution intent, and an execution authorization credential corresponding to the target execution intent is generated based on the execution path. Combining the execution path and the execution authorization credential, the local business resource system is invoked to trigger the corresponding trusted execution operation. After obtaining the corresponding execution result, it is fed back to the first intelligent agent, which then pushes the execution result to the client. In this way, this application can improve the security and trustworthiness of the multi-agent system operation process.

[0063] The following is combined Figure 2 and Figure 3 The schematic diagram shown illustrates the technical solution of the embodiments of this application in detail.

[0064] In this application, to address the problem that in multi-agent systems, after introducing large models for automatic decision-making and scheduling, agents can directly trigger data access and computation execution, and the execution process relies on ordinary computing environments and lacks hardware-level trusted constraints, a trusted execution control layer is introduced into the system. This layer unifies the execution requests generated by multiple agents into execution intentions, and the trusted execution environment verifies, schedules, and controls the execution intentions, thereby achieving unavoidable trusted control over key data operations and computation processes.

[0065] Furthermore, the multi-agent system is only responsible for parsing user business requirements and generating execution intentions, without directly performing data access or computation operations. All key execution processes involving data reading, data processing, and result generation are completed through a trusted execution environment or under the control of a trusted execution environment, making the execution path verifiable, the execution process immutable, and the execution results credible proof, thereby significantly improving the operational reliability and security of the multi-agent system.

[0066] Reference Figure 2 This application adopts a top-down layered architecture design, which is divided into three logical layers: Agent Decision Layer (ADL), Trusted Execution Control Layer (TECL), and Resource & Execution Layer (REL). In this layered architecture, the execution intent serves as the core carrier of cross-layer interaction and flows between the Agent Decision Layer, Trusted Execution Control Layer, and Resource & Execution Layer, thereby achieving unified control over the execution process of the multi-agent system.

[0067] The intelligent agent decision layer consists of a client agent (CA) and at least one function agent (FA). Located at the top layer of the application system, it receives natural language commands from the user and generates execution intentions to trigger subsequent execution processes. The client agent, serving as the user interaction entry point, is responsible for identifying, initially parsing, and uniformly generating execution intentions for the user's business needs. The function agent, under the scheduling of the client agent, performs task decomposition, business analysis, or parameter supplementation of the user's business needs and feeds the analysis results back to the client agent.

[0068] The intelligent agent decision layer interacts with the control program in the trusted execution environment through the client intelligent agent, submitting the generated execution intent to the trusted execution control layer to trigger subsequent verification, scheduling, and execution processes. Furthermore, neither the client intelligent agent nor the functional intelligent agent has direct access to resources and the execution layer or the authority to trigger execution operations; all their outputs serve only as a basis for generating or supplementing execution intents.

[0069] The Trusted Execution Control Layer is deployed in the Trusted Execution Environment (TEE) and is used to perform unified trusted verification, trusted scheduling and controlled execution of the execution intentions generated by the agent's decision layer. It is the only trusted execution entry point for all data access, data processing and result generation operations in the system, and constitutes the unbypassable control center for the system's key execution operations.

[0070] The trusted execution control layer includes at least one or a combination of an execution intent receiving module, an execution intent verification module, a trusted scheduling module, a gating control module, a trusted execution module, and a trusted result encapsulation module. Each module works collaboratively in the trusted execution environment to jointly control and execute the execution intent.

[0071] The execution intent receiving module receives execution intent requests submitted by the agent's decision layer and performs basic processing on the execution intents, including but not limited to verifying the data structure, field integrity, and version information of the execution intents. This module does not make a final judgment on the legality of the execution intents; it only serves as the entry node for the internal processing flow of the trusted execution control layer.

[0072] The execution intent verification module is used to verify the legitimacy of received execution intents in a trusted execution environment to determine whether the execution intent meets preset execution conditions. Legitimacy verification may include semantic consistency verification, preset rule verification to determine whether the execution intent conforms to the system's established usage rules or policies, authorization condition verification to determine whether the execution intent is within the allowed authorized scope, and execution constraint verification to determine whether the execution intent meets constraints such as purpose, scope, and timeliness. Furthermore, upon successful execution intent verification, a corresponding set of executable constraints can be generated. When an execution intent fails verification, the trusted execution control layer refuses to trigger subsequent execution processes and returns corresponding rejection information.

[0073] The trusted scheduling module, after the execution intent has been verified, performs path decision-making and execution planning for the execution process corresponding to the execution intent. Specifically, it generates the corresponding execution path based on the execution intent and the set of executable constraints generated by the execution intent verification module. Specifically, the trusted scheduling module determines the execution path corresponding to the execution intent, including but not limited to: the resources or interfaces to be called, the sequential relationship between execution steps, and the dependencies between execution steps. The trusted scheduling module is only responsible for determining the execution path and does not directly participate in specific data access or computational execution operations. Its scheduling decision-making process is completed within the trusted execution environment, and external entities cannot interfere with it.

[0074] The gating control module is used to manage access credentials (such as keys, tokens, certificates, access accounts, etc.) of resources and execution layer, and provides controlled access authorization that matches the execution intent when the trusted execution module performs specific execution, thereby ensuring that access to resources and execution layer can only occur under the control of the trusted execution environment, realizing unbypassable gating.

[0075] Furthermore, the gating control module should possess the following capabilities: Within the Trusted Execution Environment (TEE), the gating control module should manage and store the long-term credentials required to access the Reliable Elastic Execution System (REL). These long-term credentials include, but are not limited to, database access keys, API (Application Programming Interface) call keys, token refresh keys, and private key certificates. These long-term credentials can only be decrypted and used by the gating control module within the TEE. External entities and the decision-making layer of the intelligent agent cannot read or export these long-term credentials, thus preventing intelligent agents or external systems from bypassing the TECL to directly access the REL even if they have network reachability.

[0076] After the execution intent is verified by the execution intent verification module and the trusted scheduling module generates an execution plan, the gating control module generates a controlled authorization carrier based on the verification result and the execution plan. The controlled authorization carrier includes at least a short-term access token or a one-time session key, and a capability token describing the scope of resources allowed to be accessed, the types of actions allowed to be performed, and the execution constraints that must be met. The capability token is generated within the TEE and signed or encrypted by the gating control module using an internal TEE key, making it impossible for external entities to forge or tamper with it.

[0077] The gating control module outputs the controlled authorization carrier or short-term access material to the trusted execution module. The trusted execution module must carry the controlled authorization carrier when accessing REL's data resources or service interfaces, or initiate an access request through the gating control module. During execution, the gating control module can further inject constraints or solidify parameters on the trusted execution module's access requests.

[0078] The trusted execution module is used to complete specific execution operations within the trusted execution environment according to the execution path determined by the trusted scheduling module. These execution operations may include data reading or access, data processing, and statistical or computational operations. During execution, all raw data and intermediate results involved are isolated and protected by the trusted execution environment, and are not exposed to entities outside of it.

[0079] The trusted result encapsulation module encapsulates the execution results generated by the trusted execution module to form verifiable output results. The encapsulation process may include: binding the execution results with measurement information from the trusted execution environment, generating verification information to validate the trustworthiness of the execution process, and formatting the execution results for output. Through the trusted result encapsulation module, the execution results returned to the agent's decision layer become verifiable, thereby supporting subsequent result display, auditing, or traceability.

[0080] Additionally, refer to Figure 3As shown, the Resource and Execution Layer (REL) is located at the bottom layer of the system architecture. It carries and provides data resources, business capabilities, and execution interfaces for specific business systems, serving as the foundational support layer for multi-agent systems to complete actual business operations. The Resource and Execution Layer is not an independently operating intelligent system, but rather a unified abstraction and access mechanism for multiple heterogeneous business systems within the architecture of this invention. In this invention, the Resource and Execution Layer does not directly expose access interfaces to the agent decision-making layer; all its call requests are initiated through the Trusted Execution Control Layer, thereby ensuring that the access path of the business system is always under the control of the Trusted Execution Environment.

[0081] The resource and execution layer also includes a resource access gate module, which is used to perform trusted verification and constraint enforcement on access requests entering the REL, thereby preventing the intelligent agent decision layer or external entities of the trusted execution environment from bypassing the trusted execution control layer to directly trigger data access or business interface calls.

[0082] The technical solutions of the embodiments of this application will be described below in conjunction with specific application scenarios.

[0083] In one specific implementation, within the context of healthcare consultation, a user's business needs might manifest as natural language commands sent through a personal health assistant client, such as "Assess the urgency of my recent chest pain symptoms and recommend local hospitals where I can make an appointment." In such scenarios involving sensitive personal health information and professional medical decision-making, it is necessary to balance intent understanding, medical analysis, data compliance, and resource allocation.

[0084] In this scenario, the first intelligent agent, acting as the user interaction terminal, parses the instruction and identifies its core needs as "symptom urgency assessment" and "medical resource integration." The second intelligent agent, based on a professional medical knowledge graph and triage rules, performs in-depth analysis of the symptom description, initially classifying it as "cardiovascular-related symptoms requiring priority screening," thus generating an analysis result. Subsequently, the first intelligent agent integrates the information from both sources to generate an initial execution intent: "obtain the user's historical health records, invoke the triage model for assessment, query real-time appointment availability at designated regional hospitals, and generate a recommendation list." This intent, within a trusted execution environment, must undergo compliance verification in accordance with pre-defined medical data privacy protection regulations, cross-institutional data access authorization agreements, and ethical guidelines for medical recommendation services. Once verified, the generated target execution intent guides the secure invocation of the user's encrypted electronic medical records, certified diagnostic model interfaces, and regional medical resource platforms within the trusted execution environment. Ultimately, it generates a verifiable report containing risk assessment levels, recommended departments, and a specific hospital recommendation list, which is then securely fed back to the personal health assistant client via the first intelligent agent.

[0085] Accordingly, see Figure 4As shown, this application provides a user business requirement processing device based on a trusted execution environment and intelligent agents, applied to a multi-agent system built on a trusted execution environment, including:

[0086] The intent transmission module 11 is used to determine the user's business needs of the client through a first intelligent agent, analyze the user's business needs through a second intelligent agent, determine the initial execution intent corresponding to the user's business needs through the analysis results obtained by the first intelligent agent, and transmit the initial execution intent to the trusted execution environment.

[0087] The intent determination module 12 is used to perform compliance verification on the initial execution intent using preset execution conditions in the trusted execution environment, and to determine the verified initial execution intent as the target execution intent;

[0088] The result transmission module 13 is configured to generate a corresponding execution path based on the target execution intent in the trusted execution environment, generate an execution authorization credential about the target execution intent based on the execution path, trigger a corresponding trusted execution operation based on the execution path and the execution authorization credential and using the local business resource system to obtain the corresponding execution result, and transmit the execution result to the first intelligent agent so that the first intelligent agent can transmit the execution result to the client.

[0089] In some specific embodiments, the intent transmission module 11 specifically includes:

[0090] The requirement parsing unit is used to obtain the natural language instructions sent by the client through the first intelligent agent, determine the user service requirements of the client based on the natural language instructions, perform preliminary parsing of the user service requirements, and obtain the corresponding parsing results.

[0091] The first result transmission unit is used to obtain the user service requirements sent by the first intelligent agent through the second intelligent agent, analyze the user service requirements, and transmit the analysis results to the first intelligent agent.

[0092] The intent generation unit is used to generate an initial execution intent corresponding to the user's business needs based on the parsing results and the analysis results by the first intelligent agent.

[0093] In some specific embodiments, the intent determination module 12 further includes:

[0094] An intent acquisition unit is used to acquire the initial execution intent sent by the first intelligent agent;

[0095] The first intent verification unit is used to verify the data structure, field integrity, and version information of the initial execution intent in order to complete the preliminary verification operation of the initial execution intent.

[0096] In some specific embodiments, the intent determination module 12 specifically includes:

[0097] The second intent verification unit is used to perform semantic consistency verification on the initial execution intent and obtain the first verification result.

[0098] An intent determination unit is used to determine the initial execution intent that has passed the first verification result as the first execution intent;

[0099] The information determination unit is used to determine the corresponding operation rules, authorization scope and execution constraints from the preset execution conditions based on the intent type, resource domain and interface domain of the first execution intent;

[0100] The third intent verification unit is used to verify the first execution intent using the operation rules, the authorization scope, and the execution constraints to obtain a second verification result, so as to determine the first execution intent that has passed the verification as the second execution intent, and generate an executable constraint set based on the second execution intent; the executable constraint set is a set containing restrictive conditions for constraining trusted execution operations.

[0101] In some specific embodiments, the result transmission module 13 specifically includes:

[0102] The requirement determination unit is used to map the target execution intent to execution step nodes, and to determine the resource call requirements and interface call requirements corresponding to the execution step nodes.

[0103] The sequence determination unit is used to deduce the data dependency relationship between the execution step nodes based on the input and output variables between the execution step nodes, and to determine the execution order between the execution step nodes using the data dependency relationship;

[0104] The plan generation unit is used to generate a corresponding execution plan based on the resource call requirements, the interface call requirements, the execution step nodes, and the execution order.

[0105] The path determination unit is used to take the set of executable constraints as constraints and inject the constraints into the execution plan to obtain the execution path.

[0106] In some specific embodiments, the result transmission module 13 specifically includes:

[0107] An information binding unit is used to generate verification information for verifying the trusted execution operation, and bind the verification information, the measurement information of the trusted execution environment, and the execution result to obtain the target result;

[0108] The second result transmission unit is used to transmit the target result to the first intelligent agent so that the first intelligent agent can perform a reliable verification of the target result and transmit the execution result of the verified target result to the client.

[0109] In some specific embodiments, the result transmission module 13 specifically includes:

[0110] A trusted verification unit is used to generate a corresponding access request based on the execution path and the execution authorization credential, and transmit the access request to the local business resource system so that the business resource system can perform trusted verification on the access request, and allow the access request to call the corresponding business resources and business execution interfaces from the business resource system after the access request is verified.

[0111] An operation triggering unit is used to trigger a corresponding trusted execution operation based on the execution path, the business resources, and the business execution interface.

[0112] Furthermore, embodiments of this application also disclose an electronic device, Figure 5 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the user service requirement processing method based on a trusted execution environment and intelligent agent disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0113] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0114] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0115] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including computer programs capable of performing the user business requirement processing method based on a trusted execution environment and intelligent agent, which is executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0116] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned user business requirement processing method based on a trusted execution environment and intelligent agents. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0117] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0118] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0119] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0120] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0121] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for processing user business requirements based on a trusted execution environment and an intelligent agent, characterized in that, Applications to multi-agent systems built on trusted execution environments include: The first intelligent agent determines the user's business needs of the client, the second intelligent agent analyzes the user's business needs, the first intelligent agent uses the analysis results to determine the initial execution intent corresponding to the user's business needs, and the initial execution intent is transmitted to the trusted execution environment; In the trusted execution environment, the initial execution intent is verified for compliance using preset execution conditions, and the initial execution intent that passes the verification is determined as the target execution intent; In the trusted execution environment, a corresponding execution path is generated based on the target execution intent, and an execution authorization credential for the target execution intent is generated based on the execution path. Based on the execution path and the execution authorization credential, and using the local business resource system, a corresponding trusted execution operation is triggered to obtain the corresponding execution result. The execution result is then transmitted to the first intelligent agent, so that the first intelligent agent can transmit the execution result to the client. The compliance verification of the initial execution intent using preset execution conditions includes: Perform semantic consistency verification on the initial execution intent to obtain the first verification result; The initial execution intent that passes the first verification result is determined as the first execution intent; Based on the intent type, resource domain, and interface domain of the first execution intent, the corresponding operation rules, authorization scope, and execution constraints are determined from the preset execution conditions; The first execution intent is verified using the operation rules, the authorization scope, and the execution constraints to obtain a second verification result. The first execution intent that passes the verification according to the second verification result is determined as the second execution intent, and an executable constraint set is generated based on the second execution intent. The executable constraint set is a set containing constraints for restricting trusted execution operations. The step of generating a corresponding execution path based on the target execution intent includes: The target execution intent is mapped to execution step nodes, and the resource call requirements and interface call requirements corresponding to the execution step nodes are determined. Based on the input and output variables between the execution step nodes, the data dependencies between the execution step nodes are derived, and the execution order between the execution step nodes is determined using the data dependencies; Based on the resource call requirements, the interface call requirements, the execution step nodes, and the execution order, a corresponding execution plan is generated. The executable constraint set is used as constraint terms, and the constraint terms are injected into the execution plan to obtain the execution path.

2. The user service requirement processing method based on a trusted execution environment and intelligent agent according to claim 1, characterized in that, The process of determining the user's service requirements through a first intelligent agent, analyzing the user's service requirements through a second intelligent agent, and determining the initial execution intent corresponding to the user's service requirements through the analysis results obtained by the first intelligent agent includes: The first intelligent agent obtains the natural language instructions sent by the client, determines the user service requirements of the client based on the natural language instructions, performs preliminary analysis on the user service requirements, and obtains the corresponding analysis results. The second intelligent agent obtains the user service requirements sent by the first intelligent agent, analyzes the user service requirements, and transmits the analysis results back to the first intelligent agent. The first intelligent agent generates an initial execution intent corresponding to the user's business requirements based on the parsing results and the analysis results.

3. The user service requirement processing method based on a trusted execution environment and intelligent agent according to claim 1, characterized in that, Before performing compliance verification on the initial execution intent using preset execution conditions, the method further includes: Obtain the initial execution intent sent by the first intelligent agent; The data structure, field integrity, and version information of the initial execution intent are verified to complete the preliminary verification operation of the initial execution intent.

4. The user service requirement processing method based on a trusted execution environment and intelligent agent according to claim 1, characterized in that, The step of transmitting the execution result to the first intelligent agent, so that the first intelligent agent transmits the execution result to the client, includes: Generate verification information for verifying the trusted execution operation, and bind the verification information, the measurement information of the trusted execution environment, and the execution result to obtain the target result; The target result is transmitted to the first intelligent agent so that the first intelligent agent can perform a reliable verification of the target result and transmit the execution result of the verified target result to the client.

5. The user service requirement processing method based on a trusted execution environment and intelligent agent according to any one of claims 1 to 4, characterized in that, The step of triggering a corresponding trusted execution operation based on the execution path and the execution authorization credential, and utilizing the local business resource system, includes: Based on the execution path and the execution authorization credential, a corresponding access request is generated and transmitted to the local business resource system so that the business resource system can perform trusted verification of the access request and allow the access request to call the corresponding business resources and business execution interfaces from the business resource system after the access request is verified. The execution path is used to trigger corresponding trusted execution operations based on the business resources and the business execution interface.

6. A user service requirement processing device based on a trusted execution environment and an intelligent agent, characterized in that, Applications to multi-agent systems built on trusted execution environments include: The intent transmission module is used to determine the user's business needs of the client through a first intelligent agent, analyze the user's business needs through a second intelligent agent, determine the initial execution intent corresponding to the user's business needs through the analysis results obtained by the first intelligent agent, and transmit the initial execution intent to the trusted execution environment. The intent determination module is used to perform compliance verification on the initial execution intent using preset execution conditions in the trusted execution environment, and to determine the verified initial execution intent as the target execution intent. The result transmission module is configured to generate a corresponding execution path based on the target execution intent in the trusted execution environment, generate an execution authorization credential for the target execution intent based on the execution path, trigger a corresponding trusted execution operation based on the execution path and the execution authorization credential and using the local business resource system to obtain the corresponding execution result, and transmit the execution result to the first intelligent agent so that the first intelligent agent can transmit the execution result to the client. The intent determination module further includes: The second intent verification unit is used to perform semantic consistency verification on the initial execution intent and obtain the first verification result. An intent determination unit is used to determine the initial execution intent that has passed the first verification result as the first execution intent; The information determination unit is used to determine the corresponding operation rules, authorization scope and execution constraints from the preset execution conditions based on the intent type, resource domain and interface domain of the first execution intent; The third intent verification unit is used to verify the first execution intent using the operation rules, the authorization scope, and the execution constraints to obtain a second verification result, so as to determine the first execution intent that passes the verification as the second execution intent, and generate an executable constraint set based on the second execution intent; the executable constraint set is a set containing restrictive conditions for constraining trusted execution operations; The result transmission module includes: The requirement determination unit is used to map the target execution intent to execution step nodes, and to determine the resource call requirements and interface call requirements corresponding to the execution step nodes. The sequence determination unit is used to deduce the data dependency relationship between the execution step nodes based on the input and output variables between the execution step nodes, and to determine the execution order between the execution step nodes using the data dependency relationship; The plan generation unit is used to generate a corresponding execution plan based on the resource call requirements, the interface call requirements, the execution step nodes, and the execution order. The path determination unit is used to take the set of executable constraints as constraints and inject the constraints into the execution plan to obtain the execution path.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the user business requirement processing method based on a trusted execution environment and an intelligent agent as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, Used to store computer programs; wherein, when the computer programs are executed by a processor, they implement the user business requirement processing method based on a trusted execution environment and an intelligent agent as described in any one of claims 1 to 5.

Images