Network security penetration testing management method and system based on artificial intelligence

Abstract

The present invention provides an artificial intelligence-based network security penetration test management method and system, which relate to the technical field of network security testing. The present invention ensures the accuracy and completeness of initial task data through a strict data collection and verification process, reduces repeated testing and correction work caused by data problems, and reduces testing costs and time costs, thereby improving the reliability of the entire testing process, enhancing the ability to discover security vulnerabilities, and making the generated test report more detailed and accurate, with a higher reference value, thereby enhancing customers' trust in penetration testing services and providing a more solid foundation for network security protection.

Description

Technical Field

[0001] The present invention relates to the technical field of network security testing, and in particular to a network security penetration testing management method and system based on artificial intelligence. Background Art

[0002] With the rapid development of information technology, the security threats in cyberspace are becoming increasingly severe. As an important means of evaluating system security, network security penetration testing plays a vital role in ensuring information security. However, traditional network security penetration testing faces many challenges and limitations in practice, which seriously restrict the effectiveness and accuracy of the test.

[0003] Traditional penetration testing often relies on manual collection and processing of task test data. This process is not only time-consuming and labor-intensive, but also easily affected by human factors, making it difficult to ensure the accuracy and integrity of the data. Errors or omissions in the data may cause the test process to deviate from the expected goal, or even fail to discover key security vulnerabilities, greatly reducing the reliability and effectiveness of the test. Traditional penetration testing mostly uses manual or semi-automatic methods in data collection, vulnerability scanning, penetration testing and other links, which not only increases the workload of testers, but also limits the speed and scale of testing. With the increasing complexity and size of network systems, traditional testing methods can no longer meet the needs of fast, efficient and comprehensive testing.

[0004] Therefore, it is necessary to provide an artificial intelligence-based network security penetration testing management method and system to solve the above technical problems. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an artificial intelligence-based network security penetration testing management method and system for solving the problem of low testing efficiency caused by the difficulty in ensuring the accuracy and integrity of data.

[0006] The present invention provides an artificial intelligence-based network security penetration test management method, the management method comprising:

[0007] S100, collecting task test data for the test system, and verifying the task test data to obtain determined task test data;

[0008] The specific steps for collecting task test data for the test system and verifying the task test data are as follows:

[0009] Acquire data from different channels and use real-time monitoring technology to monitor data collection, transmission and storage in real time to identify abnormal data during the acquisition process;

[0010] Use data comparison tools to compare the data to be verified from different channels with known accurate data sets, identify differences and erroneous data, and integrate them into a preliminary erroneous data set;

[0011] Analyze the correlation of the same type of data based on historical data and establish a data correlation prediction model;

[0012] Taking the identified abnormal data and the preliminary error data set as input, the associated error data in each type of data is estimated through the established data association estimation model to form an associated error data set;

[0013] Review the identified abnormal data, preliminary error data sets, and associated error data sets that have been identified and integrated, and correct the confirmed error data based on the review results to obtain confirmed task test data;

[0014] S200, pairing an automated scanning tool with the task test data, and scanning and detecting the task data through the automated scanning tool to obtain vulnerability scanning results;

[0015] S300, based on the obtained vulnerability scanning results, matching the vulnerability with a corresponding attack code to attack and obtain a penetration test result;

[0016] S400. Evaluate the security risk level of the test system based on the obtained penetration test results, generate a test report and send it to the customer, wherein the penetration test results specifically include vulnerability information, attack paths and sensitive data.

[0017] Preferably, the specific steps of acquiring data from different channels, and using real-time monitoring technology to monitor the collection, transmission and storage of data in real time, and identifying abnormal data in the acquisition process are:

[0018] Collect the data required for the current task test from multiple channels, obtain multi-source data, and deploy real-time monitoring mechanisms in each link of multi-source data collection, transmission, and storage;

[0019] During real-time monitoring, the preset rules are used to identify abnormal behavior data during data collection, transmission and storage, including but not limited to data tampering, loss and delay;

[0020] The abnormal behavior data is classified, organized and stored to obtain abnormal data.

[0021] Preferably, the specific steps of using a data comparison tool to compare the data to be verified obtained from different channels with known accurate data sets, identifying differences and erroneous data, and integrating them into a preliminary erroneous data set are:

[0022] Filter out the data set to be verified from multiple sources and prepare a known accurate data set as a comparison benchmark;

[0023] Select the corresponding data comparison tool according to the data type and comparison requirements;

[0024] Use the selected data comparison tool to compare the data set to be verified with the known accurate data set, including the consistency, completeness and accuracy of the comparison data, to obtain the comparison results;

[0025] Based on the comparison results, the differences and errors between the data were identified and integrated into a preliminary error dataset.

[0026] Preferably, the specific steps of analyzing the correlation of the same type of data based on historical data and establishing a data correlation prediction model are:

[0027] Collect historical data of the same type as the data required for the current task test and extract basic features from it through data mining technology;

[0028] Based on the extracted features, the association relationship between the same type of data is obtained. The association relationship is used as a training set, and a neural network model is selected for training to obtain a data association prediction model.

[0029] Preferably, the specific steps of taking the identified abnormal data and the preliminary error data set as input, estimating the associated error data in each type of data through the established data association estimation model, and forming the associated error data set are:

[0030] Input the identified abnormal data and preliminary error data set into the data association prediction model, and use the model to estimate the associated error data in each type of data;

[0031] Remove duplicate and invalid data that are estimated to be associated with erroneous data in each type of data;

[0032] The formed associated error data sets are identified and integrated into associated error data sets.

[0033] Preferably, the specific steps of reviewing the identified abnormal data, the preliminary error data set, and the associated error data set integrated with the identification, and correcting the confirmed error data according to the review results to obtain the determined task test data are:

[0034] Preset audit rules and write custom scripts based on the audit rules. The preset audit rules specifically include data consistency rules, data integrity rules, data accuracy rules, and business logic rules;

[0035] Import the identified abnormal data, preliminary error data set, and associated error data set integrated with identification into the written custom script, and verify them one by one through the custom script;

[0036] Correct the data that fails the verification to obtain the confirmed task test data.

[0037] Network security penetration testing management based on artificial intelligence, the management system includes:

[0038] A data verification module is used to collect task test data for the test system, verify the task test data, and obtain confirmed task test data;

[0039] The scanning and detection module is used to pair the automated scanning tool with the task test data, and scan and detect the task data through the automated scanning tool to obtain the vulnerability scanning results;

[0040] The vulnerability attack module is used to match the corresponding attack code to the vulnerability based on the obtained vulnerability scanning results, and then attack to obtain the penetration test results;

[0041] The report generation module is used to evaluate the security risk level of the test system based on the obtained penetration test results, generate a test report and send it to the customer, where the penetration test results specifically include vulnerability information, attack paths and sensitive data.

[0042] Compared with related technologies, the network security penetration test management method and system based on artificial intelligence provided by the present invention have the following beneficial effects:

[0043] Through a strict data collection and verification process, the present invention ensures the accuracy and completeness of the initial task data, reduces repeated testing and correction work caused by data problems, reduces testing costs and time costs, thereby improving the reliability of the entire testing process, enhancing the ability to discover security vulnerabilities, and making the generated test reports more detailed and accurate, with higher reference value, thereby enhancing customers' trust in penetration testing services and providing a more solid foundation for network security protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flowchart of the network security penetration test management method based on artificial intelligence of the present invention;

[0045] Figure 2 This is a system block diagram of the artificial intelligence-based network security penetration testing management system of the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0047] Embodiment 1

[0048] like Figure 1 As shown, the network security penetration test management method based on artificial intelligence includes:

[0049] The management method comprises:

[0050] S100, collecting task test data for the test system, and verifying the task test data to obtain determined task test data;

[0051] The specific steps for collecting task test data for the test system and verifying the task test data are as follows:

[0052] S101. Acquire data from different channels, and use real-time monitoring technology to monitor data collection, transmission, and storage in real time, and identify abnormal data in the acquisition process;

[0053] In the specific implementation process, the specific steps of step S101 are:

[0054] S1011. Collect data required for current task testing from multiple channels, obtain multi-source data, and deploy real-time monitoring mechanisms in each link of multi-source data collection, transmission, and storage;

[0055] Specifically, clarify the data type, scope and quality standard required for the current task test, including but not limited to user behavior data, transaction records, system logs, social media feedback, etc., and then use API calls, crawler technology, database queries, etc. to collect data from multiple channels, and select corresponding real-time monitoring technologies and tools according to the data type, transmission protocol and storage method, including log analysis, network traffic monitoring, database triggers, message queue monitoring, etc.; deploy monitoring points at the data collection end, record the time, source, data volume and other information of each data collection, and monitor the collection frequency and success rate in real time to ensure data integrity, use network monitoring tools to track the data transmission process in real time, including data packet size, transmission speed, packet loss rate, etc., promptly discover and resolve transmission interruptions or delays, set up monitoring at the data storage layer to ensure that data is stored correctly to the specified location as expected, and monitor the health status, space usage and backup and recovery capabilities of the storage device.

[0056] S1012. During real-time monitoring, abnormal behavior data during data collection, transmission and storage is identified through preset rules, including but not limited to data tampering, loss and delay;

[0057] Specifically, the preset rules are as follows: using hash values ​​or digital signature technology to verify whether the data has been tampered with during transmission; setting thresholds for data transmission and storage; defining the normal time range for data processing, and data outside this range is considered delayed; using a real-time analysis engine to perform real-time analysis on the monitored data and compare it with the preset rules.

[0058] S1013, classifying and storing the abnormal behavior data to obtain abnormal data;

[0059] Specifically, the abnormal behavior data is classified according to the abnormal type, such as data tampering, loss, and delay, and the classified abnormal behavior data is organized into a structured format and stored as abnormal data.

[0060] S102. Use data comparison tools to compare the data to be verified obtained from different channels with known accurate data sets, identify differences and erroneous data, and integrate them into a preliminary erroneous data set;

[0061] In the specific implementation process, the specific steps of step S102 are:

[0062] S1021. Filter out the data set to be verified from multiple source data, and prepare a known accurate data set as a comparison benchmark;

[0063] Specifically, the multi-source data is cleaned, and verification targets are set to screen out the data sets to be verified, while one or more known accurate data sources are selected as comparison benchmarks.

[0064] S1022. Select a corresponding data comparison tool according to the data type and comparison requirements;

[0065] Specifically, understand the data types of the data set to be verified and the known accurate data set, and select the corresponding data comparison tool based on the data type and comparison requirements, specifically professional data quality management tools and the comparison function in ETL (Extract, Transform, Load) tools.

[0066] S1023. Using the selected data comparison tool, compare the data set to be verified with the known accurate data set, specifically including comparing the consistency, completeness and accuracy of the data, to obtain a comparison result;

[0067] Specifically, the comparison parameters are configured in the selected data comparison tool. After the comparison parameters are configured, the data comparison tool is started to compare the data set to be verified with the known accurate data set. The comparison process specifically compares the consistency, completeness and accuracy of the data to obtain the comparison result.

[0068] Exemplarily, the address field is configured as the comparison field in the data comparison tool, and the fuzzy matching rule and difference threshold are set. Then, the comparison tool is started, and the tool automatically loads two data sets and compares them according to the configured rules.

[0069] S1024. According to the comparison results, identify the differences and errors between the data and integrate them into a preliminary error data set;

[0070] Specifically, the comparison results are analyzed through existing data analysis tools to identify the differences and errors between the data, and the identified differences and errors are classified and sorted according to type, cause, etc. The classified difference and error data are integrated into a preliminary error data set.

[0071] S103, analyzing the correlation of the same type of data based on historical data, and establishing a data correlation estimation model;

[0072] In the specific implementation process, the specific steps of step S103 are:

[0073] S1031. Collect historical data of the same type as the data required for the current task test, and extract basic features from it through data mining technology;

[0074] Specifically, we use database queries, web crawlers, API calls and other methods to collect the required historical data from reliable data sources, such as meteorological stations, scientific research institutions, and government open data platforms, and extract basic features from the historical data through any one or several algorithms including statistical analysis, cluster analysis or principal component analysis (PCA).

[0075] S1032, based on the extracted features, obtaining the association relationship between the same type of data, using the association relationship as a training set, selecting a neural network model for training to obtain a data association prediction model;

[0076] Specifically, we use correlation analysis and causal analysis methods to explore the correlation between extracted features and between the same type of data prediction data targets, convert the analyzed correlation into a data set that can be used for model training, and obtain the data association prediction model through training of the long short-term memory network (LSTM).

[0077] S104, taking the identified abnormal data and the preliminary error data set as input, estimating the associated error data in each type of data through the established data association estimation model, and forming an associated error data set;

[0078] In the specific implementation process, the specific steps of step S104 are:

[0079] S1041, inputting the identified abnormal data and the preliminary error data set into the data association prediction model, and using the model to predict the associated error data in each type of data;

[0080] Specifically, the sorted abnormal data and preliminary error data sets are taken as input and imported into the data association prediction model through an appropriate interface or format (such as CSV files, database query results, etc.). After receiving the input data, the data association prediction model will process and analyze the input data according to its internal algorithm and training process results, identify errors or anomalies that may exist due to the correlation between the data, and after execution is completed, it will output the estimated associated error data in each type of data.

[0081] For example, an e-commerce company receives a batch of abnormal order data identified through preliminary verification rules and preliminary erroneous order data from user feedback, and inputs these data into a data association prediction model. The model analyzes the correlation between orders, such as user purchase history, rationality of product combinations, etc., and estimates more potential associated erroneous order data, such as incorrect price calculations, duplicate order entries, etc.

[0082] S1042, removing duplicate data and invalid data associated with erroneous data estimated in each type of data;

[0083] Specifically, for the estimated associated error data set, deduplication processing is first performed, and the validity of the deduplicated data is checked, including checking whether the format and range of the data meet the preset standards. After deduplication and validity verification, the cleaned associated error data set is output.

[0084] S1043, marking the formed associated error data set and integrating it into an associated error data set;

[0085] Specifically, identification information is added to each data record in the cleaned associated error data set, and the associated error data with the identification information is integrated to merge the error data from multiple sources into a unified data set to obtain an integrated associated error data set.

[0086] S105, review the identified abnormal data, the preliminary error data set, and the associated error data set integrated by the identification, and correct the confirmed error data according to the review result to obtain the determined task test data;

[0087] In the specific implementation process, the specific steps of step S105 are:

[0088] S1051. Preset audit rules and write custom scripts based on the audit rules, wherein the preset audit rules specifically include data consistency rules, data integrity rules, data accuracy rules, and business logic rules;

[0089] Specifically, in setting the audit rules, data consistency rules include ensuring that data remains consistent between different sources or different fields. For example, the order amount should be the same in the order details and payment records; data integrity rules include checking whether the data is complete and without missing necessary fields or information. For example, user information must contain valid contact information; data accuracy rules include verifying the correctness of the data. For example, whether the date format is correct, whether the value is within a reasonable range, etc.; business logic rules include checking whether the data conforms to specific business logic based on business scenarios and processes. For example, whether the change in order status conforms to the prescribed process; according to the established audit rules, use Python or Java in the appropriate programming language to write custom scripts to read the input data set and verify them one by one according to the preset audit rules.

[0090] S1052, importing the identified abnormal data, the preliminary error data set, and the associated error data set integrated by identification into the written custom script, and verifying them one by one through the custom script;

[0091] Specifically, the identified abnormal data, preliminary error data set, and the identified and integrated associated error data set are organized into a CSV file format that can be read by the script. The prepared data set is imported into the custom script through the input interface or function of the custom script. Then the custom script verifies the imported data set one by one according to the preset audit rules.

[0092] S1053, correcting the data that failed the verification to obtain confirmed task test data;

[0093] Specifically, the data that fails the verification is identified to determine the cause and nature of the error. Based on the cause and nature of the error, a correction method corresponding to the cause and nature of the error is automatically selected, including but not limited to modifying field values, deleting invalid records, adding missing information, etc. After the correction is completed, the custom script is used again to verify the corrected data to ensure that all errors have been corrected and the determined task test data is obtained.

[0094] S200, pairing an automated scanning tool with the task test data, and scanning and detecting the task data through the automated scanning tool to obtain vulnerability scanning results;

[0095] Specifically, according to the characteristics of the task test data, such as the IP range, open ports and service type of the target system, the system automatically selects the corresponding scanning tool from the tool library for configuration, scans and detects the task test data through the selected scanning tool, and records each operation and result of the scanning process in detail through the log.

[0096] S300, based on the obtained vulnerability scanning results, matching the vulnerability with a corresponding attack code to attack and obtain a penetration test result;

[0097] Specifically, select the attack code that matches the vulnerability from a preset library containing multiple attack codes (Exploit), and execute the attack code in a safe sandbox environment to simulate a real attack scenario to assess the actual impact and potential harm of the vulnerability. At the same time, record the successful penetration path, including the vulnerability, attack code and execution results used in each step of the attack, as well as access to sensitive data (user passwords, database information, etc.).

[0098] S400, evaluating the security risk level of the test system according to the obtained penetration test results, generating a test report and sending it to the customer, wherein the penetration test results specifically include vulnerability information, attack paths, and sensitive data;

[0099] Specifically, according to the test severity weight, attack difficulty weight, and impact range weight in the penetration test results, the comprehensive weight is obtained by multiplication and quantified into a specific security weight value. According to the preset security weight segments, the security risk level of the test system is divided accordingly to obtain the security risk level, and a report is generated, which includes the test purpose, test scope, test method and vulnerability information and transmitted to the customer.

[0100] Embodiment 2

[0101] like Figure 2 As shown, the artificial intelligence-based network security penetration test management system applied to the artificial intelligence-based network security penetration test management method specifically includes:

[0102] A data verification module is used to collect task test data for the test system, verify the task test data, and obtain confirmed task test data;

[0103] The scanning and detection module is used to pair the automated scanning tool with the task test data, and scan and detect the task data through the automated scanning tool to obtain the vulnerability scanning results;

[0104] Specifically, the scanning and detection module is used to automatically select the corresponding scanning tool from the tool library for configuration according to the characteristics of the task test data, specifically the IP range, open ports and service type of the target system, and scan and detect the task test data through the selected scanning tool, and record each step of the scanning process and the results in detail through the log.

[0105] The vulnerability attack module is used to match the corresponding attack code to the vulnerability based on the obtained vulnerability scanning results, and then attack to obtain the penetration test results;

[0106] Specifically, the vulnerability attack module selects attack codes that match the vulnerability from a preset library containing multiple attack codes (Exploit), and executes the attack code in a secure sandbox environment to simulate real attack scenarios to assess the actual impact and potential harm of the vulnerability. At the same time, it records successful penetration paths, including the vulnerabilities, attack codes and execution results used in each step of the attack, as well as access to sensitive data.

[0107] The report generation module is used to evaluate the security risk level of the test system based on the obtained penetration test results, generate a test report and send it to the customer, where the penetration test results specifically include vulnerability information, attack paths and sensitive data;

[0108] Specifically, the report generation module is used to obtain the comprehensive weight by multiplying the test severity weight, attack difficulty weight, and impact range weight in the penetration test results, and quantify it into a specific security weight value. According to the preset security weight segments, the security risk level of the test system is divided accordingly to obtain the security risk level and generate a report, which specifically includes the test purpose, test scope, test method, and vulnerability information, and transmits it to the customer.

[0109] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0110] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0111] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A network security penetration test management method based on artificial intelligence, characterized in that: The management method comprises: S100, collecting task test data for the test system, and verifying the task test data to obtain determined task test data; The specific steps for collecting task test data for the test system and verifying the task test data are as follows: Acquire data from different channels and use real-time monitoring technology to monitor data collection, transmission and storage in real time to identify abnormal data during the acquisition process; Use data comparison tools to compare the data to be verified from different channels with known accurate data sets, identify differences and erroneous data, and integrate them into a preliminary erroneous data set; Analyze the correlation of the same type of data based on historical data and establish a data correlation prediction model; Taking the identified abnormal data and the preliminary error data set as input, the associated error data in each type of data is estimated through the established data association estimation model to form an associated error data set; Review the identified abnormal data, preliminary error data sets, and associated error data sets that have been identified and integrated, and correct the confirmed error data based on the review results to obtain confirmed task test data; S200, pairing an automated scanning tool with the task test data, and scanning and detecting the task data through the automated scanning tool to obtain vulnerability scanning results; S300, based on the obtained vulnerability scanning results, matching the vulnerability with a corresponding attack code to attack and obtain a penetration test result; S400, evaluating the security risk level of the test system according to the obtained penetration test results, generating a test report and sending it to the customer, wherein the penetration test results specifically include vulnerability information, attack paths, and sensitive data; The specific steps of analyzing the correlation of the same type of data based on historical data and establishing a data correlation prediction model are as follows: Collect historical data of the same type as the data required for the current task test and extract basic features from it through data mining technology; Based on the extracted features, the association relationship between the same type of data is obtained, and the association relationship is used as a training set. A neural network model is selected for training to obtain a data association prediction model; The specific steps of taking the identified abnormal data and the preliminary error data set as input, estimating the associated error data in each type of data through the established data association estimation model, and forming the associated error data set are: Input the identified abnormal data and preliminary error data set into the data association prediction model, and use the model to estimate the associated error data in each type of data; Remove duplicate and invalid data that are estimated to be associated with erroneous data in each type of data; The formed associated error data sets are identified and integrated into associated error data sets; The specific steps of reviewing the identified abnormal data, the preliminary error data set, and the associated error data set integrated by the identification, and correcting the confirmed error data according to the review results to obtain the determined task test data are: Preset audit rules and write custom scripts based on the audit rules. The preset audit rules specifically include data consistency rules, data integrity rules, data accuracy rules, and business logic rules; Import the identified abnormal data, preliminary error data set, and associated error data set integrated with identification into the written custom script, and verify them one by one through the custom script; Correct the data that fails the verification to obtain the confirmed task test data.

2. The network security penetration test management method based on artificial intelligence according to claim 1 is characterized in that: The specific steps of acquiring data from different channels, using real-time monitoring technology to monitor the collection, transmission and storage of data in real time, and identifying abnormal data in the acquisition process are as follows: Collect the data required for the current task test from multiple channels, obtain multi-source data, and deploy real-time monitoring mechanisms in each link of multi-source data collection, transmission, and storage; During real-time monitoring, the preset rules are used to identify abnormal behavior data during data collection, transmission and storage, including but not limited to data tampering, loss and delay; The abnormal behavior data is classified, organized and stored to obtain abnormal data.

3. The network security penetration test management method based on artificial intelligence according to claim 1 is characterized in that: The specific steps of using the data comparison tool to compare the data to be verified from different channels with known accurate data sets, identify differences and erroneous data, and integrate them into a preliminary erroneous data set are as follows: Filter out the data set to be verified from multiple sources and prepare a known accurate data set as a comparison benchmark; Select the corresponding data comparison tool according to the data type and comparison requirements; Use the selected data comparison tool to compare the data set to be verified with the known accurate data set, including the consistency, completeness and accuracy of the comparison data, to obtain the comparison results; Based on the comparison results, the differences and errors between the data were identified and integrated into a preliminary error dataset.

4. An artificial intelligence-based network security penetration test management system, applied to the artificial intelligence-based network security penetration test management method according to any one of claims 1-3, characterized in that: The management system comprises: A data verification module is used to collect task test data for the test system, verify the task test data, and obtain confirmed task test data; The scanning and detection module is used to pair the automated scanning tool with the task test data, and scan and detect the task data through the automated scanning tool to obtain the vulnerability scanning results; The vulnerability attack module is used to match the corresponding attack code to the vulnerability based on the obtained vulnerability scanning results, and then attack to obtain the penetration test results; The report generation module is used to evaluate the security risk level of the test system based on the obtained penetration test results, generate a test report and send it to the customer, where the penetration test results specifically include vulnerability information, attack paths and sensitive data.

Images