Network security testing and analysis methods, devices, computer equipment and storage media

By analyzing historical cybersecurity test reports and establishing a set of attack surface correlation information, and using a risk assessment model to automatically identify potential attack surfaces, the problem of incomplete cybersecurity test coverage is solved, and the efficiency and comprehensiveness of testing are improved.

CN123122985APending Publication Date: 2026-11-17CHINA PING AN PROPERTY INSURANCE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611486822.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-22
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Current network security testing technologies struggle to cover all potential attack surfaces of the tested system, resulting in blind spots, low testing efficiency, and incompleteness.

Method used

By acquiring multiple historical cybersecurity test reports, analyzing the correlation between application types and business system information, establishing a set of correlation information between attack surfaces, and using a pre-trained risk assessment model to automatically identify potential attack surfaces not covered in the tests.

Benefits of technology

It enables the automatic discovery of potential cybersecurity risks without human intervention, improving the efficiency and comprehensiveness of cybersecurity testing and ensuring the integrity of the tested business systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122985A_ABST
    Figure CN123122985A_ABST
Patent Text Reader

Abstract

This application relates to the field of network security technology, and discloses a network security testing and analysis method, apparatus, computer equipment, and storage medium. It constructs a first set of associated information between each application type and each business system information using historical network security test reports; constructs a second set of associated information between each application type and each attack surface using a network security database; obtains the test business system information and its associated multiple test attack surfaces from the network security test report to be processed; matches the test business system information with the first and second sets of associated information to obtain multiple matching attack surfaces; identifies at least one candidate attack surface whose similarity to the test attack surface and the matching attack surface is lower than a first preset similarity; inputs the text of the candidate attack surface into a pre-trained risk assessment model, and outputs the target attack surface not involved in the test. This method can be applied to financial and medical scenarios, improving the comprehensiveness of security testing.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of cybersecurity technology and is applied in financial and medical scenarios. In particular, it relates to a cybersecurity testing and analysis method, apparatus, computer equipment, and storage medium. Background Technology

[0002] Cybersecurity testing is a crucial technical means of cybersecurity risk assessment. Cybersecurity test reports comprehensively record security vulnerabilities, risk levels, and remediation recommendations found in the tested system or application. On one hand, this guides technical operations personnel in locating security flaws and completing closed-loop remediation; on the other hand, it provides enterprise management with a clear view of the system's security status, offering a basis for decision-making regarding security resource allocation and security strategy formulation. Examples include cybersecurity testing of financial or medical systems.

[0003] In related technologies, testers can only conduct security testing within a limited scope, making it difficult to cover all potential attack surfaces of the system under test. This easily leads to omissions of boundary scenarios and failure to include new business risks in the testing scope. Furthermore, the business models of different systems vary greatly, and when testers lack experience, some critical attack surfaces may be missed in the cybersecurity testing scope, resulting in blind spots in the cybersecurity test report. In other words, related technologies suffer from low testing efficiency and incomplete testing in cybersecurity testing. Summary of the Invention

[0004] This application provides a network security testing and analysis method, apparatus, computer equipment, and storage medium, which can improve the testing efficiency and comprehensiveness of network security testing.

[0005] Firstly, a network security testing and analysis method is provided, including: Multiple historical network security test reports are obtained, and the multiple historical network security test reports are analyzed to obtain the first set of related information between each application type and each business system information; Based on the network security database, establish a second set of associated information between each application type and the attack surface corresponding to each application type; Obtain the test business system information of the network security test report to be processed, as well as the multiple test attack surfaces associated with the test business system information; The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system, and the target application type is matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. Determine the similarity between each test attack surface and each matched attack surface, and among the multiple matched attack surfaces, identify at least one candidate attack surface whose similarity is lower than a first preset similarity. The target application type, the test business system information, and the description text corresponding to the at least one candidate attack surface are input into the pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report to be processed.

[0006] Secondly, it provides network security testing and analysis equipment, including: The first acquisition module is used to acquire multiple historical network security test reports and analyze the multiple historical network security test reports to obtain a first set of related information between each application type and each business system information. The generation module is used to establish a second set of associated information between each application type and the attack surface corresponding to each application type, based on the network security database. The second acquisition module is used to acquire the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information; The matching module is used to match the test business system information with the first set of associated information to obtain the target application type corresponding to the test business system, and to match the target application type with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. The determination module is used to determine the similarity between each of the test attack surfaces and each of the matched attack surfaces, and to determine at least one candidate attack surface whose similarity is lower than a first preset similarity among the multiple matched attack surfaces; The prediction module is used to input the target application type, the test business system information, and the description text corresponding to the at least one candidate attack surface into a pre-trained risk assessment model, and output the target attack surface that is not covered in the network security test report to be processed.

[0007] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the aforementioned network security testing and analysis method.

[0008] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the aforementioned network security testing and analysis method.

[0009] In the network security testing and analysis method, apparatus, computer equipment, and storage medium provided in this application embodiment, the following steps are taken: First, a first set of associated information is obtained by acquiring multiple historical network security test reports and analyzing them; second, a second set of associated information is established based on the network security database, corresponding to each application type and its corresponding attack surface; third, the test business system information of the network security test report to be processed, and multiple test attack surfaces associated with the test business system information, are acquired; fourth, the test business system information and the first set of associated information are matched to obtain the target application type corresponding to the test business system, and the target application type and the second set of associated information are matched to determine multiple matching attack surfaces corresponding to the test business system; fifth, the similarity between each test attack surface and each matching attack surface is determined, and at least one candidate attack surface with a similarity lower than a first preset similarity is identified among the multiple matching attack surfaces; sixth, the description text corresponding to the target application type, test business system information, and at least one candidate attack surface is input into a pre-trained risk assessment model, and the target attack surface not involved in the test in the network security test report to be processed is output.

[0010] Therefore, by acquiring and analyzing multiple historical network security test reports, a first set of correlation information is obtained between each application type and each business system information. This allows for the discovery of the correlation between application types and business system information. Based on the network security database, a second set of correlation information is established between each application type and its corresponding attack surface. This allows for the discovery of the correlation between application types and attack surfaces.

[0011] Then, the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information, are obtained. The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system. The target application type is then matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. This allows for the accurate identification of multiple matching attack surfaces that may be vulnerable to attacks corresponding to the test business system.

[0012] Finally, the similarity between each test attack surface and each matched attack surface is determined. Among multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified. The target application type, test business system information, and the descriptive text corresponding to each candidate attack surface are input into a pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report. This process identifies at least one candidate attack surface that has not been tested. Finally, the pre-trained risk assessment model identifies potential target attack surfaces that may pose a network attack risk, thus promptly identifying potential attack risks that have not been tested.

[0013] Compared to related technologies where security testing is conducted by testers within a limited scope, this application eliminates the need for manual intervention. It can automatically identify target attack surfaces that have not undergone cybersecurity testing and may pose cybersecurity risks. These target attack surfaces can be used to retest the tested business systems or to detect potential cyberattack risks in advance, thereby improving the efficiency and comprehensiveness of cybersecurity testing. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the application environment for the network security testing and analysis method provided in the embodiments of this application; Figure 2 This is a schematic diagram of a scenario for the network security testing and analysis method provided in the embodiments of this application; Figure 3 This is a flowchart illustrating the network security testing and analysis method provided in the embodiments of this application; Figure 4 This is another flowchart illustrating the network security testing and analysis method provided in this application embodiment; Figure 5 This is a schematic diagram of the network security testing and analysis device provided in the embodiments of this application; Figure 6 This is a schematic diagram of the structure of the computer device provided in the embodiments of this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] It should be noted that in all specific embodiments of this application, when data related to user identity or characteristics needs to be processed, the user's permission or consent will be obtained first. Furthermore, the collection, use, and processing of this data will comply with relevant laws, regulations, and standards. In addition, when embodiments of this application need to obtain sensitive personal information of users, separate permission or consent from the user will be obtained through pop-ups or redirects to confirmation pages. Only after obtaining the user's separate permission or consent will the necessary user-related data for the normal operation of the embodiments of this application be obtained.

[0018] In some processes described in the specification, claims and the foregoing drawings, there are multiple steps that appear in a specific order. However, it should be clearly understood that these steps may be performed in any order or in parallel. The step numbers are only used to distinguish the different steps and do not represent any execution order.

[0019] The technical problems existing in the relevant technology are as follows: Cybersecurity testing is a crucial technical means of cybersecurity risk assessment. Cybersecurity test reports comprehensively record security vulnerabilities, risk levels, and remediation recommendations found in the tested system or application. On one hand, this guides technical operations personnel in locating security flaws and completing closed-loop remediation; on the other hand, it provides enterprise management with a clear view of the system's security status, offering a basis for decision-making regarding security resource allocation and security strategy formulation. Examples include cybersecurity testing of financial or medical systems.

[0020] In related technologies, testers can only conduct security testing within a limited scope, making it difficult to cover all potential attack surfaces of the system under test. This easily leads to omissions of boundary scenarios and failure to include new business risks in the testing scope. Furthermore, the business models of different systems vary greatly, and when testers lack experience, some critical attack surfaces may be missed in the cybersecurity testing scope, resulting in blind spots in the cybersecurity test report. In other words, related technologies suffer from low testing efficiency and incomplete testing in cybersecurity testing.

[0021] To address the aforementioned problems, this application embodiment obtains and analyzes multiple historical network security test reports to generate a first set of correlation information between each application type and each business system information. This allows for the discovery of the correlation between application types and business system information. Based on a network security database, a second set of correlation information is established between each application type and its corresponding attack surface. This allows for the discovery of the correlation between application types and attack surfaces.

[0022] Then, the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information, are obtained. The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system. The target application type is then matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. This allows for the accurate identification of multiple matching attack surfaces that may be vulnerable to attacks corresponding to the test business system.

[0023] Finally, the similarity between each test attack surface and each matched attack surface is determined. Among multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified. The target application type, test business system information, and the descriptive text corresponding to each candidate attack surface are input into a pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report. This process identifies at least one candidate attack surface that has not been tested. Finally, the pre-trained risk assessment model identifies potential target attack surfaces that may pose a network attack risk, thus promptly identifying potential attack risks that have not been tested.

[0024] Compared to related technologies where security testing is conducted by testers within a limited scope, this application eliminates the need for manual intervention. It can automatically identify target attack surfaces that have not undergone cybersecurity testing and may pose cybersecurity risks. These target attack surfaces can be used to retest the tested business systems or to detect potential cyberattack risks in advance, thereby improving the efficiency and comprehensiveness of cybersecurity testing.

[0025] Please see Figure 1 , Figure 1 This is a schematic diagram of the application environment for the network security testing and analysis method provided in the embodiments of this application.

[0026] The network security testing and analysis method provided in this application embodiment can be applied to, for example, Figure 1In this application environment, the client communicates with the server via a network. The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The following detailed description of specific embodiments further illustrates this application.

[0027] The server or client can obtain multiple historical network security test reports and analyze them to obtain a first set of correlation information between each application type and each business system information. This allows for the discovery of the relationships between application types and business system information. Based on the network security database, a second set of correlation information is established between each application type and its corresponding attack surface. This allows for the discovery of the relationships between application types and attack surfaces.

[0028] Then, the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information, are obtained. The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system. The target application type is then matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. This allows for the accurate identification of multiple matching attack surfaces that may be vulnerable to attacks corresponding to the test business system.

[0029] Finally, the similarity between each test attack surface and each matched attack surface is determined. Among multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified. The target application type, test business system information, and the descriptive text corresponding to each candidate attack surface are input into a pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report. This process identifies at least one candidate attack surface that has not been tested. Finally, the pre-trained risk assessment model identifies potential target attack surfaces that may pose a network attack risk, thus promptly identifying potential attack risks that have not been tested.

[0030] As can be seen from the above, this application can automatically identify target attack surfaces that have not undergone network security testing and may pose network security risks without human intervention. These target attack surfaces can be used to retest the tested business system or to detect potential network attack risks in advance, thereby improving the testing efficiency and comprehensiveness of network security testing.

[0031] Please see Figure 2 , Figure 2 This is a schematic diagram of a scenario for the network security testing and analysis method provided in the embodiments of this application.

[0032] This includes obtaining historical network security test reports, which record all test content and results for the relevant business systems. Each historical network security test report can be analyzed to determine the application type and corresponding business system information for each report. Finally, by combining the application types and corresponding business system information from all historical network security test reports, a first set of related information between each application type and each business system can be obtained.

[0033] The preset application types include several basic application types such as "internal applications," "user-end applications," "AI applications," "content generation applications," "mobile applications," "financial payment applications," and "new technology applications." The application types in historical cybersecurity test reports can include one or more of these basic application types. In other words, application types are abstract classification labels used to summarize the business characteristics of business systems.

[0034] Business system information can be a textual description of the business system under test, such as system name, business function, service object, deployment method, business scenario, etc.

[0035] The first set of related information is essentially a standard database that can characterize the relationships between application types and business system information.

[0036] Based on a cybersecurity database, a second set of correlated information can be established between each application type and its corresponding attack surface. For example, the cybersecurity database contains standardized security testing items (attack surface entries) that different application types should cover, along with corresponding standard terms and standardized descriptive text. By analyzing the cybersecurity database, the relationship between application types and attack surfaces can be obtained, thus constructing a second set of correlated information between each application type and each attack surface.

[0037] The attack surface is a set of potential entry points, attack paths, and risk detection items that attackers can use to launch attacks within an application / business system. It represents various risk dimensions that the system may be vulnerable to attack. In this application, it refers to the security detection items that need to be performed on the business system, characterizing the risk dimensions that attackers can use to launch attacks on the system.

[0038] From the pending network security test report of the tested business system, the test business system information and test attack surface can be extracted. The test attack surface can be understood as the attack surface that has already undergone network security testing.

[0039] Then, the information level of the test business system can be vectorized and matched with the vectors of business system information in the first association set. The application type corresponding to the vector of the matched business system information is then used as the target application type.

[0040] Then, the vector corresponding to the target application type is matched with the vector corresponding to the application type in the second associated information set, thereby determining the attack surface associated with the vector of the matched application type as the matched attack surface.

[0041] After obtaining the matching attack surfaces, the similarity between each test attack surface and each matching attack surface is determined, and at least one candidate attack surface with a similarity lower than the first preset similarity is identified among the multiple matching attack surfaces.

[0042] Finally, the target application type, test business system information, and description text corresponding to each candidate attack surface are input into the pre-trained risk assessment model, which outputs the target attack surfaces not covered in the network security test report to be processed.

[0043] In other words, among multiple matching attack surfaces, those with a similarity greater than or equal to a first preset similarity level are identical or similar to the test attack surface and do not require repeated testing. At least one candidate attack surface with a similarity lower than the first preset similarity level has not undergone cybersecurity testing. Finally, a pre-trained risk assessment model is used to evaluate the candidate attack surfaces, filtering out target attack surfaces that can affect the test business system. Attack surfaces that have not been cybersecurity tested can be considered potential network threats, and subsequent cybersecurity testing of the test business system can be conducted based on these target attack surfaces. This approach automatically and efficiently analyzes potential threats, improving the efficiency and comprehensiveness of cybersecurity testing.

[0044] The network security testing and analysis method provided in this application can be applied to financial networks to ensure financial security, and can also be applied to medical networks to ensure the security of medical information.

[0045] The network security testing and analysis methods, apparatus, computer equipment, and storage media provided in the embodiments of this application will be described in detail below.

[0046] Please see Figure 3 , Figure 3 This is a flowchart illustrating the network security testing and analysis method provided in this application. The network security testing and analysis method may include the following steps: Step 210: Obtain multiple historical network security test reports and analyze them to obtain the first set of related information between each application type and each business system information. Step 220: Based on the network security database, establish a second set of related information between each application type and the corresponding attack surface; Step 230: Obtain the test business system information of the network security test report to be processed, as well as the multiple test attack surfaces associated with the test business system information; Step 240: Match the test business system information with the first set of related information to obtain the target application type corresponding to the test business system, and match the target application type with the second set of related information to determine multiple matching attack surfaces corresponding to the test business system; Step 250: Determine the similarity between each test attack surface and each matched attack surface, and among multiple matched attack surfaces, identify at least one candidate attack surface whose similarity is lower than the first preset similarity. Step 260: Input the target application type, test business system information, and description text corresponding to each candidate attack surface into the pre-trained risk assessment model, and output the target attack surfaces not covered in the network security test report to be processed.

[0047] Steps 210 to 260 will be described in detail below.

[0048] In step 210, multiple historical network security test reports are obtained and analyzed to obtain the first set of related information between each application type and each business system information.

[0049] This includes obtaining historical network security test reports, which record all test content and results for the relevant business systems. Each historical network security test report can be analyzed to determine the application type and corresponding business system information for each report. Finally, by combining the application types and corresponding business system information from all historical network security test reports, a first set of related information between each application type and each business system can be obtained.

[0050] The preset application types include several basic application types such as "internal applications," "user-end applications," "AI applications," "content generation applications," "mobile applications," "financial payment applications," and "new technology applications." The application types in historical cybersecurity test reports can include one or more of these basic application types. In other words, application types are abstract classification labels used to summarize the business characteristics of business systems.

[0051] Business system information can be a textual description of the business system under test, such as system name, business function, service object, deployment method, business scenario, etc.

[0052] The first set of related information is essentially a standard database that can characterize the relationships between application types and business system information.

[0053] In some implementations, multiple historical network security test reports are analyzed to obtain a first set of correlation information corresponding to each application type and each business system information, including: (1.1) Perform text cleaning on the text corresponding to each historical network security test report to obtain the target report text corresponding to each historical network security test report; (1.2) Determine the application type corresponding to each business system and the business system information associated with the application type based on each target report text; (1.3) Vectorize each application type and associated business system information to obtain the first set of associated information between each application type and associated business system information.

[0054] This process involves obtaining multiple original files of historical cybersecurity test reports. These reports can be in document format and contain textual content such as system overview, description of the tested business system, and test scope. The report text is then extracted from these historical reports, focusing on paragraphs describing the tested business system. Redundant content, such as vulnerability details, vulnerability reproduction steps, and remediation suggestions, which are not system attribute descriptions, is filtered out to obtain intermediate processed text.

[0055] It can also perform text cleaning operations on the intermediate processed text, such as removing special characters, tabs, line breaks, garbled characters, and document tags; removing function words, modifiers, and templated fixed phrases without actual business meaning; and eliminating repetitive and redundant text fragments. This yields the target report text, which mainly retains the effective text content that can characterize the attributes of the business system.

[0056] Then, word segmentation is performed on the target report text, dividing the continuous text into independent word units. The word frequency of each word is counted, and target text with a word frequency higher than a preset word frequency threshold is selected (high-frequency feature words, such as "OA", "internal approval", "user payment", "large model inference", etc.); high-frequency words can reflect the business attributes of the business system.

[0057] Based on high-frequency feature words, the system matches preset basic application types (internal applications, C-end applications, AI applications, content generation applications, mobile applications (APPs), financial payment applications, and new technology applications) to determine the application type to which the business system in the historical cybersecurity test report belongs; it supports one business system corresponding to multiple application types (cross-types).

[0058] The text fragments in the target report that describe the system's business attributes are treated as business system information and bound to the identified application types to form a set of "application types". The process of mapping application types to business system information is repeated across all historical network security test reports to obtain multiple sets of mappings between application types and business system information.

[0059] Finally, for each group of bound "application type - business system information", the business system information text is segmented and pooled; through the word vectorization model, the text of the business system information is converted into a high-dimensional numerical vector; this vector is the vector representation of the business system information, and business system information with similar semantics has vectors that are close in spatial distance.

[0060] The business system information vectors and their corresponding application type tags are associated and stored. All processed historical reports containing <business system information vectors, application types> data are aggregated to form a structured dataset, namely the first associated information set. Each record within the first associated information set has a key of a business system information vector and a value of one or more corresponding application type tags. This set is stored offline for matching and retrieval in subsequent reports.

[0061] As explained above, the advantages of this approach are twofold: firstly, it constructs a first set of related information through offline analysis of existing historical cybersecurity test reports; secondly, it eliminates report noise interference through text cleaning, utilizes high-frequency words to mine the correspondence between business system information and application types, and then vectorizes this data to obtain a computable set of associations. On the one hand, it fully reuses existing data from historical penetration test reports, reducing the workload of manual annotation; on the other hand, it transforms the business system description text into vector representations, providing a data foundation for automatically identifying target application types in subsequent reports, achieving automated application type determination, and improving the automation level of the entire security test analysis process.

[0062] In some implementations, the application type corresponding to each business system and the associated business system information are determined based on each target report text, including: (1.2.1) Identify target texts in each target report text whose word frequency is higher than the preset word frequency; (1.2.2) Determine the application type corresponding to the business system and the business system information associated with the application type based on the target text.

[0063] The process involves acquiring the target report text after processing a single historical report. This text has been cleaned of special symbols, invalid function words, and template phrases, primarily retaining the business descriptions of the tested business system. Chinese word segmentation is then performed on the target report text, dividing the continuous natural language text into independent word units and removing stop words (particles, conjunctions, adverbs, etc., that lack business differentiation capabilities). The frequency of each word after segmentation is calculated, yielding a word frequency value for each word. A pre-configured preset word frequency threshold is read, and words with frequency values ​​exceeding this threshold are selected to form the target text (a set of high-frequency feature words).

[0064] For example, in a target report text from an internal OA system, after word segmentation and statistical analysis, the words "OA", "internal approval", and "employee management" have a higher frequency than the preset word frequency, and these words are the target text; while general words such as "test" and "report" are filtered out because their frequency does not meet the criteria.

[0065] By repeating this process, the target report text corresponding to each historical network security test report is traversed in a loop, and the target text (a set of high-frequency feature words) corresponding to each report is obtained one by one.

[0066] This application sets up multiple basic application types, which can match and compare high-frequency feature words in the target text with feature reference words of each application type. If the high-frequency feature words have a high degree of overlap with the reference words of a certain application type, then the application type is determined to be the application type corresponding to the business system. It supports matching one, two or more application types to adapt to composite business systems (such as mini programs that have both AI application and mobile application APP attributes).

[0067] Then, extract complete text fragments from the target report text that describe the business scenarios, service objects, and business functions of the system under test, as business system information; this business system information is used to characterize the overall business picture of the business system, and is not limited to the high-frequency words themselves.

[0068] Then, the identified application types are bound and associated with the information of the business system to form a set of application types. Business system information correspondence. If the historical report contains manually labeled application type tags, the automatic identification results will be compared and corrected with the manually labeled tags; obviously erroneous samples will be manually corrected to ensure sample quality. By traversing all historical reports, multiple sets of application types are obtained. Correspondence between business system information.

[0069] The advantage of this approach is that by segmenting and counting the frequency of words in the cleaned target report text, target texts with a frequency higher than a preset frequency are selected. Based on the target text matching, the application type corresponding to the business system is determined, and the corresponding business system information is bound. This method fully leverages the business characteristics contained in historical cybersecurity test reports, reducing the workload of manual annotation. Simultaneously, it supports multiple overlapping application types for a business system, taking into account complete business system information text, and providing high-quality raw sample data for subsequent vectorization and automatic identification of application types in reports to be processed.

[0070] In step 220, a second set of associated information is established based on the network security database, corresponding to each application type and the attack surface of each application type.

[0071] In some implementations, a second set of correlation information is established based on a network security database, corresponding to each application type and its corresponding attack surface, including: (1.1) Identify at least one attack surface for each application type and the standard terms for each attack surface based on the network security database; (1.2) Input the standard words corresponding to each application type and each attack surface into the pre-trained large language model, and output the corresponding key-value pairs between each application type and each attack surface; (1.3) Establish a second set of associated information between each application type and the attack surface corresponding to each application type based on key-value pairs.

[0072] This includes reading a network security database containing preset basic application types. For each basic application type, attack surface items that should be subject to security testing are extracted from the database. For example, for internal applications, attack surfaces include: unauthorized access detection, session timeout failure detection, and internal interface authentication detection; for AI applications, attack surfaces include: prompt word injection, model data poisoning, and training data leakage detection.

[0073] For each attack surface, standard terms are extracted. These standard terms are standardized words / phrases that can uniquely represent the attack surface, serving as a standardized expression of the attack surface. For example, the standard terms for the attack surface "session timeout without exit detection" are: session timeout, session not exited.

[0074] After completing a full traversal of all application types, the following raw data is obtained: Application Type - Attack Surface List - Attack Surface Corresponding Standard Terms. The attack surfaces and standard terms extracted from the database are manually reviewed to remove attack surface entries that do not match the application type, are duplicates, or are outdated. Ambiguous standard terms are corrected to ensure the accuracy of the baseline data.

[0075] Then, the processed application type, attack surface, and attack surface standard words are fed into a pre-trained large language model as input. The large language model normalizes and processes the input content, removes redundant descriptions and colloquial expressions, and outputs structured key-value pairs. The key is the application type, and the value is a set of standardized description texts of all attack surfaces under that application type.

[0076] After obtaining the output of the large model, text cleaning is performed on the key-value pairs, such as removing meaningless words, special symbols, redundant line breaks, and template phrases; non-existent attack surface entries generated by the model illusion are also removed. Manual secondary verification can also be performed to check the key-value pairs output by the large model, confirming the correct correspondence between application types and attack surfaces, and correcting any errors generated by the model.

[0077] Finally, for each standardized attack surface description text in the cleaned and verified application type-attack surface key-value pairs, the attack surface text is segmented and pooled; word vectorization is then performed, converting the natural language attack surface text into high-dimensional numerical vectors to obtain standard attack surface word vectors. The application type, standardized attack surface description text, and standard attack surface word vectors are then mutually bound and stored. All bound data corresponding to all application types are collected to form a structured dataset, i.e., the second set of associated information.

[0078] The advantage of this approach is that it obtains the attack surface and standard terms corresponding to each application type from a cybersecurity database, generates key-value pairs between application types and attack surfaces using a pre-trained large language model, and further constructs a second set of associated information. This solution relies on a security compliance knowledge base to ensure the professionalism and completeness of the attack surface, utilizes a large model to complete the structured transformation, and reduces manual processing costs. Simultaneously, it binds and stores the standardized attack surface text with the vectorized results, which not only supports rapid vector similarity comparison but also provides raw text input for subsequent risk assessment models, providing a reliable benchmark dataset for automatically verifying the completeness of attack surface coverage in penetration testing.

[0079] In step 230, the test business system information of the network security test report to be processed, as well as the multiple test attack surfaces associated with the test business system information, are obtained.

[0080] This system can receive pending network security test reports, which include a complete overview of the system, business description, test scope, test items, vulnerability detection process, and test conclusions. It employs text extraction to extract system attribute text from the system introduction, business overview, and tested system description sections of the report. Specifically, this includes: the tested system's business functions and target users; system deployment scenarios and user groups; system business characteristics and system positioning. All extracted descriptive text is then summarized into the test business system information.

[0081] From the test content, test items, test scope, and security detection project sections of the report, extract all security detection dimensions actually performed in this penetration test: security test items clearly carried out in the report; attack dimensions and detection types actually detected; and summarize all the security detection items actually tested in this test to form multiple test attack surface sets.

[0082] Then, the two types of information to be extracted are cleaned, such as removing special symbols, garbled characters, line breaks, and redundant template statements; deleting meaningless function words and repeated descriptions; retaining business feature words and test item keywords; finally, a clean and noise-free test business system information text and a structured collection of multiple test attack surface texts are obtained.

[0083] Finally, vectorization processing is performed. For example, test business system information is used to generate a test system information vector; each test attack surface generates a corresponding test attack surface vector, which is used for subsequent cosine similarity matching and attack surface comparison processes.

[0084] The advantage of doing this is that it accurately separates system business attribute information from actual test attack surface information from the pending report, enabling structured parsing of the report and providing effective input for subsequent automatic identification of application types and automatic comparison of attack surface integrity.

[0085] In step 240, the test business system information and the first set of associated information are matched to obtain the target application type corresponding to the test business system. The target application type and the second set of associated information are then matched to determine multiple matching attack surfaces corresponding to the test business system.

[0086] In some implementations, the test business system information and the first set of associated information are matched to obtain the target application type corresponding to the test business system, including: (1.1) Obtain the first vector corresponding to the information of the test business system; (1.2) Match the first vector with the vector corresponding to each business system information in the first associated information set to obtain the matching target business system information; (1.3) The application type associated with the target business system information is determined as the target application type corresponding to the test business system.

[0087] For example, a word vector mapping model can be used to convert the structured text corresponding to the information of the test business system into a fixed-dimensional high-dimensional semantic vector, thereby generating the first vector of the test business system.

[0088] Then, iterate through the standard vectors corresponding to all historical business system information in the first set of associated information, and calculate the cosine similarity between each standard vector and the first vector of the current system under test to obtain the similarity score for each set of vectors. Sort all similarity scores in descending order, and select the top three historical business system information with the highest similarity scores as candidate matching samples.

[0089] Then, a preset similarity threshold is set, and the credibility of the top three candidate samples is judged. If the highest similarity score among the top three is greater than the preset threshold, the match is considered valid, and the historical business system information corresponding to the high similarity among the top three is determined as the target business system information. If the highest similarity score is still lower than the preset threshold, the credibility of automatic matching is determined to be insufficient, and the application type is manually determined.

[0090] Finally, the application type tags bound to the successfully matched target business system information in the first set of associated information are retrieved. Based on the earlier matching results, cross-matching of single, two, or three application types is supported, meaning that the current test business system can correspond to multiple target application types simultaneously, adapting to complex business systems. All valid matching application type tags are summarized and uniformly organized into the final target application type of the system under test, for use in subsequent retrieval of the standard attack surface in the second set of associated information.

[0091] The advantage of doing this is that by converting business system text into standardized vectors and relying on cosine similarity to achieve semantic-level intelligent matching, it can replace manual subjective classification, improve the automation level and consistency of application type recognition, and ensure recognition stability and accuracy.

[0092] Next, the second set of associated information, which has been stored offline in a persistent storage, is read. This set contains the binding relationships between application type, standardized attack surface description text, and standard attack surface term vectors. Using the target application type as the search keyword, the second set of associated information is traversed to filter out all attack surface entries that completely correspond to the target application type. If multiple target application types exist, the attack surfaces corresponding to each type are retrieved separately, the search results are merged, and duplicate attack surface entries are removed.

[0093] Finally, the standardized description text and standard word vectors of all attack surfaces after retrieval and merging are extracted and summarized to form multiple matching attack surface sets. These sets represent all standard attack surfaces that the current tested business system can achieve network security detection.

[0094] In step 250, the similarity between each test attack surface and each matched attack surface is determined, and among the multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified.

[0095] In some implementations, among multiple matching attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified, including: (1.1) Among multiple matching attack surfaces, identify multiple initial attack surfaces with a similarity lower than the first preset similarity and a similarity greater than the second preset similarity; (1.2) Select at least one candidate attack surface from multiple initial attack surfaces that is suitable for testing business system information.

[0096] This process involves obtaining the word vectors of the standard attack surfaces corresponding to all matching attack surfaces, as well as the word vectors of the test attack surfaces corresponding to the test attack surfaces in the test report. For each standard matching attack surface, a cosine similarity is calculated between it and all test attack surfaces to obtain the maximum matching similarity of that standard attack surface, which is used to characterize the coverage of that standard attack surface in this test report.

[0097] Then, the standard attack surface with the maximum matching similarity less than the first preset similarity and greater than the second preset similarity is selected as the initial attack surface, where the second preset similarity is less than the first preset similarity. In other words, this interval represents vectors that are not completely matched, but not completely unrelated. These are fuzzy samples that the machine cannot directly determine whether they have been missed, and require further refined reasoning.

[0098] Attack surfaces with a similarity higher than the first preset similarity are directly determined to be covered; attack surfaces with a similarity lower than the second preset similarity are directly determined to be irrelevant to the test business system and will no longer be included in the model inference.

[0099] Finally, for all the initial attack surfaces selected, a business compatibility verification is performed by combining the information of the current test business system with the target application type. It is determined whether the security detection scenario and applicable objects of each initial attack surface match the business attributes, system architecture, and business functions of the tested business system, eliminating attack surfaces that are not applicable to the current system (e.g., attack surfaces that are only applicable to financial systems are not applicable to ordinary portal websites). Attack surface entries that are compatible with the current test business system scenario and are indeed required for detection by the current system are retained, ultimately resulting in at least one candidate attack surface, which serves as the final inference sample for the risk assessment model.

[0100] The advantage of this approach is that by using dual threshold intervals for screening, attack surfaces with high matching scores that have already been covered and attack surfaces with extremely low matching scores that have obviously been missed are filtered out in advance. Only ambiguous and difficult-to-determine intermediate samples are retained as the initial attack surfaces, which significantly reduces the computational load of model inference. At the same time, business adaptation screening is performed in conjunction with test business system information to eliminate incompatible and invalid attack surfaces, preventing irrelevant attack surfaces from entering the model and causing misjudgments. This effectively improves the accuracy of candidate attack surfaces and the accuracy of subsequent risk inference, balancing overall detection efficiency and detection accuracy.

[0101] In step 260, the target application type, test business system information, and description text corresponding to each candidate attack surface are input into the pre-trained risk assessment model, and the target attack surface not covered in the network security test report to be processed is output.

[0102] After selecting candidate attack surfaces, the standardized description text for each candidate attack surface is obtained. Simultaneously, the target application type and test business system information obtained during the detection are retrieved to form the multi-dimensional input data for the model. The target application type undergoes label normalization, the test business system information undergoes key business feature extraction and simplification, and the candidate attack surface description text retains complete and standardized semantic content while removing redundant noise to ensure that the input data is well-organized, consistent, and adaptable to the model's input interface.

[0103] The standardized target application type, test business system information, and candidate attack surface description text are then input into the pre-trained risk assessment model. The target application type provides the system security testing specification scenario, the test business system information provides the specific business attribute context, and the candidate attack surface text provides potential missed detection items to be assessed.

[0104] The pre-trained risk assessment model, based on security testing rules learned during training, risk characteristics of different application types, attack surface adaptation logic, and distribution of missed detection risks, performs comprehensive semantic reasoning and risk assessment on the input content, and calculates the probability of compliance missed detection risk for each candidate attack surface in the current tested system scenario.

[0105] When the output risk probability of a candidate attack surface exceeds a preset risk threshold, it is determined that the attack surface is not effectively covered in this test report and has a compliance risk due to test omissions. Finally, all attack surfaces determined to have test omission risks are summarized and uniformly output as target attack surfaces. These target attack surfaces are the set of attack surfaces that are not effectively tested and have compliance coverage deficiencies in the pending network security test reports, thus completing the integrity analysis of this network security test.

[0106] The pre-trained risk assessment model is trained as follows: First, a large number of historical cybersecurity test report samples were collected to construct a training dataset. Each sample included application type, business system information, candidate attack surface description text, and manually labeled whether a test was missed. All sample texts were cleaned, normalized, and vector-encoded, and then divided into training and validation sets. A risk assessment model integrating semantic features and scenario rules was built to predict the probability of missed attack surface risks using input features. Model training was completed by iteratively updating network parameters through backpropagation, and the validation set was used to continuously optimize the model's generalization ability, suppress false positives, and improve the accuracy of boundary sample resolution. After training convergence, the model parameters were solidified, resulting in a pre-trained risk assessment model that can be directly used for online inference.

[0107] In some implementations, after inputting the target application type, test business system information, and descriptive text corresponding to each candidate attack surface into a pre-trained risk assessment model, and outputting the target attack surfaces not covered in the network security test report to be processed, the method further includes: (1.1) In the preset test environment, determine the attack tactics corresponding to the test business system information based on the target attack surface; (1.2) Conduct network security tests on the test business system according to the attack tactics, and obtain the corresponding updated network security test report for the test business system.

[0108] For example, in a pre-defined simulation testing environment, the target attack surface output by the model is obtained, which represents the undetected items not covered in the report and posing compliance risks. Combining the security detection dimensions, attack principles, and vulnerability triggering conditions of the target attack surface with the business architecture, functional characteristics, and deployment environment of the current test system, an attack tactic adapted to that system is automatically generated. This attack tactic includes the corresponding testing approach, attack path, detection method, and testing steps, used to specifically complete the security detection of the undetected attack surface.

[0109] Finally, based on the generated attack tactics, supplementary cybersecurity penetration testing was conducted on the test business system, completing security checks corresponding to each target attack surface. The testing process, results, and risks of this new test were recorded, and the new test content was merged into the original pending cybersecurity test report. The report content was then organized and summarized to generate an updated cybersecurity test report with complete supplementary test items, achieving a closed-loop completion of any missed test items in the original report.

[0110] The advantage of doing this is that it allows for the retesting of the business system by targeting untested attack surfaces, thereby achieving comprehensiveness in network security testing.

[0111] In the network security testing and analysis method provided in this application embodiment, multiple historical network security test reports are acquired and analyzed to obtain a first set of associated information between each application type and each business system information. This allows for the discovery of the correlation between application types and business system information. Based on the network security database, a second set of associated information between each application type and its corresponding attack surface is established. This allows for the discovery of the correlation between application types and attack surfaces.

[0112] Then, the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information, are obtained. The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system. The target application type is then matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. This allows for the accurate identification of multiple matching attack surfaces that may be vulnerable to attacks corresponding to the test business system.

[0113] Finally, the similarity between each test attack surface and each matched attack surface is determined. Among multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified. The target application type, test business system information, and descriptive text corresponding to at least one candidate attack surface are input into a pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report. This process identifies at least one candidate attack surface that has not been tested. Finally, the pre-trained risk assessment model identifies potential target attack surfaces that may pose a network attack risk, thus promptly identifying potential attack risks that have not been tested.

[0114] Compared to related technologies where security testing is conducted by testers within a limited scope, this application eliminates the need for manual intervention. It can automatically identify target attack surfaces that have not undergone cybersecurity testing and may pose cybersecurity risks. These target attack surfaces can be used to retest the tested business systems or to detect potential cyberattack risks in advance, thereby improving the efficiency and comprehensiveness of cybersecurity testing.

[0115] Please see Figure 4 , Figure 4 This is another flowchart illustrating the network security testing and analysis method provided in this application embodiment. The network security testing and analysis method may include the following steps: Step 301: Obtain multiple historical network security test reports, perform text cleaning on the text corresponding to each historical network security test report, and obtain the target report text corresponding to each historical network security test report; Step 302: Determine the application type corresponding to each business system and the business system information associated with that application type based on each target report text; Step 303: Vectorize the information of each application type and associated business system to obtain the first set of associated information between each application type and associated business system information; Step 304: Determine at least one attack surface for each application type and the standard terms corresponding to each attack surface based on the network security database; Step 305: Input the standard words corresponding to each application type and each attack surface into the pre-trained large language model, and output the corresponding key-value pairs between each application type and each attack surface; Step 306: Establish a second set of association information between each application type and the corresponding attack surface based on key-value pairs; Step 307: Obtain the test business system information of the network security test report to be processed, as well as the multiple test attack surfaces associated with the test business system information; Step 308: Match the test business system information with the first set of related information to obtain the target application type corresponding to the test business system, and match the target application type with the second set of related information to determine multiple matching attack surfaces corresponding to the test business system; Step 309: Determine the similarity between each test attack surface and each matched attack surface, and among multiple matched attack surfaces, identify at least one candidate attack surface whose similarity is lower than the first preset similarity. Step 310: Input the target application type, test business system information and description text corresponding to at least one candidate attack surface into the pre-trained risk assessment model, and output the target attack surface not covered in the network security test report to be processed; Step 311: In the preset test environment, determine the attack tactics corresponding to the test business system information based on the target attack surface; Step 312: Conduct network security testing on the test business system according to the attack tactics, and obtain the corresponding updated network security test report for the test business system.

[0116] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed description of the above network security testing and analysis method, which will not be repeated here.

[0117] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed description of the above network security testing and analysis method, which will not be repeated here.

[0118] Please see Figure 5 , Figure 5 This is a schematic diagram of the network security testing and analysis apparatus provided in this application embodiment. This apparatus can execute the aforementioned network security testing and analysis method. In this application embodiment, the terms "module" or "unit" refer to a computer program or part of a computer program with a predetermined function, which works together with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0119] Network security testing and analysis device 400, including: The first acquisition module 410 is used to acquire multiple historical network security test reports and analyze the multiple historical network security test reports to obtain the first set of related information corresponding to each application type and each business system information. The generation module 420 is used to establish a second set of related information between each application type and the attack surface corresponding to each application type based on the network security database. The second acquisition module 430 is used to acquire the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information; The matching module 440 is used to match the test business system information with the first set of associated information to obtain the target application type corresponding to the test business system, and to match the target application type with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. The determination module 450 is used to determine the similarity between each test attack surface and each matched attack surface, and to determine at least one candidate attack surface with a similarity lower than a first preset similarity among multiple matched attack surfaces; The prediction module 460 is used to input the target application type, test business system information and description text corresponding to at least one candidate attack surface into the pre-trained risk assessment model, and output the target attack surface not covered in the network security test report to be processed.

[0120] In some implementations, the first acquisition module 410 is used for: Text cleaning is performed on the text corresponding to each historical network security test report to obtain the target report text corresponding to each historical network security test report; Based on each target report text, determine the application type corresponding to the business system and the business system information associated with the application type; Vectorize the information of each application type and its associated business system to obtain the first set of associated information between each application type and its associated business system.

[0121] In some implementations, the first acquisition module 410 is used for: Identify target texts within each target report text that have a word frequency higher than a preset word frequency; Based on the target text, determine the application type corresponding to the business system and the business system information associated with the application type.

[0122] In some implementations, the generation module 420 is used for: Based on the network security database, at least one attack surface corresponding to each application type was identified, along with the standard terms corresponding to each attack surface; Input the standard words corresponding to each application type and each attack surface into the pre-trained large language model, and output the corresponding key-value pairs between each application type and each attack surface; Based on key-value pairs, a second set of associated information is established between each application type and the corresponding attack surface.

[0123] In some implementations, the determining module 450 is used for: Among multiple matching attack surfaces, several initial attack surfaces with similarity lower than the first preset similarity and greater than the second preset similarity are identified; Select at least one candidate attack surface from multiple initial attack surfaces that is suitable for testing business system information.

[0124] In some implementations, the matching module 440 is used for: Obtain the first vector corresponding to the information of the test business system; Match the first vector with the vector corresponding to each business system information in the first associated information set to obtain the matching target business system information; The application type associated with the target business system information is determined as the target application type corresponding to the test business system.

[0125] In some implementations, the network security test analysis device 400 further includes a prediction module 460, which is used to: after inputting the target application type, test business system information and the description text corresponding to each candidate attack surface into a pre-trained risk assessment model and outputting the target attack surface not involved in the test in the network security test report to be processed, determine the attack tactics corresponding to the test business system information based on the target attack surface in a preset test environment. Based on the attack tactics, conduct cybersecurity tests on the test business system and obtain an updated cybersecurity test report for the corresponding test business system.

[0126] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed description of the above network security testing and analysis method, which will not be repeated here.

[0127] In this application, the first acquisition module 410 acquires multiple historical network security test reports and analyzes them to obtain a first set of associated information between each application type and each business system information; the generation module 420 establishes a second set of associated information between each application type and the attack surface corresponding to each application type based on the network security database; the second acquisition module 430 acquires the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information; the matching module 440 matches the test business system information with the first set of associated information to obtain the target application type corresponding to the test business system, and matches the target application type with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system; the determination module 450 determines the similarity between each test attack surface and each matching attack surface, and determines at least one candidate attack surface with a similarity lower than a first preset similarity among the multiple matching attack surfaces; the prediction module 460 inputs the target application type, test business system information, and description text corresponding to at least one candidate attack surface into a pre-trained risk assessment model, and outputs the target attack surface not involved in the test in the network security test report to be processed.

[0128] Therefore, by acquiring and analyzing multiple historical network security test reports, a first set of correlation information is obtained between each application type and each business system information. This allows for the discovery of the correlation between application types and business system information. Based on the network security database, a second set of correlation information is established between each application type and its corresponding attack surface. This allows for the discovery of the correlation between application types and attack surfaces.

[0129] Then, the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information, are obtained. The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system. The target application type is then matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. This allows for the accurate identification of multiple matching attack surfaces that may be vulnerable to attacks corresponding to the test business system.

[0130] Finally, the similarity between each test attack surface and each matched attack surface is determined. Among multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified. The target application type, test business system information, and descriptive text corresponding to at least one candidate attack surface are input into a pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report. This process identifies at least one candidate attack surface that has not been tested. Finally, the pre-trained risk assessment model identifies potential target attack surfaces that may pose a network attack risk, thus promptly identifying potential attack risks that have not been tested.

[0131] Compared to related technologies where security testing is conducted by testers within a limited scope, this application eliminates the need for manual intervention. It can automatically identify target attack surfaces that have not undergone cybersecurity testing and may pose cybersecurity risks. These target attack surfaces can be used to retest the tested business systems or to detect potential cyberattack risks in advance, thereby improving the efficiency and comprehensiveness of cybersecurity testing.

[0132] This application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned network security testing and analysis method. This computer device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0133] Please see Figure 6 , Figure 6 The hardware structure of a computer device according to another embodiment is illustrated. The computer device includes: The processor 501 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 502 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 502 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 502 and is called and executed by the processor 501 using the network security testing and analysis method of the embodiments of this application. The input / output interface 503 is used to implement information input and output; The communication interface 504 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 505 transmits information between various components of the device (e.g., processor 501, memory 502, input / output interface 503, and communication interface 504); The processor 501, memory 502, input / output interface 503, and communication interface 504 are connected to each other within the device via bus 505.

[0134] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned network security testing and analysis method.

[0135] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0136] The network security testing and analysis method, network security testing and analysis device, computer equipment, and storage medium provided in this application embodiment acquire multiple historical network security test reports and analyze them to obtain a first set of correlation information corresponding to each application type and each business system information. This allows for the discovery of the correlation between application types and business system information. Based on the network security database, a second set of correlation information corresponding to each application type and its corresponding attack surface is established. This allows for the discovery of the correlation between application types and attack surfaces.

[0137] Then, the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information, are obtained. The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system. The target application type is then matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. This allows for the accurate identification of multiple matching attack surfaces that may be vulnerable to attacks corresponding to the test business system.

[0138] Finally, the similarity between each test attack surface and each matched attack surface is determined. Among multiple matched attack surfaces, at least one candidate attack surface with a similarity lower than a first preset similarity is identified. The target application type, test business system information, and descriptive text corresponding to at least one candidate attack surface are input into a pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report. This process identifies at least one candidate attack surface that has not been tested. Finally, the pre-trained risk assessment model identifies potential target attack surfaces that may pose a network attack risk, thus promptly identifying potential attack risks that have not been tested.

[0139] Compared to related technologies where security testing is conducted by testers within a limited scope, this application eliminates the need for manual intervention. It can automatically identify target attack surfaces that have not undergone cybersecurity testing and may pose cybersecurity risks. These target attack surfaces can be used to retest the tested business systems or to detect potential cyberattack risks in advance, thereby improving the efficiency and comprehensiveness of cybersecurity testing.

[0140] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0141] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0142] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0143] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0144] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0145] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0146] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0147] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0148] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0149] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0150] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0151] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0152] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A network security testing and analysis method, characterized in that, include: Multiple historical network security test reports are obtained, and the multiple historical network security test reports are analyzed to obtain the first set of related information between each application type and each business system information; Based on the network security database, establish a second set of associated information between each application type and the attack surface corresponding to each application type; Obtain the test business system information of the network security test report to be processed, as well as the multiple test attack surfaces associated with the test business system information; The test business system information is matched with the first set of associated information to obtain the target application type corresponding to the test business system, and the target application type is matched with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. Determine the similarity between each test attack surface and each matched attack surface, and among the multiple matched attack surfaces, identify at least one candidate attack surface whose similarity is lower than a first preset similarity. The target application type, the test business system information, and the description text corresponding to the at least one candidate attack surface are input into the pre-trained risk assessment model, which outputs the target attack surface not covered in the network security test report to be processed.

2. The network security testing and analysis method according to claim 1, characterized in that, The analysis of multiple historical network security test reports yields a first set of correlation information between each application type and each business system information, including: Text cleaning is performed on the text corresponding to each of the historical network security test reports to obtain the target report text corresponding to each of the historical network security test reports; Based on each target report text, determine the application type corresponding to the business system and the business system information associated with the application type; Vectorize each application type and associated business system information to obtain a first set of associated information between each application type and associated business system information.

3. The network security testing and analysis method according to claim 2, characterized in that, The step of determining the application type corresponding to the business system and the business system information associated with the application type based on each target report text includes: In each target report text, identify target texts with word frequencies higher than a preset word frequency; Based on the target text, determine the application type corresponding to the business system and the business system information associated with the application type.

4. The network security testing and analysis method according to claim 1, characterized in that, The step of establishing a second set of association information between each application type and the corresponding attack surface based on the network security database includes: Based on the network security database, at least one attack surface corresponding to each application type is identified, as well as the standard term corresponding to each attack surface; Input the standard words corresponding to each application type and each attack surface into a pre-trained large language model, and output the corresponding key-value pairs between each application type and each attack surface; Based on the key-value pairs, a second set of associated information is established between each application type and the attack surface corresponding to each application type.

5. The network security testing and analysis method according to claim 1, characterized in that, The step of identifying at least one candidate attack surface with a similarity lower than a first preset similarity among multiple matched attack surfaces includes: Among the multiple matching attack surfaces, multiple initial attack surfaces with a similarity lower than a first preset similarity and greater than a second preset similarity are identified; At least one candidate attack surface suitable for the test business system information is selected from the plurality of initial attack surfaces.

6. The network security testing and analysis method according to claim 1, characterized in that, The step of matching the test business system information with the first associated information set to obtain the target application type corresponding to the test business system includes: Obtain the first vector corresponding to the information of the test business system; The first vector is matched with the vector corresponding to each business system information in the first associated information set to obtain the matching target business system information; The application type associated with the target business system information is determined as the target application type corresponding to the test business system.

7. The network security testing and analysis method according to claim 1, characterized in that, After inputting the target application type, the test business system information, and the descriptive text corresponding to each candidate attack surface into the pre-trained risk assessment model, and outputting the target attack surfaces not covered in the network security test report to be processed, the method further includes: In a preset test environment, the attack tactics corresponding to the test business system information are determined based on the target attack surface; Based on the attack tactics described, network security testing is conducted on the test business system to obtain an updated network security test report corresponding to the test business system.

8. A network security testing and analysis device, characterized in that, include: The first acquisition module is used to acquire multiple historical network security test reports and analyze the multiple historical network security test reports to obtain a first set of related information between each application type and each business system information. The generation module is used to establish a second set of associated information between each application type and the attack surface corresponding to each application type, based on the network security database. The second acquisition module is used to acquire the test business system information of the network security test report to be processed, as well as multiple test attack surfaces associated with the test business system information; The matching module is used to match the test business system information with the first set of associated information to obtain the target application type corresponding to the test business system, and to match the target application type with the second set of associated information to determine multiple matching attack surfaces corresponding to the test business system. The determination module is used to determine the similarity between each of the test attack surfaces and each of the matched attack surfaces, and to determine at least one candidate attack surface whose similarity is lower than a first preset similarity among the multiple matched attack surfaces; The prediction module is used to input the target application type, the test business system information, and the description text corresponding to the at least one candidate attack surface into a pre-trained risk assessment model, and output the target attack surface that is not covered in the network security test report to be processed.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to execute the network security testing and analysis method according to any one of claims 1 to 7.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the network security testing and analysis method according to any one of claims 1 to 7.