Automated Web Vulnerability Verification Methods, Devices, Electronic Equipment, and Storage Media

By combining multidimensional differential analysis and signal fusion logic calculation with semantic transformation templates, the false positive and false negative problems of Web vulnerability scanners are solved, and highly accurate automated verification of Web vulnerabilities is achieved.

CN123122998APending Publication Date: 2026-11-17SHENZHEN IBOX INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611528325.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-29
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing web vulnerability scanners suffer from high false positive and false negative rates during the verification phase. In particular, single-request keyword matching cannot distinguish between injected responses and normal pages, and it cannot effectively verify injection attacks such as blind injection attacks that do not display error messages.

Method used

By acquiring the baseline response data and injection response data of the target injection point to be verified, multidimensional differential analysis is performed to quantify the differences in multiple dimensions. The continuous evidence quality score of the vulnerability is obtained through multidimensional signal fusion logic calculation, and the business logic vulnerability is verified by combining semantic transformation templates.

Benefits of technology

Significantly reduces false positive rate, improves verification accuracy, covers blind injection missed detection, realizes automated detection of business logic vulnerabilities, and provides fine-grained verification results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122998A_ABST
    Figure CN123122998A_ABST
Patent Text Reader

Abstract

This application provides a method, apparatus, electronic device, and storage medium for automated verification of web vulnerabilities. The method includes: acquiring baseline response data and injection response data corresponding to the target injection point to be verified; performing multidimensional difference analysis on the baseline response data and the injection response data to obtain multidimensional difference representations; calculating a continuous evidence quality score indicating the degree of confidence in the existence of the vulnerability based on the difference representations through multidimensional signal fusion logic; and determining the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score. This application eliminates environmental noise interference by comparing the baseline and injection response of the same injection point, and uses multidimensional difference analysis and signal fusion generation to replace traditional binary judgment, effectively distinguishing between real vulnerability triggers and normal pages containing keywords, significantly reducing the false positive rate, and improving the detection accuracy and coverage of blind injection and business logic vulnerabilities without error feedback.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of network security technology, and specifically to a method, apparatus, electronic device, and storage medium for automated verification of web vulnerabilities. Background Technology

[0002] With the widespread adoption of web applications, security attacks targeting web applications are increasing. Enterprises typically use automated web vulnerability scanners to detect application security vulnerabilities. The workflow of a web vulnerability scanner usually includes three phases: reconnaissance, detection, and verification. The reconnaissance phase collects the target's Uniform Resource Locator (URL) and injection point; the detection phase sends attack payloads to the injection point; and the verification phase confirms whether the vulnerability actually exists.

[0003] Currently, most scanners use a single-request keyword matching mode in their verification phase. This means that after sending an attack payload to the injection point, the returned HTTP response body is searched for preset error keywords (such as SQL syntax errors, stack traces, etc.). A match indicates the existence of a vulnerability. In addition, some fuzzing tools use a single dimension, such as HTTP status code or response body length, to filter abnormal responses for verification.

[0004] However, the above verification methods have at least the following technical drawbacks: On the one hand, single-request keyword matching only checks whether the injected response itself contains the keyword, and cannot distinguish whether the error was triggered by the injected payload or whether the normal page naturally contains the keyword. When a normal page happens to contain the keyword, it will be mistakenly identified as a vulnerability, resulting in a high false positive rate. On the other hand, for injection attacks such as blind injection that do not produce error responses, no erroneous keywords are generated in the response of the injection request, and keyword matching completely fails, resulting in a high false negative rate. At the same time, the single-dimensional filtering verification method cannot distinguish between the anomalies caused by injection and the differences in responses caused by normal business logic, and the responses are not obtained in pairs in the same context. Environmental differences such as changes in server status will introduce noise, further affecting the accuracy of verification. Summary of the Invention

[0005] In view of the aforementioned problems, this application is made to provide an automated web vulnerability verification method, apparatus, electronic device, and storage medium that overcomes or at least partially solves the aforementioned problems, comprising: An automated web vulnerability verification method includes: Obtain the baseline response data and injection response data corresponding to the target injection point to be verified; Multidimensional difference analysis was performed on the baseline response data and the injected response data to obtain a multidimensional characterization of the differences. Based on the aforementioned difference characterization, a continuous evidence quality score representing the degree of certainty of vulnerability existence is obtained through multi-dimensional signal fusion logic calculation. Based on the continuous evidence quality score, the vulnerability verification result corresponding to the target injection point is determined.

[0006] Furthermore, the multidimensional difference analysis performed on the baseline response data and the injected response data yields a multidimensional characterization of the differences, including: The difference characterization is obtained by quantifying the difference between the baseline response data and the injected response data in at least two independent dimensions. The at least two independent dimensions include: The response body structure difference dimension characterizes the degree of change in the size or structure of the response body content. Status code differences dimension, representing the consistency changes of HTTP status codes; Error content difference dimension represents the addition of specific error patterns in the response body; The response header difference dimension represents the variation in the number of response header fields.

[0007] Furthermore, the quantification methods corresponding to the error content difference dimension include: Identify the number of first erroneous signatures in the injected response data and the number of second erroneous signatures in the baseline response data; Calculate the difference between the first number of erroneous signatures and the second number of erroneous signatures, and use the difference as a difference representation of the error content difference dimension; The error signature includes at least one of the following: database error message, programming language stack trace information, or framework exception information.

[0008] Furthermore, the continuous evidence quality score, which calculates the degree of certainty about the existence of a vulnerability based on the difference characterization through multi-dimensional signal fusion logic, includes: The independent score for each dimension is determined based on the difference representation and the preset scoring mapping rules; The independent scores for all dimensions are summed, and the summation result is truncated to a preset numerical range to obtain the continuous evidence quality score.

[0009] Further, obtaining the baseline response data and injection response data corresponding to the target injection point to be verified includes: A baseline request is sent to the target injection point using the original parameter values, and the baseline response data is received. While maintaining consistency with the baseline request context, the original parameter values ​​are changed to test parameter values ​​as the injection request; Send the injection request to the target injection point and receive the injection response data; Contextual consistency includes consistency in Uniform Resource Locators, HTTP methods, and non-test request headers.

[0010] Furthermore, after determining the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score, the method further includes: Apply a preset semantic transformation template to the business logic parameters to be verified to generate an image request; Obtain the image response data corresponding to the image request; The mirror response data is compared with the expected rejection behavior and the baseline response data, and the business logic vulnerability verification signal is determined based on the comparison results. Update or determine the vulnerability verification result based on the business logic vulnerability verification signal.

[0011] Furthermore, the semantic transformation template includes at least one of the following: The sign-flipping template inverts the sign of numerical parameters. Zero value bypass template, setting numerical parameters to zero; Extreme value overflow template, which replaces numerical parameters with extreme values ​​that exceed the normal range; A state transition template that replaces state parameters with unexpected target state values; Type obfuscation template, which replaces numeric parameters with strings containing numeric content.

[0012] An automated web vulnerability verification device includes: The acquisition module is used to acquire the baseline response data and injection response data corresponding to the target injection point to be verified. The analysis module is used to perform multidimensional difference analysis on the baseline response data and the injected response data to obtain a multidimensional difference characterization; The calculation module is used to calculate the continuous evidence quality score of the degree of certainty of the existence of the vulnerability based on the difference characterization through multi-dimensional signal fusion logic. The determination module is used to determine the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score.

[0013] A computer electronic device includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the automated web vulnerability verification method as described above.

[0014] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the automated web vulnerability verification method described above.

[0015] This application has the following advantages: In the embodiments of this application, in view of the problem of "high false positive rate and high false negative rate for blind injection and business logic vulnerabilities caused by single request keyword matching" in the prior art, this application provides a solution, specifically: "obtaining baseline response data and injection response data corresponding to the target injection point to be verified; performing multi-dimensional difference analysis on the baseline response data and the injection response data to obtain multiple dimensions of difference characterization; based on the difference characterization, calculating the continuous evidence quality score of the degree of confidence in the existence of the vulnerability through multi-dimensional signal fusion logic; and determining the vulnerability verification result corresponding to the target injection point according to the continuous evidence quality score." This application eliminates environmental noise interference and accurately captures subtle changes caused by injection by acquiring baseline response data and injection response data from the same injection point and performing multidimensional differential analysis. It obtains continuous evidence quality scores through multidimensional signal fusion logic calculation, replacing traditional binary judgment, enabling the system to quantify the degree of certainty regarding the existence of vulnerabilities. This effectively solves the false positive problem caused by single-request keyword matching in existing technologies, as erroneous signatures already present in the baseline are deducted in the differential calculation and no longer contribute to the score. Simultaneously, through comprehensive analysis of response body structure, status codes, and other dimensions, even in scenarios like blind injection where there are no obvious erroneous keywords, anomalies can be detected through signals such as length ratios or status code divergence, solving the problem of missed detections in blind injection. Furthermore, by applying semantic transformation templates to business logic parameters and comparing expected rejection behavior with actual responses, it fills the gap in traditional methods that cannot detect business logic vulnerabilities without error feedback, significantly improving the accuracy and coverage of automated verification. Attached Figure Description

[0016] To more clearly illustrate the technical solution of this application, the drawings used in the description 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.

[0017] Figure 1 This is a flowchart illustrating the steps of an automated web vulnerability verification method according to an embodiment of this application; Figure 2 This is a structural block diagram of an automated web vulnerability verification device provided in one embodiment of this application; Figure 3 This is a schematic diagram of the structure of a computer electronic device provided in an embodiment of the present invention; 1. A computer electronic device; 2. An external device; 3. A processing unit; 4. A bus; 5. A network adapter; 6. An I / O interface; 7. A display; 8. Memory; 9. Random access memory; 10. A cache memory; 11. A storage system; 12. A program / utility; 13. A program module. Detailed Implementation

[0018] To make the objectives, features, and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] The automated web vulnerability verification method provided in this application can be applied to the verification stage of an automated web vulnerability scanner to perform secondary confirmation on suspected vulnerabilities discovered during the detection stage. This verification stage takes over the target injection point output by the detection stage, and obtains paired response data by sending baseline requests and injection requests sequentially to the same injection point. Multidimensional differential analysis is then performed, and finally, a quantitative vulnerability verification result is output. In one specific implementation, the execution entity of this method can be a verification engine deployed on a security detection server or an independent verification module integrated into the scanner; this application does not specifically limit this.

[0020] It should be noted that, for ease of understanding of the technical solution of this application, the following explanations are provided for the key terms involved in the embodiments of this application: Paired requests: refers to sending two HTTP requests to the same injection point one after the other. The baseline request uses the original parameter values ​​and does not contain the attack payload, while the injection request uses the transformed test parameter values. The two are completely identical except for the parameters being tested, forming the data basis for differential analysis.

[0021] Evidence quality score: refers to a continuous value between 0.0 and 1.0, used to indicate the degree of confidence that the injection caused an abnormal response, replacing the traditional binary judgment method of "vulnerability exists / does not exist".

[0022] Error signature: refers to specific text patterns that appear in the response body, such as database error messages, programming language stack traces, and framework exception messages.

[0023] Mirror probe: refers to a probe request sent after applying parameter value transformation to business logic parameters. It discovers business logic vulnerabilities by comparing the actual response behavior with the expected rejection behavior.

[0024] Injection points: These are locations in an HTTP request where user input can be controlled, including URL query parameters, POST form fields, JSON attributes, HTTP headers, cookie values, and RESTful path segments.

[0025] Reference Figure 1 This application illustrates an embodiment of an automated web vulnerability verification method, specifically including: S101. Obtain the baseline response data and injection response data corresponding to the target injection point to be verified; S102. Perform multidimensional difference analysis on the baseline response data and the injection response data to obtain a multidimensional difference characterization; S103. Based on the difference characterization, the continuous evidence quality score of the degree of certainty of the existence of the vulnerability is obtained by multi-dimensional signal fusion logic calculation. S104. Determine the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score.

[0026] In the embodiments of this application, the above steps work together to achieve the technical effect of extracting multidimensional differential signals from paired response data and fusing multiple signals into a single continuous confidence value. Specifically, step S101 obtains baseline response data and injected response data through a paired request mechanism, solving the problem of lack of reference benchmark in single-request verification; step S102 quantifies the differences between the two sets of response data in multiple independent dimensions, so that the differences are no longer simply "different" but are decomposed into multiple difference representations with independent semantic sources; step S103 transforms the independent difference representations of each dimension into continuous evidence quality scores between 0 and 1 through multidimensional signal fusion logic, replacing the binary judgment of "hit equals vulnerability" in traditional keyword matching; step S104 determines the final vulnerability verification result based on the interval of the continuous evidence quality score, providing an interpretable quantitative basis for vulnerability labeling. The data flow between the above steps is as follows: the baseline response data and injected response data output in step S101 serve as inputs to step S102; the multi-dimensional difference representation output in step S102 serves as input to step S103; and the continuous evidence quality score output in step S103 serves as the judgment criterion for step S104. Throughout the entire chain, each step does not rely on the precondition of whether a preset keyword appears in a single response, but rather makes judgments based on the differential signals between paired responses, fundamentally avoiding the false alarm problem caused by keyword coincidence.

[0027] Taking a specific application scenario as an example: During the reconnaissance phase, the automated vulnerability scanner identifies the keyword parameter of the product search interface as a potential injection point. During the detection phase, an SQL injection test payload is sent to this injection point, followed by the verification phase. In the verification phase, step S101 obtains the baseline response data when the keyword parameter uses the original value "shirt" and the injection response data when the injection payload "shirt' OR '1'='1--'" is used. Step S102 quantifies the differences between the two sets of response data across four independent dimensions: response body structure, status code, error content, and response header. Step S103 scores and accumulates the independent differences in each dimension, truncating them to a preset numerical range to obtain the continuous evidence quality score corresponding to the injection point. Step S104 outputs the vulnerability verification result based on the judgment range of the score—when the score is in a lower range, even if the injection response contains the phrase "Database error occurred," it will not be considered a vulnerability because the same text already exists in the baseline response, and the error content difference dimension count is zero, thus automatically eliminating coincidental keywords.

[0028] This application's embodiments, by acquiring baseline response data and injection response data from the same injection point and performing multidimensional differential analysis, can eliminate environmental noise interference and accurately capture subtle changes caused by injection. By calculating continuous evidence quality scores through multidimensional signal fusion logic, replacing traditional binary judgment, the system can quantify the degree of certainty regarding the existence of vulnerabilities, effectively solving the false positive problem caused by single-request keyword matching in existing technologies. This is because erroneous signatures already present in the baseline are deducted during differential calculation and no longer contribute to the score. Simultaneously, through comprehensive analysis of response body structure, status codes, and other dimensions, even in scenarios like blind injection where there are no obvious erroneous keywords, anomalies can be detected through signals such as length ratios or status code divergence, solving the problem of missed detections in blind injection. Furthermore, by applying semantic transformation templates to business logic parameters and comparing expected rejection behavior with actual responses, the gap in traditional methods' inability to detect business logic vulnerabilities without error feedback is filled, significantly improving the accuracy and coverage of automated verification.

[0029] The automated web vulnerability verification method in this exemplary embodiment will be further described below.

[0030] In one embodiment of the present invention, the specific process of "performing multidimensional difference analysis on the baseline response data and the injected response data to obtain a multidimensional difference characterization" in step S102 can be further explained in conjunction with the following description.

[0031] As described in the following steps, the differences between the baseline response data and the injected response data are quantified in at least two independent dimensions to obtain the difference characterization; wherein, the at least two independent dimensions include: response body structure difference dimension, characterizing the degree of change in the size or structure of the response body content; status code difference dimension, characterizing the consistency change of HTTP status codes; error content difference dimension, characterizing the addition of specific error patterns in the response body; and response header difference dimension, characterizing the change in the number of response header fields.

[0032] In this embodiment, four independent dimensions characterize the server behavior deviation caused by the injection request from different observable perspectives: the response body structure difference dimension focuses on the relative changes in the size of the response body content between the baseline response data and the injected response data, used to capture significant expansion or contraction of page content caused by the injection payload; the status code difference dimension focuses on whether the HTTP status codes carried by the two sets of response data are consistent, used to capture abnormal server states caused by injection; the error content difference dimension focuses on whether the injected response data adds specific error patterns that are not present in the baseline response data, used to capture backend anomalies actually triggered by the injection; and the response header difference dimension focuses on the changes in the number of response header fields in the two sets of response data, used to capture anomalies at the response construction level caused by injection. By adopting the above multi-dimensional quantification method, even if one dimension does not generate a signal under a specific vulnerability type, the other dimensions can still provide evidence independently, avoiding the overall failure of single-dimensional methods when facing blind injection or logical vulnerabilities.

[0033] In one embodiment of the present invention, the above-mentioned "error content difference dimension" can be further described in detail with reference to the following description.

[0034] As described below, the quantification method corresponding to the error content difference dimension includes: identifying the number of first error signatures in the injected response data and the number of second error signatures in the baseline response data; calculating the difference between the number of first error signatures and the number of second error signatures, and using the difference as the difference representation of the error content difference dimension; wherein, the error signature includes at least one of database error messages, programming language stack trace information, or framework exception information.

[0035] In this embodiment, the core design principle of this dimension is "incremental counting" rather than "existence counting": only error signatures that appear newly in the injected response data compared to the baseline response data are counted as differences. This way, if the baseline response data itself contains error-related text (e.g., "Database error occurred" is included in the normal error message of an e-commerce platform), this text will not contribute any increment when it reappears in the injected response data, thus preventing misjudgments caused by inherent page text at the mechanism level. As a specific implementation, the error signature library can cover 12 cross-engine patterns, including SQL dialect error patterns (including MySQL, PostgreSQL, Oracle, MSSQL, SQLite), programming language exception patterns (including Java, .NET, Node.js, PHP), and infrastructure error patterns (including ODBC, JDBC). These patterns are organized in the form of regular expressions and can be selectively enabled based on the target application's runtime technology stack. Simultaneously, for some technology stacks, an error page structure identification method based on Abstract Syntax Trees (AST) can be used instead of regular expression matching as a supplementary implementation method for error signature identification.

[0036] In one embodiment of the present invention, the specific process of step S103, "based on the difference characterization, calculating the continuous evidence quality score of the degree of certainty of the existence of the vulnerability through multi-dimensional signal fusion logic", can be further explained in conjunction with the following description.

[0037] As described in the following steps, the independent score for each dimension is determined based on the difference representation and the preset scoring mapping rule; the independent scores for all dimensions are accumulated, and the accumulated result is truncated to a preset numerical range to obtain the continuous evidence quality score.

[0038] In this embodiment, a preset scoring mapping rule defines a mapping relationship from difference representation to independent score for each dimension, ensuring that each dimension contributes to the final score with a uniform scale. The independent scores of all dimensions are accumulated and then truncated to the [0.0, 1.0] interval, guaranteeing that the continuous evidence quality score is always bounded and comparable. Using continuous numerical values ​​instead of binary judgments allows the verification results to express intermediate states of "evidence strength," providing security analysts with more granular decision-making information and avoiding boundary misjudgments caused by hard threshold judgments. In one example, the scoring mapping rule can use the ratio of the injected response body length to the baseline response body length in the response body structure difference dimension as the difference representation of that dimension, and combine it with a preset ratio interval to map different scores; the status code difference dimension is assigned a fixed score when the status codes are inconsistent; the error content difference dimension is assigned an increasing score according to the interval of the incremental number of error signatures; and the response header difference dimension is assigned a score according to the interval of the absolute value of the difference in the number of response headers. Preferably, the independent scoring weights of each dimension are configured with preset values ​​in the initial stage. After the system accumulates a certain number of verification records (e.g., greater than or equal to 100 verifications), the weights can be dynamically adjusted based on the Pearson correlation coefficient between the signals of each dimension and the actual vulnerability results, replacing the fixed weight configuration.

[0039] In one embodiment of the present invention, the specific process of "obtaining baseline response data and injection response data corresponding to the target injection point to be verified" in step S101 can be further described in conjunction with the following description.

[0040] As described in the following steps, a baseline request is sent to the target injection point using the original parameter values, and the baseline response data is received; while maintaining the consistency of the baseline request context, the original parameter values ​​are changed to test parameter values ​​as the injection request; the injection request is sent to the target injection point, and the injection response data is received; wherein, the consistency of context includes consistency of Uniform Resource Locator, HTTP method, and non-test request headers.

[0041] In this embodiment, the key to the paired request mechanism is that the baseline request and the injection request remain completely identical in URL, HTTP method, and non-test request headers, except for the parameters being tested. This ensures that any differences between the two sets of response data are attributed to changes in test parameter values ​​rather than environmental fluctuations. This mechanism eliminates noise interference from server state drift, session context differences, etc., at the source, and forms the basis for the reliability of subsequent multidimensional differential analysis. In one example, if the target injection point is located in a POST form field, the baseline request and the injection request share the same URL path, Request Method, Cookie, User-Agent, Content-Type, and all other form fields except the field being tested. If the target injection point is located in a RESTful path segment, the path prefix and other path segments of the two requests remain consistent, with only the content of the tested path segment differing.

[0042] In one embodiment of the present invention, the specific features of the "Web vulnerability automated verification method" after step S104 can be further described in conjunction with the following description.

[0043] As described in the following steps, a preset semantic transformation template is applied to the business logic parameters to be verified to generate an image request; the image response data corresponding to the image request is obtained; the image response data is compared with the expected rejection behavior and the baseline response data, and a business logic vulnerability verification signal is determined based on the comparison result; the vulnerability verification result is updated or determined according to the business logic vulnerability verification signal.

[0044] In this embodiment, the verification path for business logic parameters is relatively independent of the aforementioned verification path based on multidimensional difference: business logic vulnerabilities typically do not generate any error signatures, and the injected response data is highly similar to the baseline response data in terms of status code and response body structure. Therefore, it is difficult to effectively identify vulnerabilities using only four-dimensional difference scores. Mirror request verification introduces "expected rejection behavior" as a reference benchmark, transforming the judgment dimension from "whether the response is abnormal" to "whether the server should have rejected but did not," thereby capturing illegal business operations that cannot be distinguished by traditional difference dimensions. Specifically, after generating a mirror request using a semantic transformation template and obtaining mirror response data, the mirror response data is compared with both the expected rejection behavior and the baseline response data: if the server does not actually reject the transformed parameter value (e.g., returning status code 200 instead of 400 / 403 / 422), and the mirror response data is highly consistent with the baseline response data in terms of response body structure, a strong business logic vulnerability verification signal is generated. Based on this business logic vulnerability verification signal, existing vulnerability verification results are corrected or directly determined, thereby achieving a unified output of vulnerability verification results.

[0045] In one embodiment of the present invention, the above-mentioned "semantic transformation template" can be further described in detail with reference to the following description.

[0046] In this embodiment, the semantic transformation template includes at least one of the following: a sign-flipping template, which inverts the sign of the numerical parameter; a zero-value bypass template, which sets the numerical parameter to zero; an extreme value overflow template, which replaces the numerical parameter with an extreme value that exceeds the normal range; a state transition template, which replaces the state parameter with an unexpected target state value; and a type obfuscation template, which replaces the numerical parameter with a string containing numeric content. These five semantic transformation templates target different types of business logic defects: the sign-flipping template is used to detect negative value acceptance vulnerabilities, i.e., the server incorrectly accepts negative prices or amounts; the zero-value bypass template is used to detect vulnerabilities related to free orders or zero-dollar payments; the extreme value overflow template is used to detect vulnerabilities where extreme values ​​exceeding the normal business range are accepted; the state transition template is used to detect state machine bypass vulnerabilities; and the type obfuscation template is used to detect logic vulnerabilities caused by the backend's lenient handling of parameter types. As a specific implementation, taking the payment amount parameter `amount=100` as an example, the sign-flipping template transforms the parameter value to `amount=-100`, the zero-value bypass template transforms it to `amount=0`, the extreme value overflow template transforms it to `amount=999999999`, and the type confusion template transforms it to `amount=100_string`. Taking the order status parameter `status=pending` as an example, the status transition template can directly replace the parameter value with the unexpected target status value `completed`. The expected rejection behavior corresponding to each template is that the server should return a rejection status code such as 400, 403, or 422; the actual behavior is a 200 status code, and a consistent response body constitutes a strong business logic vulnerability verification signal. In one example, the business logic vulnerability verification signal is also output in the form of a continuous evidence quality score, for example, mapping the case of "should be rejected but accepted and the response body is consistent" to a numerical output of 0.85.

[0047] As can be seen from the above embodiments, the automated Web vulnerability verification method disclosed in this application has at least the following technical effects: First, it significantly reduces the false positive rate at the mechanism level. Because the error content difference dimension uses incremental counting logic, error signatures that already exist in the baseline response data will not generate any differential signals when they reappear in the injected response data, thus making coincidental matching caused by inherent error text on regular pages no longer a basis for judgment.

[0048] Second, it achieves effective coverage of blind SQL injection vulnerabilities without feedback. The response body structure difference dimension, status code difference dimension, and optional response time dimension can independently provide evidence signals without any erroneous signatures emerging, compensating for the complete failure of keyword matching methods in blind SQL injection scenarios. In one example, server latency or empty responses triggered by blind SQL injection payloads can be reflected as a signal of a significant downward trend in the response body length ratio, thereby obtaining an evidence quality score that can enter the judgment interval.

[0049] Third, it achieves automated detection of business logic vulnerabilities for the first time. By comparing five semantic transformation templates with expected rejection behaviors, error-free keyword scenarios such as accepting negative payment amounts, bypassing zero values, and state transitions are all transformed into quantifiable business logic vulnerability verification signals, enabling the automatic identification of illegal business operations that traditional scanners cannot distinguish from normal requests.

[0050] Fourth, the evidence quality score is represented by a continuous numerical value, combined with a preset interval division (e.g., not less than 0.5 is marked as suspected, not less than 0.7 is marked as high confidence), which enables the verification results to have fine-grained expression from low to high, allowing security teams to set different confirmation thresholds according to actual security strategies, and avoiding the fragility problem of binary judgment near the boundary.

[0051] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0052] Reference Figure 2 This illustration shows an automated web vulnerability verification device 200 provided in an embodiment of this application, used to implement the automated web vulnerability verification method provided in the embodiments of this application. The device specifically includes: The acquisition module 201 is used to acquire the baseline response data and injection response data corresponding to the target injection point to be verified. Analysis module 202 is used to perform multidimensional difference analysis on the baseline response data and the injected response data to obtain a multidimensional difference characterization; Calculation module 203 is used to calculate the continuous evidence quality score of the degree of certainty of vulnerability existence based on the difference characterization through multi-dimensional signal fusion logic. The determination module 204 is used to determine the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score.

[0053] In one embodiment of the present invention, the analysis module 202 is further configured to: The difference characterization is obtained by quantifying the difference between the baseline response data and the injected response data in at least two independent dimensions. The at least two independent dimensions include: The response body structure difference dimension characterizes the degree of change in the size or structure of the response body content. Status code differences dimension, representing the consistency changes of HTTP status codes; Error content difference dimension represents the addition of specific error patterns in the response body; The response header difference dimension represents the variation in the number of response header fields.

[0054] In one embodiment of the present invention, the quantification method corresponding to the error content difference dimension includes: Identify the number of first erroneous signatures in the injected response data and the number of second erroneous signatures in the baseline response data; Calculate the difference between the first number of erroneous signatures and the second number of erroneous signatures, and use the difference as a difference representation of the error content difference dimension; The error signature includes at least one of the following: database error message, programming language stack trace information, or framework exception information.

[0055] In one embodiment of the present invention, the calculation module 203 is further configured to: The independent score for each dimension is determined based on the difference representation and the preset scoring mapping rules; The independent scores for all dimensions are summed, and the summation result is truncated to a preset numerical range to obtain the continuous evidence quality score.

[0056] In one embodiment of the present invention, the acquisition module 201 is further configured to: A baseline request is sent to the target injection point using the original parameter values, and the baseline response data is received. While maintaining consistency with the baseline request context, the original parameter values ​​are changed to test parameter values ​​as the injection request; Send the injection request to the target injection point and receive the injection response data; Contextual consistency includes consistency in Uniform Resource Locators, HTTP methods, and non-test request headers.

[0057] In one embodiment of the present invention, the automated web vulnerability verification device 200 is further configured to: Apply a preset semantic transformation template to the business logic parameters to be verified to generate an image request; Obtain the image response data corresponding to the image request; The mirror response data is compared with the expected rejection behavior and the baseline response data, and the business logic vulnerability verification signal is determined based on the comparison results. Update or determine the vulnerability verification result based on the business logic vulnerability verification signal.

[0058] In one embodiment of the present invention, the semantic transformation template includes at least one of the following: The sign-flipping template inverts the sign of numerical parameters. Zero value bypass template, setting numerical parameters to zero; Extreme value overflow template, which replaces numerical parameters with extreme values ​​that exceed the normal range; A state transition template that replaces state parameters with unexpected target state values; Type obfuscation template, which replaces numeric parameters with strings containing numeric content.

[0059] Reference Figure 3 The illustration shows a computer electronic device 1 for implementing the automated web vulnerability verification method of the present invention, which may specifically include: The computer electronic device 1 is manifested in the form of a general computing device, including: one or more processors or processing units 3, memory 8, and a bus 4 connecting different system components (including memory 8 and processing unit 3).

[0060] Bus 4 represents one or more of several bus architectures, including memory buses or memory controllers, peripheral buses, graphics acceleration ports, processors, or local buses using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Audio / Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0061] Computer electronic device 1 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by computer electronic device 1, including volatile and non-volatile media, removable and non-removable media.

[0062] Memory 8 may include computer system readable media in the form of volatile memory, such as random access memory 9 and / or cache memory 10. A computer electronic device 1 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 11 may be used to read and write non-removable, non-volatile magnetic media (commonly referred to as a "hard disk drive"). Although Figure 3As not shown, a disk drive for reading and writing to a removable non-volatile disk (such as a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (such as a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 4 via one or more data media interfaces. The memory may include at least one program product having a set (e.g., at least one) of program modules 13 configured to perform the functions of the embodiments of this application.

[0063] A program / utility 12 having a set (at least one) of program modules 13 may be stored, for example, in memory. Such program modules 13 include—but are not limited to—an operating system, one or more application programs, other program modules 13, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 13 typically perform the functions and / or methods described in the embodiments of this application.

[0064] The computer electronic device 1 can also communicate with one or more external devices 2 (e.g., keyboard, pointing device, display 7, camera, etc.), and with one or more devices that enable an operator to interact with the computer electronic device 1, and / or with any device that enables the computer electronic device 1 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through the I / O interface 6. Furthermore, the computer electronic device 1 can also communicate with one or more networks (e.g., local area network (LAN)), wide area network (WAN), and / or public networks (e.g., the Internet) through the network adapter 5. Figure 3 As shown, network adapter 5 communicates with other modules of computer electronic device 1 via bus 4. It should be understood that, although... Figure 3 Not shown, it may be combined with other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processing unit 3, external disk drive array, RAID system, tape drive and data backup storage system 11, etc.

[0065] The processing unit 3 executes various functional applications and data processing by running programs stored in memory 8, such as implementing the automated Web vulnerability verification method provided in the embodiments of this application.

[0066] That is, when the above-mentioned processing unit 3 executes the above-mentioned program, it achieves the following: obtaining the baseline response data and injection response data corresponding to the target injection point to be verified; performing multidimensional difference analysis on the baseline response data and the injection response data to obtain the difference characterization in multiple dimensions; based on the difference characterization, calculating the continuous evidence quality score of the degree of confidence in the existence of the vulnerability through multidimensional signal fusion logic; and determining the vulnerability verification result corresponding to the target injection point according to the continuous evidence quality score.

[0067] In this application embodiment, the application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the automated Web vulnerability verification method provided in all embodiments of the application.

[0068] That is, when the program is executed by the processor, it implements the following: acquiring baseline response data and injection response data corresponding to the target injection point to be verified; performing multidimensional difference analysis on the baseline response data and the injection response data to obtain multiple dimensions of difference characterization; based on the difference characterization, calculating the continuous evidence quality score of the degree of confidence in the existence of the vulnerability through multidimensional signal fusion logic; and determining the vulnerability verification result corresponding to the target injection point according to the continuous evidence quality score.

[0069] Any combination of one or more computer-readable media may be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0070] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0071] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the operator's computer, partially on the operator's computer, as a standalone software package, partially on the operator's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the operator's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider). The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably.

[0072] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0073] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0074] The above provides a detailed description of the automated Web vulnerability verification method, apparatus, electronic device, and storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for automated verification of web vulnerabilities, characterized in that, include: Obtain the baseline response data and injection response data corresponding to the target injection point to be verified; Multidimensional difference analysis was performed on the baseline response data and the injected response data to obtain a multidimensional characterization of the differences. Based on the aforementioned difference characterization, a continuous evidence quality score representing the degree of certainty of vulnerability existence is obtained through multi-dimensional signal fusion logic calculation. Based on the continuous evidence quality score, the vulnerability verification result corresponding to the target injection point is determined.

2. The method according to claim 1, characterized in that, The multidimensional difference analysis of the baseline response data and the injected response data yields a multidimensional characterization of the differences, including: The difference characterization is obtained by quantifying the difference between the baseline response data and the injected response data in at least two independent dimensions. The at least two independent dimensions include: The response body structure difference dimension characterizes the degree of change in the size or structure of the response body content. Status code differences dimension, representing the consistency changes of HTTP status codes; Error content difference dimension represents the addition of specific error patterns in the response body; The response header difference dimension represents the variation in the number of response header fields.

3. The method according to claim 2, characterized in that, The quantification methods corresponding to the error content difference dimensions include: Identify the number of first erroneous signatures in the injected response data and the number of second erroneous signatures in the baseline response data; Calculate the difference between the first number of erroneous signatures and the second number of erroneous signatures, and use the difference as a difference representation of the error content difference dimension; The error signature includes at least one of the following: database error message, programming language stack trace information, or framework exception information.

4. The method according to claim 1, characterized in that, The continuous evidence quality score, which calculates the degree of certainty about the existence of a vulnerability based on the aforementioned difference characterization through multi-dimensional signal fusion logic, includes: The independent score for each dimension is determined based on the difference representation and the preset scoring mapping rules; The independent scores for all dimensions are summed, and the summation result is truncated to a preset numerical range to obtain the continuous evidence quality score.

5. The method according to claim 1, characterized in that, The process of obtaining the baseline response data and injection response data corresponding to the target injection point to be verified includes: A baseline request is sent to the target injection point using the original parameter values, and the baseline response data is received. While maintaining consistency with the baseline request context, the original parameter values ​​are changed to test parameter values ​​as the injection request; Send the injection request to the target injection point and receive the injection response data; Contextual consistency includes consistency in Uniform Resource Locators, HTTP methods, and non-test request headers.

6. The method according to any one of claims 1 to 5, characterized in that, After determining the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score, the method further includes: Apply a preset semantic transformation template to the business logic parameters to be verified to generate an image request; Obtain the image response data corresponding to the image request; The mirror response data is compared with the expected rejection behavior and the baseline response data, and the business logic vulnerability verification signal is determined based on the comparison results. Update or determine the vulnerability verification result based on the business logic vulnerability verification signal.

7. The method according to claim 6, characterized in that, The semantic transformation template includes at least one of the following: The sign-flipping template inverts the sign of numerical parameters. Zero value bypass template, setting numerical parameters to zero; Extreme value overflow template, which replaces numerical parameters with extreme values ​​that exceed the normal range; A state transition template that replaces state parameters with unexpected target state values; Type obfuscation template, which replaces numeric parameters with strings containing numeric content.

8. A web vulnerability automated verification device, characterized in that, include: The acquisition module is used to acquire the baseline response data and injection response data corresponding to the target injection point to be verified. The analysis module is used to perform multidimensional difference analysis on the baseline response data and the injected response data to obtain a multidimensional difference characterization. The calculation module is used to calculate the continuous evidence quality score of the degree of certainty of the existence of the vulnerability based on the difference characterization through multi-dimensional signal fusion logic. The determination module is used to determine the vulnerability verification result corresponding to the target injection point based on the continuous evidence quality score.

9. A computer electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the automated web vulnerability verification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the Web vulnerability automated verification method as described in any one of claims 1 to 7.