CN-122027251-A - Weak password scanning and cleaning system based on machine learning and automatic blasting

CN122027251ACN 122027251 ACN122027251 ACN 122027251ACN-122027251-A

Abstract

The invention discloses a weak password scanning and cleaning system based on machine learning and automatic blasting, which is characterized by comprising a data acquisition layer, an analysis processing layer, a detection execution layer, a treatment closed loop layer and a visual display layer, wherein the data acquisition layer is used for acquiring asset information, internet leakage passwords, open social database data and internal password strategies, the analysis processing layer is used for generating a dynamic codebook, fusing multi-source data and applying an AI algorithm to infer the weak passwords, the detection execution layer comprises a distributed detection engine and is used for executing weak password detection on target assets, the treatment closed loop layer is used for distributing a correction task, checking password intensity and executing retest and confirmation to form a treatment closed loop, and the visual display layer is used for displaying situation maps, trend reports and compliance audit information. The invention combines an intelligent detection means with a systematic treatment flow, not only solves the bottleneck of the prior art in technical capability, but also realizes upgrading from the aspect of safety management concept, provides a set of efficient, comprehensive and landable weak password comprehensive treatment solution for enterprises and public institutions, and has extremely high practical value and popularization prospect.

Inventors

CHEN HAN
YANG ZHICHONG
WANG TAO

Assignees

浙江海瑞网络科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (10)

1. A weak password scanning and cleaning system based on machine learning and automatic blasting, comprising the following modules: The data acquisition layer is used for acquiring asset information, internet leakage passwords, open social database data and internal password policies; the analysis processing layer is used for generating a dynamic codebook, fusing multi-source data and applying an AI algorithm to infer a weak password; the detection execution layer comprises a distributed detection engine and is used for executing weak password detection on the target asset; the control closed loop layer is used for distributing a rectification task, checking password strength, and executing retest and confirmation to form a control closed loop; The visual display layer is used for displaying situation maps, trend reports and compliance audit information.
2. A machine learning and automated blasting-based weak password scanning and cleaning system according to claim 1, wherein the analysis processing layer comprises: a machine learning speculation module to generate a weak password based on PCFG rules and a deep learning model; And the dynamic codebook generating module is used for generating a dynamic codebook by combining the static dictionary, the leakage password and the AI presumption password.
3. The machine learning and automated blasting-based weak password scanning and cleaning system of claim 2, wherein the machine learning speculation module comprises: The PCFG rule generation submodule is used for generating a weak password according to the probability context-free grammar; and the deep learning prediction sub-module is used for generating a personalized weak password based on the organization characteristics by adopting an LSTM or a transducer model.
4. The machine learning and automated blasting-based weak password scanning and cleaning system of claim 1, wherein the detection execution layer comprises: The dynamic page automatic blasting module is used for identifying and blasting Web login points; the verification code automatic identification module is used for identifying character type or slider type verification codes; and the multi-asset weak password detection platform supports unified detection of a Web system, middleware, network equipment and an operating system.
5. The machine learning and automated blasting-based weak password scanning and cleaning system of claim 4, wherein the dynamic page automated blasting module comprises: the pre-rendering sub-module is used for dynamically rendering the target page; the login point identification sub-module performs semantic identification on the page elements by using the large model; and the blasting execution sub-module is used for executing automatic blasting and supporting failed retry and rate control.
6. The weak password scanning and cleaning system based on machine learning and automatic blasting according to claim 4, wherein the verification code automatic identification module adopts a CNN or Transformer architecture, and is trained to identify multiple types of verification codes.
7. The machine learning and automated blasting-based weak password scanning and cleaning system of claim 1, wherein the governance closed loop layer further comprises: modifying a task distribution mechanism; A password strength verification module; and the retest and confirmation module is used for verifying the correction effect and realizing weak password zero clearing.
8. A machine learning and automated blasting-based weak password scanning and cleaning system according to claim 1, wherein the system further comprises a security protection mechanism comprising: Detecting a rate limit; A rights control mechanism; And a data desensitization processing module.
9. The weak password scanning and cleaning system based on machine learning and automatic blasting according to claim 1, comprising the weak password detection and cleaning method comprising the steps of: Collecting whole network asset information and identifying login points; generating a dynamic codebook, and combining a static dictionary, a leakage password and an AI (advanced technology association) presumption password; performing weak password detection using a distributed detection engine; rectifying, checking and retesting the detection result to form a treatment closed loop; and displaying the detection result and the treatment progress through a visual interface.
10. The machine learning and automated blasting-based weak password scanning and cleaning system of claim 9, wherein the step of generating a dynamic codebook comprises: Generating a common weak password pattern using PCFG rules; a personalized weak password is generated based on the tissue features using a deep learning model.

Description

Weak password scanning and cleaning system based on machine learning and automatic blasting Technical Field The invention relates to the technical field of intelligent wearing, in particular to a weak password scanning and cleaning system based on machine learning and automatic blasting. Background In the fields of network and information security, the problem of weak passwords is always one of the main hidden hazards threatening the security of the information system of enterprises and public institutions. Weak passwords generally refer to passwords that are low in strength, easily guessed or broken by automated machinery, such as simple combinations of "123456", "admin", etc., or "company@2024" that are highly correlated to tissue information, etc. In real network attack and defense exercise, an attacker often takes a weak password as a first-choice break, and once the attacker gets his hands, the attacker can bypass complex peripheral safety protection and directly acquire system rights, so that serious consequences such as data leakage, service interruption and the like are caused. Although the risk of weak passwords is generally known in the industry, the existing weak password detection and management technology still has significant shortages, so that the problem cannot be effectively solved for a long time. The prior art mainly has the following two layers of defects of 1 1. The discovery capability of the asset and login points is insufficient, and a detection blind area exists. Currently, IT assets for enterprises and institutions are large in scale, complex in system type and distributed in deployment. Existing weak password detection techniques mostly rely on asset inventory manually maintained by an operator or network scan tools with limited functionality (such as Nmap port scanners). The method can not realize automation, persistent discovery and carding of the whole network assets, and is difficult to comprehensively identify all exposed login interfaces (such as Web login pages, database management ports, network equipment management interfaces and the like). These undiscovered "shadow assets" and login points become dead zones for security detection, which are very vulnerable to exploitation by attackers, constituting a serious security threat. 2. The detection means is static and one-sided, and the coverage rate and the intelligent degree are low. Currently, the mainstream weak password detection tools (such as hydro, medusa, etc.) generally adopt a "blasting" mode based on a fixed codebook. This static detection mechanism has inherent limitations: The detection range is limited, and a preset static dictionary usually only contains common general weak passwords and cannot cover real user passwords which are continuously leaked on the Internet. Pertinence is lacking in that a speculative password with a high hit rate cannot be intelligently generated according to characteristics of a target organization (such as company name, domain name, year of establishment, business characteristics). Dynamic update and deletion, namely, the content of the codebook is stiff, and is difficult to dynamically evolve along with the change of social engineering skills and user secret setting habits, so that the detection effect of the novel weak password is poor. In summary, in the prior art, due to the key defects of both comprehensive discovery and intelligent detection of the assets, the treatment work of the weak password is always remained in the stage of treating the symptoms without treating the root causes, and effective closed loop management cannot be formed. Therefore, there is an urgent need in the art for a systematic solution that can fully discover login points and perform weak password detection and management in a dynamic intelligent manner. Disclosure of Invention The invention provides a weak password scanning and cleaning system based on machine learning and automatic blasting, which aims to overcome the defects existing in the prior art. The technical scheme of the invention is realized as follows: A weak password scanning and cleaning system based on machine learning and automatic blasting, comprising the following modules: The data acquisition layer is used for acquiring asset information, internet leakage passwords, open social database data and internal password policies; the analysis processing layer is used for generating a dynamic codebook, fusing multi-source data and applying an AI algorithm to infer a weak password; the detection execution layer comprises a distributed detection engine and is used for executing weak password detection on the target asset; the control closed loop layer is used for distributing a rectification task, checking password strength, and executing retest and confirmation to form a control closed loop; The visual display layer is used for displaying situation maps, trend reports and compliance audit information. Preferably, the analysis processing layer