All Products
Search
Document Center

Web Application Firewall:Fraud Detection

Last Updated:Aug 20, 2025

Web Application Firewall (WAF) has a built-in mobile phone number reputation library to help prevent threats, such as spam user registration and marketing fraud. WAF compares mobile phone numbers or the MD5 hash values of phone numbers with the reputation library. If specific requests match suspicious behavior tags, WAF requires slider CAPTCHA verification, blocks the requests, records the requests, or adds tags to the requests.

Prerequisites

  • Web services are added to WAF on the Website Configuration page. For more information, see Website configuration overview.

  • You have enabled bot management and created a web protection template or an app protection template.

Billing

  • If you configure rules and traffic hits the rules, you are charged based on the number of hits. You are charged USD 0.007 per hit.

  • You are not charged if you enable the Fraud Detection feature but do not configure any rules, or if no traffic matches the configured rules.

Note

Fees for the fraud detection feature are included in the bills of WAF. Bills are generated on a daily basis.

Configure Fraud Detection

Configuration item

Description

Account Extraction

Customize the Account Type and Location. You can add a maximum of five conditions. The conditions are evaluated using a logical OR.

Account fetch: If a logon request uses the GET method and the request parameter is username=158***&password=***, set Location to Query Parameter and enter username for Parameter Name. This allows WAF to fetch the phone number for subsequent matching.

Risk Tag

  • Suspicious Sock Puppet Account: A phone number that seems to be provided by a third-party platform and does not belong to an individual user. The default risk level is High.

  • Fraud Risk: A phone number that seems to have been involved in fraudulent activities. The default risk level is High.

  • Spam User Registration: A phone number that seems to be used with malicious tools to register user accounts for subsequent marketing campaigns. The default risk level is High.

  • Marketing Fraud: A phone number that seems to be used with malicious tools to participate in marketing campaigns, such as registering multiple accounts to claim coupons. The default risk level is High.

  • Auto-purchase Bot: A phone number that seems to be used with malicious tools to participate in ticket grabbing or other flash sale events. The default risk level is High.

The risk levels are High, Medium-high, and Medium. Select a risk level as needed.

Rule Type

Suspected Bot: Exhibits some characteristics of a crawler but lacks direct evidence, such as attack traces or clear intent. Further verification of its intent is required.

Malicious Bot: An automated program that uses automated methods to attack a target system or network, steal data, or perform malicious operations for illicit purposes.

Actions

  • Block: Blocks requests that hit the rule and returns a block page to the client that initiated the requests.

    Note

    By default, WAF returns a preconfigured block page. You can use the custom response feature to configure a custom block page. For more information, see Configure a custom block page.

  • Monitor: Does not block requests that hit the rule. It only records the requests in logs. You can query the logs in WAF to find requests that hit the rule and analyze the rule's effectiveness, for example, to check for false positives.

    The monitor mode lets you test newly configured rules. After you confirm that no false positives are generated, you can set the action to another mode.

  • JavaScript Validation: WAF returns a JavaScript snippet to the client that a normal browser can automatically execute. If the client executes the JavaScript code, WAF allows all subsequent requests from the client for a period (30 minutes by default) without further challenges. Otherwise, the request is blocked.

    Note

    If you enable Fraud Detection for app protection, you cannot select JavaScript Challenge as the action.

  • Slider CAPTCHA: WAF returns a slider verification page to the client. If the client passes the slider verification, WAF allows all subsequent requests from the client for a period (30 minutes by default) without further verification. Otherwise, the request is blocked.

  • Strict Slider CAPTCHA Verification: WAF returns a slider verification page to the client. If the client passes the slider verification, WAF allows the current request. Otherwise, the request is blocked. In Strict Slider mode, every request from the client requires verification.

  • Add Tag: Lets you customize the header name and content (such as rule type, rule ID, and web UMID). WAF does not process the request directly. Instead, it adds a header to forward the hit information to the origin server. You can integrate this with your backend risk control system for business-side processing.

Canary Release

Configure the percentage of objects of a specific dimension to which the rule applies.

After you enable the grayscale rule, you must set the grayscale Dimension and Canary Release Proportion. The grayscale Dimension can be IP, Custom Header, Custom Parameter, Custom Cookie, Session, or Web UMID.

Note

A grayscale rule is applied based on the Dimension you set, not randomly to a percentage of requests. For example, if you select IP as the dimension, all requests from the IPs that trigger the grayscale rule will hit the protection rule.

Effective Mode

  • Permanently Effective (Default): The rule is always in effect when the protection template is enabled.

  • Fixed Schedule: To prevent attackers from identifying and bypassing updated protection rules before an event, new rules are typically enabled at the start of the event. You can set the rule to take effect for a specific period in a specific time zone.

  • Recurring Schedule: Suitable for recurring events where stricter protection rules are needed during the event and looser rules are used at other times. You can also deploy new rules during low-traffic periods to verify their effectiveness and check for false positives. You can set the rule to take effect for a specific period each day in a specific time zone.