LogInjector: Detecting Web Application Log Injection Vulnerabilities

Web applications widely use the logging functionality, but improper handling can bring serious security threats. An attacker trigger execution of malicious data by writing to web application logs and then accessing view–logs interface, resulting in a vulnerability log injection. However, detecting this type requires automatic discovery log-injectable interfaces interfaces, which is difficult. In addition, bypasssing application-specific input-filtering checks write an effective payload also challenging. This paper proposes LogInjector, efficient injection detection method. First, it obtains storage form location finds through extended dynamic crawler. Second, automatically identifies interfaces. Finally, LogInjector utilizes testing approach based on feedback-guided mutation detect vulnerabilities. To verify effectiveness we test 14 popular real-world cases compare with Black Widow, state-of-the-art scanner. detects 16 vulnerabilities, including 6 zero-day while Widow only three demonstrating practice.