SUMMARY
Among the various types of software vulnerabilities, command injection is the most common type of threat in web applications. In command injection, SQL injection type of attacks areextremely prevalent, and ranked as the second most common form of attack on web. SQL injection attacks involve the construction of application’s input data that will result in the execution of malicious SQL statements. Most of the SQL injection detection techniques involve the code to be written along with the actual scripting code. These techniques do not detect errors in SQL statements. Hence, this paper proposes a mechanism to identify invalid SQL statements, to analyze the query for invalid non SQL key words, and to customize the captured errors. This mechanism is different from others by means of separation of the main scripting code and SQL injection code.