SQL injection (SQLI) remains one of the most dangerous threats to database-driven applications. Existing SQL injection prevention techniques suffer from limitations such as high false positives, processing overhead, and weak runtime validation. This study proposes an N-GRAM-driven Transact-SQL (T-SQL) security framework for detecting and preventing SQL injection attacks within Microsoft SQL Server stored procedures. The framework integrates N-GRAM character analysis with T-SQL validation functions to analyse user inputs before query execution. Malicious query patterns are detected through character sequence matching against predefined SQL injection signatures.
Download