The principal goal of software security is to design software to be resistant to malicious cyber-attacks. If software security is not handled at every stage of the life cycle of software, many businesses, organizations, and even governments are exploited due to security vulnerabilities in their software systems. Analyzing the source code in developing secure software systems is an important step. Identifying the insecure source code parts helps to reduce or remove vulnerabilities in software. The C programming language is one of the most-commonly used programming languages, but it is insecure and vulnerable to code injection attacks. C programming language's most common security vulnerabilities are buffer overflows, input validations, and resource management errors. In this paper, a software tool called SecureC has been developed, which can detect insecure source codes in the C programming language. Firstly, the new source code templates using the Regular expressions (Regex) were extracted, then the source codes in the dataset by using these new templates were marked. Then, it has been focused on finding whether there is any performance difference between the marked dataset and the unmarked dataset while detecting insecure parts from the source code that has hundreds of thousands of lines. As a result, the usage of a marked dataset ensured the performance improvement in both binary and multiple classification.