Secure Code Analyzer

An AI-powered tool that analyses source code and flags security vulnerabilities automatically — the kind of issues that slip through code review and end up in CVE databases.

What it detects

SQL / command injection patterns
Hardcoded secrets and credentials
Insecure dependency usage
Unsafe deserialization
Common OWASP Top 10 patterns

How it works

The model is trained/fine-tuned to understand code context — not just pattern matching, but understanding intent and data flow to reduce false positives. Input is raw source code; output is a structured vulnerability report with severity levels and remediation hints.

Why I built it

Security is often an afterthought. This project was an experiment in applying ML to a problem that's usually solved with static analysis rules — and seeing whether a language model can catch things that rule-based scanners miss.