SafeCode AI

The Problem: AI tools now let developers write code faster than ever — but more AI-generated code means more potential vulnerabilities, produced at a speed no manual security review can keep up with. Our Solution: SafeCode AI is a Spring Boot platform that takes any GitHub repository URL and runs a full AI Security Audit using IBM Bob 2.0. It analyzes five OWASP ASVS categories: access control, input validation, API security, secrets handling, and security configuration. The result is a Security Score (0-100) and a structured list of findings, each with severity, file location, line number, issue description, and a recommended fix. What makes it different: Beyond detection, SafeCode AI offers AI Fix — a one-click button that uses IBM Bob to automatically patch the vulnerable code, re-verify the fix, and instantly recalculate the security score. We demonstrate this end-to-end on hardcoded secrets: the platform clones the repo, identifies the exposed secret, applies a corrected version of the file, confirms the fix resolved the issue, and updates the score live. Target users: Development teams and solo developers who ship AI-assisted code quickly and need a fast, automated safety net before deployment — without slowing down their workflow. How it works: repoUrl → clone repository → security scan → AI analysis → automatic fix → re-verification → updated score, all built on a Spring Boot backend and Angular frontend, persisted in PostgreSQL for audit history. SafeCode AI proves that speed and security don't have to be a trade-off when AI is used correctly — for both writing code and reviewing it.