LLM powered Git-Scan

Description

llm-powered-git-scan is an AI-assisted Git repository scanner that analyzes commit diffs for potential secrets, API keys, or other sensitive data using the OpenAI API. The tool clones or scans a given Git repository, iterates through recent commits, and sends code diffs to an LLM (like GPT-4o-mini or GPT-5-mini) for intelligent, context-aware security analysis. It detects sensitive code patterns that traditional regex scanners often miss — such as access tokens, credentials, or embedded API keys — and generates a structured JSON report for auditing and review.

Core Features:
- AI-powered detection of sensitive data and secrets.
- Uses OpenAI GPT models (no local inference required).
- Automatically installs missing dependencies.
- Outputs detailed JSON reports to your Documents folder.
- Works on Windows, macOS, and Linux.
- Scans both local and remote (GitHub) repositories.

Use case example:
> You provide a GitHub repo URL or local path. The program clones or reads it, analyzes recent commits using an LLM, and saves a clean JSON report showing which code lines might expose secrets.

Extra notes:
- Project completed within 5 days (120 hours).
- Developed as part of a JetBrains Internship assignment — built to demonstrate integrating LLMs into a cybersecurity automation workflow using Python.

> Attention — More images coming soon!