Datenschutz
A scanner that flags risky code patterns and explains how they might be fixed.
- Status
- Security-analysis prototype
- Year
- 2025
- Role
- Solo
- Area
- Developer tools
In plain English
Security mistakes in code, like leaked passwords or unsafe database queries, are easy to make and costly to miss. Datenschutz scans code, flags patterns that may be unsafe and explains possible fixes. It gives a human reviewer a head start.
How it works
- Point the scanner at a file or folder of code.
- Rule-based detectors (readable pattern checks) flag possible issues such as SQL injection or hardcoded secrets.
- Each finding is classified and mapped to standards like the OWASP Top 10 (a common list of web risks).
- A language model adds explanations and suggested fixes, shown in a web interface.
What I did
I built it solo, combining my own detection rules with existing security standards and model providers.
What I learned
An automated scanner works best as the starting point for a human review. Keeping the rules readable makes every finding easy to check.