SQL Injection Pattern Scanner (Logs & Inputs)
Find suspicious requests in a log excerpt and see which signature each one matched.
Use the tool
Why this tool exists
Each line is checked as written and after URL-decoding (up to twice) against signatures for tautologies (' OR 1=1), UNION SELECT, stacked statements, sleep and benchmark calls, schema enumeration, error-based functions, comment terminators, command and file access, hex or CHAR() payloads and generic SQL keywords. Hits are ranked high, medium or low.
A match means the request looked like an attack, not that it worked. Check response codes and sizes, then fix the root cause with parameterised queries; filters and WAF rules are a second layer. Obfuscated or novel payloads will be missed.
No data is sent anywhere: the work happens in your browser, so you can use it offline once the page has loaded.
How to use it
- Paste log lines or input values, one per line.
- Run the scan.
- Review high-severity lines first, including the decoded view.
- Check whether the endpoint uses parameterised queries.
Frequently asked questions
Can it prove an attack succeeded?
No. It only matches signatures; look at the response and database logs.
Why are there false positives?
Legitimate text such as "select a size from the list" can match keyword rules, which is why those are marked low.
What fixes SQL injection?
Prepared statements with bound parameters, least-privilege database accounts and input validation.
Is my data sent anywhere?
No. Everything is processed in your browser, and the page makes no network requests with your input.