SanityCode runs a fixed set of rules on your code.
plain-language checks for AI-written code
What it does:
SanityCode compares AI-written code against a fixed set of risk types. Any part that matches a rule is shown on screen in fixed wording. Issues outside those rules do not appear.
How to use it:
- Paste your code.
- Run the check.
- View the result.
Check types:
- Information exposure: cases where information that should stay hidden becomes visible.
- Unverified external input: cases where content from outside is used as-is without checking.
- Values only the server should decide: cases where the browser decides a value that should be decided by the server.
Your code is processed only on this device and is not sent to a server or anywhere else. The check does not call a model.