Notes from inside the code.
How application security actually fails, how to check your own app, and how AIpento decides what counts as a finding. Written by the people who build it.
Why we run your tests three times before we accept a fix
A test suite that passes once proves less than it looks, and one that passes twice can still be flaky. Three clean runs are what separate a fix from a lucky run.
6 October 2026 · 2 min readBefore you launch an app built with AI tools: nine checks
Lovable, Cursor, Replit, Bolt and Claude Code make building fast. These are the nine problems we see most often in what they produce, and how to check for each one yourself.
6 October 2026 · 2 min readReproduce or withdraw: what counts as a finding
A security report full of maybes costs more than it saves. Inside AIpento, a finding that cannot be proven again does not reach you.
6 October 2026 · 2 min readSAST, DAST, pentest, AI pentest: what each one catches
Four names for testing an application, and they find different things. A plain guide to what each does, what it misses, and where AIpento fits.
6 October 2026 · 3 min read