A security scanner built specifically for code written by AI tools like Cursor, Copilot, Claude Code, Lovable, and Bolt. Rather than a general-purpose SAST tool, it narrowly targets the vulnerability classes AI repeatedly introduces: hardcoded credentials, missing access control, injection, path traversal, unverified JWTs, weak randomness, disabled TLS, and prototype pollution. It works in three modes: paste a public GitHub repo URL for a no-account scan against 15 rules; install the GitHub App or Action for inline PR comments (Pro adds fail-on-high to block high-severity merges); and enable live app probing, which requests routes like /dashboard, /admin, and /api/users to check whether unauthenticated access exposes private data. Fixes are behavior-preserving—Prbl baselines function behavior, applies the minimal fix, then re-scans to verify nothing broke. Its published research scanned 2,148 AI-built repos and 4,740 live apps, finding 25.7% of repos carried at least one high-severity issue and 5.5% of live apps exposed private data without login; the dataset is public on GitHub under CC BY 4.0. Pricing is freemium, with Free covering public-repo scanning and non-blocking Action comments, and Pro adding private-repo dashboards, blocking merges, and audit logs. Launched on Product Hunt in July 2026.
Pros & Cons:
Pros: Narrow scope is a deliberate design choice—by targeting the vulnerabilities AI tools actually repeat, it avoids the false-positive noise that makes general SAST tools unusable on AI-generated code (the founder cites a test where 23 of 24 high-severity alerts were false positives); behavior-preserving fixes with a re-scan verification step are more trustworthy than “apply patch and hope”; live app probing checks unauthenticated access to real routes, catching exposed data that static analysis alone misses; public research dataset (2,148 repos, 4,740 apps, CC BY 4.0) makes its claims checkable rather than marketing assertions; Free tier supports public repo scanning with no account required; GitHub App and Action integrate into existing PR workflows.
Cons: 15 rules only, by design—it is not a replacement for CodeQL, Dependabot, or Semgrep, and users expecting full SAST coverage will be disappointed; private-repo dashboards, merge blocking, and audit logs require Pro, and Pro pricing is not clearly published in surfaced documentation; independent user reviews are thin—Product Hunt and Future Tools listings exist, but no G2/Capterra/Trustpilot ratings surfaced; company entity, funding, and team size are not disclosed; name collision with Proto Labs (ticker PRLB) means search results mix in unrelated manufacturing stock news.
Who is it for?:Developers shipping code generated by Cursor, Copilot, Claude Code, Lovable, or Bolt who want a security check before merge; solo builders and indie hackers who need a free public-repo scan without setting up a full SAST pipeline; teams that want PR-level inline security comments with an optional merge-blocking gate; security-conscious developers who prefer a narrow, low-noise tool over a broad scanner full of false positives.