Sentrint
Scans AI-generated code for hardcoded secrets and exploitable vulnerabilities, with AI-filtered false positives
🔗 Visit Sentrint
Description
AI coding assistants write a lot of code very fast, and fast code sometimes ships a hardcoded API key or a sloppy input-handling bug nobody caught in review. Sentrint scans repositories specifically for the kinds of mistakes AI coding tools tend to introduce, across whichever assistant you used to write the code.
Sentrint scans GitHub repositories for vulnerabilities introduced by AI coding assistants — Claude, Copilot, Cursor, Gemini and 12 others. It runs a three-pass analysis: a security engine that scans for hardcoded secrets and exploitable code paths, an AI-powered filtering pass to cut false positives, and generation of platform-specific fix prompts you can hand back to your AI assistant. Scans run on ephemeral Google Cloud containers that are destroyed after completion, with read-only GitHub access, live API-key validation, and a portfolio dashboard with README security-grade badges.
💬 Our review
The short version: purpose-building a scanner around the specific mistakes AI coding assistants make — and generating a fix prompt tailored to the tool that wrote the bug — is a smart, current-moment angle on code security scanning.
Compared to general static analyzers like Semgrep or dependency scanners like Snyk, Sentrint's pitch is narrower but more actionable for this specific failure mode: instead of a generic finding, you get a fix prompt written for the exact AI platform (Claude, Copilot, etc.) that introduced the issue. It's early (beta v0.9) and the free tier is limited to one scan per month, with the useful tier (36 scans/month) at $12.35/month during the current launch discount. Worth it if you ship a lot of AI-generated code and want a security pass tailored to that workflow; a general scanner like Semgrep is more mature if you want broader, tool-agnostic coverage.
📊 Global score
🤖 AI-enriched data
Gratuit : 1 scan/mois. Recharge : 3,90 $ pour 6 scans. Founder Plan : 12,35 $/mois ou 123,50 $/an (tarif de lancement, 36 scans/mois, export CSV/JSON, SBOM).
Pros
Palier gratuit sans carte bancaire, avec rapport complet et suggestions de correction par IA
Compatible avec 16+ plateformes IA, avec prompts de correction adaptés à chaque outil
Filtrage des faux positifs par une couche IA d'analyse d'exploitabilité
Dashboard portefeuille avec notes de sécurité en temps réel et badges README
Cons
Palier gratuit limité à 1 scan par mois
Version bêta (v0.9), l'API et les fonctionnalités peuvent encore changer
Nécessite un accès en lecture au dépôt GitHub (dépôts scannés puis supprimés)