VibeGuard
A security linter built specifically to catch the vulnerability patterns AI coding assistants like Copilot and Cursor repeatedly produce.
🔗 Visit VibeGuard
Description
AI coding assistants are great at getting something working fast, but they also have well-known bad habits, hardcoding a secret here, skipping input sanitization there, that a rushed developer might not catch in review. VibeGuard is a linter built specifically to catch those AI-specific mistakes, grading your codebase and explaining each problem in plain language rather than cryptic security jargon.
VibeGuard scans code for patterns AI assistants like Copilot, Cursor, and ChatGPT are known to repeat, including SQL injection, hardcoded secrets, JWT bypass vulnerabilities, and command injection, using 47 rules across 6 categories. It outputs a letter grade (A-F) for the codebase along with plain-English remediation guidance, integrates into CI/CD with configurable exit codes, and requires zero configuration to run.
💬 Our review
The short version: VibeGuard fills a real, narrow niche, catching the specific bad habits AI coding tools produce, and its zero-config setup and plain-English grading make it easy to try, but it's a two-star project competing against much more established general-purpose scanners.
Bandit and Semgrep already catch most of the same vulnerability classes (SQL injection, hardcoded secrets, command injection) with far larger, battle-tested rule sets and broader language support; VibeGuard's pitch is being tuned specifically to AI-generated code patterns and giving letter-grade, plain-English output non-security-focused developers can actually act on. If you already run Semgrep or Bandit in CI, VibeGuard is a lightweight, free complementary layer rather than a replacement; if you're a team leaning heavily on AI coding tools with no security linting yet, it's an easy, zero-config starting point.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Gratuit, open-source
Pros
ciblé spécifiquement sur les erreurs des assistants IA (Copilot, Cursor, ChatGPT)
note en lettre (A-F) et explications en langage clair
zéro configuration pour démarrer
gratuit, open-source
intégration CI/CD avec codes de sortie configurables
Cons
projet très jeune (2 stars), rule-set moins large que Bandit/Semgrep
couvre 47 règles vs des centaines chez les scanners établis
complète un scanner généraliste plutôt que de le remplacer