VulnShield
Catch code vulnerabilities before they ship.
8 followers
Catch code vulnerabilities before they ship.
8 followers
VulnShield is an AI-powered, agentic security scanner for your codebase. It automatically detects SQL injection, XSS, hardcoded secrets, and weak cryptography — before they reach production. No manual audits, no security expertise required. Just connect your repo and get clear, prioritized findings with fix guidance. Built for small dev teams who need enterprise-grade code security without the overhead. VulnShield is live today.

How does VulnShield handle false positives, especially for things like weak cryptography where the context matters a lot?
@burhanmpux Hey Burhan Right now, it doesn't — by design. Findings come from a deterministic regex engine, and the AI layer is only allowed to explain findings, never add, drop, or re-score them (this is verified mechanically). So a false positive like md5(cacheKey) gets flagged the same as md5(password), since the rule can't yet read context. But I am working on it to fix this issue Hope so u enjoyed rest of the product :)
How does it handle false positives on things like test fixtures or seed data that might contain intentional hardcoded secrets? That tends to be a pain point with most scanners I have tried.
@merdavudoljrdf Hey Omer! It doesn't distinguish these today. The SEC-SECRET rule is a regex pattern that matches key/token-like strings on any line of code, regardless of file path or context. It has no awareness of whether a match sits in src/config.js or tests/fixtures/seed-data.js — an intentional test API key is flagged identically to a real leaked credential, at the same Critical severity.
Love that the findings come with prioritized fix guidance rather than just dumping a wall of CVEs on you. That kind of opinionated triage shows real understanding of how small teams actually triage security work.
@cemile19671 So glad to hear you're enjoying the product!
The sign-in was painless and the scan finished way faster than I expected on a messy Node repo. Loved how the findings actually ranked by real risk instead of just dumping everything on me at once.
@tugayabaylungn Thanks for the kind words — thrilled you like it!
The repo-to-scan flow looks really clean. Love that findings come out prioritized with actual fix steps instead of a wall of CVEs to triage.
@esraderebayvda Really happy the product is working well for you!