FAQ
Questions, answered
Everything people ask before they scan their first app.
Getting Started
It helps vibe coders review security risks, accessibility barriers, technical SEO/AEO readiness and browser bundles on public pages. Findings include evidence and fix guidance. The anonymous scan covers a limited security scope; a free account adds all four tools without domain verification.
Yes. Submit the public deployed URL. The checks can identify supported exposure and configuration patterns, regardless of the builder. Platform-specific guides help you test data permissions, secrets and deployment settings separately.
The instant scan on the homepage is a free, lightweight security pass with no account — headers, TLS, cookies, and secrets visible in public HTML/JS only. The result includes all recorded findings, individual check outcomes, limitations, timing, request counts, severity filters and a copyable summary. Unsaved results expire after 30 days. It does not probe .env/.git paths, query Supabase RLS, or run active checks. Create a free account, then run all four tools (Security, Accessibility, SEO & AEO, Bundle & Tech).
Scanning and Security
Checks are read-only by default. Optional active checks on Pro and Business send crafted requests after hostname verification and are designed to avoid destructive testing, credential guessing and denial of service. Every target behaves differently, so use active testing only on systems you are authorized to test. If bot protection blocks a scan, configure an allow rule and retry.
SEO, accessibility, bundle and passive security audits work on website URLs without verification. Active security probes require ownership proof via an HTML meta tag or a .well-known verification file. The free homepage instant scan stays a lightweight public check (headers, TLS, cookies, secrets in shipped JS) with no signup.
Every account scan checks for a bot-challenge page first. If one is still on, the scan stops before Security, Accessibility, SEO, or Bundle run. Use a scoped allow rule for the scanner if appropriate for your host. The documented User-Agents are: VibeShield-Verification/1.0 (+https://usevibeshield.com), VibeShield-Scanner/0.1 (+passive security check; https://usevibeshield.com), and VibeShield-SiteAudit/0.1 (+non-destructive accessibility/SEO/bundle scan; https://usevibeshield.com). A User-Agent is not proof of identity; keep the rest of your host protections in place.
Alongside on-page SEO, we inspect technical readiness signals such as structured data, AI crawler rules, thin text and snippet directives. llms.txt is optional; its absence is not proof that AI search cannot use your site. We do not measure rankings, brand mentions or citations in answer engines.
When a page ships a Supabase URL and public key, VibeShield can check discovered tables using HEAD requests for row counts, without fetching row data. A readable table shows anonymous visibility; it does not prove RLS is disabled. Review whether a policy intentionally permits those rows and whether any sensitive data should be restricted.
Some checks finished, while others were blocked, timed out, exceeded a budget or need manual review. Recorded findings still deserve attention. Unfinished checks do not count as passes, affected scores stay unavailable, and comparisons must not label untested issues resolved. A completed tool can still show its own score.
A tool can be incomplete when a page is blocked, times out or exceeds a scan budget. The report lists coverage notes instead of treating unfinished checks as passes. Accessibility manual-review notes identify checks the automated engine could not decide; they need inspection of the affected page.
No. Bundle & Tech observes LCP, CLS, FCP and TTFB during the initial desktop Chromium page load. These are lab measurements, separate from the bundle grade. We do not fetch PageSpeed Insights results or provide INP, a Lighthouse score or real-user Core Web Vitals.
No scanner can promise that. A clean report means the completed checks found no issue in their scanned scope at that moment. Blocked or skipped checks remain coverage gaps. Keep scanning after changes — every new prompt can change your app.
AI Fix Prompts
Anonymous reports show evidence and fix guidance. Account reports also provide a "Copy AI fix prompt" button. Review and test the proposed patch, deploy, then re-scan with comparable coverage. Only completed compatible checks can establish resolution. Pro adds a fix-everything prompt.
Pricing and Limits
The default Basic plan is $0 with 10 scans/month, 1 saved domain and 1 page per scan. Pro is $19/month with 50 scans, 5 domains and up to 3 pages; it adds verified active checks, weekly monitoring and exports. Business is $39/month with 150 scans, 15 domains and up to 5 pages; it adds daily/deploy scans, GitHub and team alerts. The pricing page shows current configured allowances and included features.
In Settings → Plan & billing, use Manage subscription or Cancel subscription (our payment provider's customer portal), or the link in any payment receipt. Your plan stays on until the end of the month you've paid for, then your account moves to Free. Your scan history stays.
Yes. Pro and Business come with a 14-day money-back guarantee — email contact@usevibeshield.com within 14 days of your purchase and we'll refund you in full, no reason needed. Refunds are issued by our payment provider (the merchant of record) to your original payment method, usually within 5–10 business days, and your account returns to Free.
Privacy and Data
Website audits use the URL and do not require the target app's login or database credentials. GitHub integration is separate: a connected token provides access to private repositories and enables pull-request comments. That token is encrypted at rest.
Integrations
Business repository scans combine source/configuration analysis, Gitleaks potential-secret detection and OSV.dev known npm advisories. Findings show files, lines and redacted secret evidence. We inspect fetched files at the selected commit, not Git history. Dependency coverage currently uses the root package.json and package-lock.json. A token raises the file cap from 40 to 2,000; file-size, total-size and service limits still apply. Dependency checks need supported exact installed versions; missing lockfiles, skipped versions and advisory-service failures are reported as limitations. A repository with no supported files is Not verified, not a clean scan.
Still wondering? Run a free scan and explore the results.