VibeArmor Documentation
Everything you need to scan your apps, integrate with CI/CD, and connect VibeArmor to your AI coding assistant.
Quick Start
Get your first security scan in under 3 minutes. No account required.
1. Scan your first app
Go to vibearmor.ai/scan and paste any live URL. Vercel previews, Netlify deploys, Railway apps, or any public endpoint. The scan takes about 2-3 minutes and runs 100 security checks against your app.
2. What the score means
Every scan produces a score from 0 to 100, mapped to a letter grade. The score reflects real hackability, not security hygiene trivia.
3. Understanding finding tiers
Every finding is classified into one of three tiers. This is the core of how VibeArmor differs from generic scanners: we separate what is actually exploitable from what is noise.
API Reference
Trigger scans programmatically. Available on Security Report ($499) and Pentest ($2,500) tiers, or with rate limits on free tier.
CI/CD Integration
Add a security gate to your deployment pipeline. Block deploys that fail your score threshold. Zero dependencies required.
GitHub Actions
Add this step to any workflow. It uses only curl and jq — no action marketplace dependencies, no supply chain risk.
name: Security Scan
on:
push:
branches: [main]
pull_request:
jobs:
vibearmor:
runs-on: ubuntu-latest
steps:
- name: Wait for deployment
uses: actions/github-script@v7
with:
script: |
// Wait for your preview deploy to be ready
// Replace with your deployment check logic
- name: VibeArmor Security Scan
run: |
RESULT=$(curl -s -X POST https://vibearmor.ai/api/scan \
-H "Content-Type: application/json" \
-d '{"url": "${{ env.DEPLOY_URL }}"}')
SCORE=$(echo $RESULT | jq -r '.score')
GRADE=$(echo $RESULT | jq -r '.grade')
CRITICAL=$(echo $RESULT | jq -r '.criticalCount')
echo "## Security Scan Results" >> $GITHUB_STEP_SUMMARY
echo "Score: **$SCORE** (Grade: **$GRADE**)" >> $GITHUB_STEP_SUMMARY
echo "Critical findings: **$CRITICAL**" >> $GITHUB_STEP_SUMMARY
if [ "$SCORE" -lt "70" ]; then
echo "::error::Security score $SCORE is below threshold (70)"
exit 1
fi
echo "Security scan passed with score $SCORE ($GRADE)"Vercel Deploy Hook
Automatic scans triggered on every Vercel deployment. Add your Vercel project ID in the VibeArmor dashboard and scans run on every preview and production deploy. No workflow files needed.
Tips
- Set the threshold to match your risk tolerance. 70 blocks anything below a C. 90 requires an A or better.
- For pull request checks, scan the preview URL (Vercel generates one per PR).
- Scans take 2-3 minutes. Add a timeout of
5mto your CI step to avoid hanging builds.
MCP Server (AI Assistant Integration)
Connect VibeArmor directly to your AI coding assistant. Scan URLs, check security, and get scores without leaving your editor.
Claude Code
Add the VibeArmor MCP server to your .mcp.json file:
{
"mcpServers": {
"vibearmor": {
"command": "npx",
"args": ["-y", "vibearmor-mcp"],
"env": {
"VIBEARMOR_API_URL": "https://vibearmor.ai",
"VIBEARMOR_API_KEY": "your-api-key"
}
}
}
}Cursor
Open Settings → MCP → Add Server and enter the same configuration as above.
Available tools
scan_urlRun a full 100-check security scan on any URL. Returns score, grade, findings, and fix suggestions.
check_securityFocused scan on a single security category (e.g., auth, secrets, headers). Faster than a full scan.
security_scoreQuick pass/fail score check for CI/CD gates. Returns score and grade without full finding details.
get_scan_reportFetch a previous scan result by scan ID. Use this to review historical scans.
Usage example
Once connected, you can ask your AI assistant naturally:
Scoring Model
Our scoring model is calibrated so that well-secured apps score high and genuinely hackable apps score low. Stripe scores A+. Shopify scores A.
Point deductions by tier and severity
Every scan starts at 100. Points are deducted based on the tier and severity of each failed check. Tier 3 findings are shown but never deduct points.
Score floors
Floors prevent misleadingly low scores when an app has many minor issues but nothing exploitable. They ensure the grade reflects actual hackability.
No app can score below 40, regardless of findings.
If no Tier 1 findings, the score cannot drop below 75.
If no Tier 1 or Tier 2 findings, the score stays at 90+.
Calibration reference
We calibrate our scoring against well-known production apps to ensure grades are meaningful.
Ready to scan?
Paste a URL and get your security score in 3 minutes. Free, no signup.
Scan Your App Free