Technical Specifications
Framework-aware security scanning that starts with JavaScript/TypeScript. It runs where you work, speaks your language, and never uploads your source code.
We built VibeSec because code is shipping faster than ever, especially with AI-assisted development and automated review workflows. Our mission is to catch framework-specific security problems early, so teams can move fast without breaking safety.
The CLI Engine
A local-first static analysis CLI built on Node.js. Installs quickly and runs where you work.
Installation
npm install -g @reliabilityworks/vibesecScan Command
Scan the current directory or a specific path.
vibesec scan .Output Formats
Native support for multiple output formats for easy integration.
$ vibesec scan .
✔ Scanning 142 files...
✔ Found 0 vulnerabilities.
# Export to SARIF for GitHub
$ vibesec scan . --output sarif --out-file results.sarifFramework Aware
Automatically detects and adapts to your project structure.
VibeSec understands the nuances of modern JS frameworks. It knows that a file inpages/api is different from a component in src/components.
Supported Detectors
CI/CD Integration
Gate your builds with zero friction. First-class GitHub Actions support.
Block PRs based on severity thresholds using the --fail-on flag or use our dedicated GitHub Action.
- uses: Reliability-Works/vibesec@v1
with:
path: .
framework: auto
output: sarif
out-file: vibesec.sarif
fail-on: high
- uses: github/codeql-action/upload-sarif@v3
with:
sarif_file: vibesec.sarifConfiguration & Baselines
Declarative configuration that lives in your repo.
Config File
Configure ignores in .vibesec.yaml (or .vibesec.yml).
Baselines
Suppress existing issues to focus on new code.
vibesec scan . --write-baselineCreates .vibesec.baseline.yaml
Custom Rules
Write your own security rules using simple YAML or JSON.
Drop rule files into .vibesec/rules or specify a custom directory with --rules-dir.
- id: no-eval
severity: critical
title: "Avoid eval"
matcher:
type: regex
fileGlobs:
- "**/*.{js,jsx,ts,tsx}"
pattern: "eval\("
message: "Avoid using eval()"Safety Model
Designed for trust and privacy.
- Passive Scanning: Never executes your repository code.
- Local First: Analysis happens on your machine or CI runner. Source code never leaves your environment.
- Zero-Config: Safe defaults out of the box.
Coming Soon
We are building a hosted platform to complement the CLI, plus more languages & frameworks.
Hosted Dashboard
Centralized view of security posture across all repos.
SSO & IAM
Enterprise-grade access control and team management.
Compliance Reports
Audit-friendly reports and SOC2-style evidence bundles.