Skip to content

Security: DevStudio-AI/VybeGrid-DNS

Security

SECURITY.md

VybeGrid DNS — Security Model

How VybeGrid protects provider credentials, secures operations, and ensures compliance


Overview

VybeGrid manages access to critical infrastructure — DNS records and CDN configurations that control where internet traffic flows. The security model reflects this responsibility, with defense-in-depth protections at every layer.


Encrypted Secrets Vault

The most critical security component. All provider API keys and credentials are encrypted at rest using AES-256-GCM.

How It Works

User enters API key
    │
    ▼
┌────────────────────────────────────────┐
│  1. Key derivation from master secret   │
│     (PBKDF2 / HKDF)                    │
├────────────────────────────────────────┤
│  2. AES-256-GCM encryption             │
│     • Authenticated encryption          │
│     • Unique nonce per encryption       │
│     • Integrity tag prevents tampering  │
├────────────────────────────────────────┤
│  3. Store encrypted blob in PostgreSQL  │
│     • Never stored in plaintext         │
│     • Per-provider isolation            │
├────────────────────────────────────────┤
│  4. Decrypt only at point of use        │
│     • Held in memory briefly            │
│     • Wiped after API call              │
└────────────────────────────────────────┘

Why AES-256-GCM

Property Benefit
256-bit key Computationally infeasible to brute-force
Galois/Counter Mode Authenticated encryption — detects tampering
Unique nonce Same key encrypts different data safely
AEAD Associated data verified without decryption

Authentication

JWT Token System

Feature Implementation
Token format JSON Web Tokens (python-jose)
Signing HMAC-SHA256
Expiration Short-lived access tokens + refresh rotation
Claims User ID, organization, roles, scopes

Multi-Factor Authentication

Feature Implementation
Method TOTP (RFC 6238)
Compatibility Google Authenticator, Authy, etc.
Enrollment QR code + backup codes
Enforcement Configurable per organization

Password Security

Feature Implementation
Hashing bcrypt (passlib)
Cost factor Adaptive (increases with hardware)
Salt Unique per password

Rate Limiting

Protection against API abuse and DDoS:

Endpoint Category Limit Window
Authentication 5 requests per minute
General API 100 requests per minute
Webhook events 50 requests per minute
Health check API 200 requests per minute

Implemented via SlowAPI middleware on FastAPI. Limits are per-IP with optional per-user tracking.


Signed Webhooks

When VybeGrid sends webhook notifications (failover events, alerts), they are cryptographically signed:

Webhook payload
    │
    ▼
┌─────────────────────────────────────────┐
│  1. Generate HMAC-SHA256 signature       │
│     using shared webhook secret          │
│                                          │
│  2. Add timestamp header                 │
│     (prevents replay attacks)            │
│                                          │
│  3. Include signature in headers:        │
│     X-Webhook-Signature: sha256=...      │
│     X-Webhook-Timestamp: 1710268800     │
│                                          │
│  Receiver:                               │
│  4. Verify signature matches payload     │
│  5. Check timestamp within window        │
│     (reject if > 5 minutes old)          │
└─────────────────────────────────────────┘

AI Security

When AI agents are enabled, additional security controls apply:

Control Implementation
Data sanitization API keys, IPs, and PII stripped before LLM calls
No direct execution AI recommendations require policy engine approval
Opt-in only AI features completely disabled by default
Provider isolation AI agents cannot access provider credentials directly
Audit logging All AI recommendations and decisions logged

Audit Trail

Every operation in VybeGrid produces an audit event:

Event Types Examples
Authentication Login, logout, MFA enrollment, failed attempts
Provider management Add/remove provider, credential update
DNS operations Record create/update/delete
Failover events Automatic failover, manual trigger, rollback
Configuration changes Policy updates, threshold adjustments
AI decisions Anomaly detected, recommendation issued

Properties:

  • Append-only (no modification or deletion)
  • Timestamped with actor identification
  • Filterable and searchable in dashboard
  • Exportable for compliance reporting

Access Control

Role-Based Permissions

Role Capabilities
Viewer Dashboard read, metrics, audit log view
Operator DNS operations, manual failover, health check config
Admin Provider management, policy configuration, user management
Owner All capabilities + billing + credential management

Provider-Level Permissions

Fine-grained control over which providers each user can access:

  • Read-only access to specific providers
  • Full CRUD on permitted providers only
  • Credential visibility restricted to Admin/Owner

Network Security

Control Implementation
HTTPS Required for all API communication
CORS Strict origin allowlist
Request validation Pydantic schemas on all endpoints
Input sanitization SQL injection and XSS prevention
Dependency scanning Regular security audit of packages

Compliance Readiness

Standard Status
SOC 2 Architecture supports audit, access control, encryption requirements
HIPAA Encryption at rest + in transit, access logging, role-based access
GDPR Data export capabilities, access logging
PCI-DSS Encrypted credential storage, audit trail

© 2024-2026 DevStudio AI Inc.. All rights reserved.

There aren’t any published security advisories