Security Overview

Enver’s layered security model and threat mitigations.

Security Overview

Enver is designed with a defence-in-depth approach. Each layer independently protects secrets, so a breach of one layer doesn’t compromise your data.

Client → Authentication → Authorisation → Input Validation → Database
  ↑            ↑                ↑                 ↑              ↑
Zero-       Token scopes    Rate limiting    Zod schemas     Encrypted
Knowledge   IP binding      IP validation   Sanitise input   at rest

Layers

1. Zero-Knowledge encryption

The server stores only ciphertext. See Zero-Knowledge Architecture.

2. Token-based authentication

Every API request requires a valid env_live_... token, hashed with SHA-256 before storage. Tokens can expire and are scoped to specific operations.

3. IP binding

Tokens are tied to the IP address used at creation. A stolen token is unusable from a different machine. See IP Binding.

4. Scope enforcement

Tokens have the minimum permissions required. Read-only CI tokens cannot modify secrets.

5. Rate limiting

100 requests per minute per IP prevents brute-force attacks.

6. Input validation

Every API request is validated against strict Zod schemas — no SQL injection, no malformed payloads.

7. Audit logging

Every secret push, pull, member change, and project creation is logged with a timestamp, user ID, and action type.

Reporting a vulnerability

Email [e-mail]. We respond within 48 hours.