Security Overview
Sovereign is designed for deployment in security-conscious environments — from commercial SaaS to defense and aerospace. This page summarizes the platform's security posture. See the sub-pages for details on Data Protection and Authorization & Compliance.
Authentication
All API access requires authentication. The platform supports the following modes today:
- Bypass — no end-user authentication enforced. Intended for air-gap, single-operator, or pre-IdP production deployments. Not appropriate for multi-user deployments. Service-to-service calls remain authenticated via service keys.
- Local — built-in admin account with password authentication.
- OIDC — federated authentication via your corporate identity provider (Okta, Azure AD, Google Workspace, Keycloak, and others)
For commercial multi-user deployments, OIDC mode is recommended. See Authentication Setup for configuration details.
Federation for Air-Gapped and High-Side Environments [v1]
Classified and other air-gapped facilities typically cannot reach the Internet-discovery endpoints that OIDC deployments rely on. The platform's federation support targets these environments:
- SAML federation
[v1] — assertion-based federation with on-premises identity providers (ADFS, Keycloak, Shibboleth). No Internet egress required; metadata exchange can be performed manually.
- LDAP / Active Directory
[v1] — on-premises directory binding, including bridge configurations where AD fronts SAML for downstream federation.
- PIV / CAC smartcard
[v1] — X.509 client-certificate authentication for U.S. federal civilian (PIV) and DoD (CAC) environments. Certificate identity is extracted at TLS termination on the reverse proxy and mapped to platform users; OCSP/CRL revocation checking is enforced per session.
These modes will ship with v1. See Authorization & Compliance for advanced authorization features also committed to v1.
Network Isolation
The deployment architecture provides multiple layers of network isolation:
- Auth database — runs on a dedicated database instance with no external port exposure. Only the Auth Service can access it over an isolated network.
- Application database — accessible to application services but isolated from the auth network.
- Service-to-service communication — authenticated via service keys. Internal endpoints are not exposed through the reverse proxy.
Tenant Data Isolation
All data is scoped to tenants at every layer:
- Workflow definitions, executions, configurations, and schedules are tenant-scoped
- The orchestrator validates tenant context before dispatching any steps
- Cross-tenant data access is prevented at the query level
- Connection ownership is verified on every credential operation
Rate Limiting
Authentication endpoints are protected with rate limiting to prevent brute-force attacks:
- Strict limits on credential endpoints (login, token exchange)
- Relaxed limits on metadata endpoints (JWKS, discovery)
- Configurable thresholds per environment
Production Security Checklist
- Change all default passwords and service keys
- Enable TLS on all public-facing endpoints
- Configure OIDC mode for authentication (not Bypass or Local)
- Use a dedicated key management service for encryption keys
- Restrict database network access (no exposed ports)
- Configure TLS for the message queue
- Enable audit logging (requires audit capability in your license)
- Implement regular backup procedures
- Monitor health check endpoints
- Rotate service keys and encryption keys on a regular schedule
Next Steps
- Data Protection — credential encryption, message queue security, license enforcement
- Authorization & Compliance — RBAC, audit logging, and the defense-grade authorization system coming in beta
- Authentication Setup — configure your authentication mode
- Deployment Guide — production deployment with security hardening