Security and Governance

How authentication, tenant scoping, roles, automation controls, and auditability are handled

This is not a formal security whitepaper. It is a practical guide to the security and governance posture that already exists in the product.

Identity And Session Security

The platform protects product surfaces with authenticated sessions.

Important behaviors:

signed-in state is required for protected workflows
the current organization is stored as active session context
account-level and organization-level access are separated in the product model
optional 2FA flows exist in the authentication layer

This reduces ambiguity about who is acting and in which tenant context they are acting.

Tenant Isolation

Organization data is intended to be accessed in an organization-scoped way.

In practice, the product enforces this through:

active organization checks
membership validation before organization-scoped data is returned
organization-aware server procedures
organization filters in database-backed product workflows

This is one of the most important architectural guarantees in the product.

Role Model

Obelisk uses two layers of authorization:

Platform Roles

user
admin

Platform roles control access to platform-wide administration.

Organization Roles

owner
admin
member

Organization roles control tenant-local operations such as:

team management
billing actions
invitation flows
API and webhook controls
destructive organization changes

Automation Security

The product exposes automation, but not as an ungoverned free-for-all.

Key controls include:

scoped API keys
hashed API key storage
revocation support
optional expiration timestamps
public API access through bearer credentials
outbound webhooks with delivery metadata and optional signing secret

This is the right tradeoff for teams that need machine access without abandoning governance.

Webhook Trust Model

Outbound webhook delivery includes identifying metadata such as:

event type
delivery ID
timestamp
signature

If a signing secret is configured, receivers can validate payload authenticity and freshness before accepting a delivery.

Auditability

Security is stronger when actions are inspectable.

The platform supports that through:

provisioning audit logs
activity feeds
recent API key visibility
alert delivery history
platform alert audit trails
request-aware logging context in server procedures

This makes incident review and ownership handoff materially easier.

Governance Model

The product is designed around a clear division:

organization workspaces for tenant operations
admin surfaces for platform governance

That split matters because it keeps tenant work, platform intervention, and operational escalation from collapsing into one ambiguous interface.

What Good Governance Looks Like

In a healthy deployment, teams typically:

keep platform admin access limited
assign organization admin roles intentionally
rotate automation credentials instead of reusing them indefinitely
separate production and non-production workflows
use alerts and audit logs instead of relying on memory or chat history