Security

Effective date: 25 September 2026

This page describes controls that are actually in place on the platform. OKFLW does not currently hold any external compliance certification.

Workspace isolation

Each workspace has its own data boundary, and the tenant and workspace are determined on the server from the session, never from the browser.

Server and database authorization

Every request passes a permission check on the server, and sensitive operations are enforced inside the database itself.

Roles and permissions

Team roles (such as owner, admin, operator and viewer) determine what each member can see or change.

Human approval

Sensitive actions pause for human approval before they run.

Data-loss prevention and outbound controls

Outgoing content is checked before it leaves to prevent leaking sensitive data.

Runtime limits

Runtime quotas and plan limits prevent unexpected consumption in each workspace.

Abuse controls

Security signal detectors watch for suspicious patterns and open security incidents for review.

Kill switch

Specific capabilities can be stopped immediately for a company, workspace, agent or channel when needed.

Tamper-evident audit log

Important events are recorded in a hash-chained log that is sealed periodically, so later changes can be detected.

Monitoring and incidents

Service-level objectives are monitored continuously, and incidents are opened and followed until closed. Current status is on the Status page.

Backups and recovery

Database backups are taken daily and intraday on the platform infrastructure, and restore procedures are tested with regular drills.

Credential protection

Channel credentials are stored encrypted and never displayed after they are saved, passwords are stored hashed, and the site is served over HTTPS only.

Reporting a vulnerability

A dedicated public security-reporting channel is not currently configured. It will be listed on this page and on the Contact page once the platform owner configures it.

Last updated: 2026-09-24