Trust & governance
Trust & Governance
This is a control model for responsible intelligence — not a badge wall. Every control below carries an honest status. We describe only what is actually in place, and we state plainly what is not yet earned.
Control model
The control ledger
Six governance layers, from human accountability down to auditable evidence. Each layer records its controls with a status you can hold us to.
Status key
- Implemented
In place today.
- Engagement-dependent
Available within an engagement's scope or plan.
- Planned
On the roadmap, not yet in place.
- Not yet certified
Practiced, but not independently certified.
- Not publicly disclosed
Shared under NDA during due diligence, not published here.
Layer 01
Human accountability & oversight
A person stays accountable for consequential decisions. Automation never becomes the final authority.
- ImplementedNamed accountability
- Every engagement assigns a human owner accountable for consequential decisions.
- ImplementedAutonomy limits
- Automated systems act within defined boundaries; actions with material impact require human review before execution.
- ImplementedReview gates
- Delivery advances through governed gates — nothing reaches production without a cleared review.
Layer 02
Data boundaries & confidentiality
Where client and customer data may flow — and where it may not.
- ImplementedPayment data isolation
- Payment data is processed by Stripe (PCI DSS Level 1) and is never stored on our servers.
- ImplementedRetention windows
- Analytics events 400 days, audit logs 2 years, sent notification bodies 30 days.
- ImplementedRight to erasure
- Account deletion from dashboard settings cancels Stripe and cascades a database purge.
- ImplementedCookie consent
- Managed through a dynamic consent banner under the visitor’s control.
- Engagement-dependentClient confidentiality
- Client material is handled under the confidentiality terms of each engagement.
- Not publicly disclosedInfrastructure topology
- Detailed infrastructure and vendor topology is shared under NDA during due diligence — not published on this page.
Layer 03
Security posture
The technical controls that protect data in transit, at rest, and at the edge.
- ImplementedTransport encryption
- TLS 1.2+ across all traffic, enforced by HSTS with preload.
- ImplementedCredential hashing
- Passwords are handled by Clerk (Argon2id); API keys are stored as SHA-256 hashes and the full secret is shown exactly once.
- ImplementedSigned webhooks
- Every webhook is verified via HMAC-SHA256 (Svix for Clerk, Stripe signatures, Sanity HMAC) before any payload is parsed.
- ImplementedHardened headers
- HSTS preload, X-Frame-Options, a strict CSP, Referrer-Policy, and Permissions-Policy.
- ImplementedValidated configuration
- Secret configuration is Zod-validated at boot before any service starts; distributed rate limiting runs via Upstash Redis.
Layer 04
Identity & authorization
Who may act — and the limits of what each role may do.
- ImplementedAuthentication
- Managed by Clerk with MFA and passkeys.
- Engagement-dependentSingle sign-on (SSO)
- Available on Clerk's Organizations plan, scoped to the engagement.
- ImplementedRequest integrity
- CSRF protection via a double-submit cookie plus an HMAC token on every mutating request.
- ImplementedRole-scoped access
- Team-scoped RBAC (OWNER → ADMIN → EDITOR → MEMBER → VIEWER), enforced by server-side session gates.
Layer 05
Validation & independent review
How the system is tested before and after it ships — and what has not yet been independently verified.
- ImplementedContinuous integration
- Every pull request runs typecheck, lint, unit tests, and a full build.
- ImplementedPerformance & accessibility budgets
- Lighthouse CI enforces measurable budgets; Playwright covers critical user journeys end to end.
- PlannedThird-party penetration test
- Planned before production launch. Not yet performed.
- Not yet certifiedSOC 2 Type II
- On the roadmap. Not yet achieved.
Layer 06
Auditability & evidence
What is recorded, how issues are traced, and the honest maturity of our recovery posture.
- ImplementedAudit trail
- Authentication, billing, API keys, team operations, and webhook receipts are recorded in an audit log.
- ImplementedObservability
- Sentry captures errors when enabled; structured JSON logs run with PII redaction always on.
- ImplementedTraceable delivery evidence
- Delivery artifacts and cleared gates are retained so decisions remain traceable.
- ImplementedRegulatory alignment
- Practices are aligned with GDPR and PDPL (Saudi Arabia). Alignment is not a third-party certification.
- PlannedRecovery targets (RPO/RTO)
- Documented in the operations runbook, but not yet drilled or certified.
Disclosure integrity
What we do not yet have
- Not yet certifiedSOC 2 Type II
- On the roadmap — not yet achieved.
- PlannedFormal third-party pen test
- Planned before production launch.
- PlannedCertified RPO/RTO
- Documented in the ops runbook, but not yet drilled.
This page is updated the moment any of the above is genuinely earned. We will never claim a control we have not achieved.
Responsible disclosure
Security contact
To report a vulnerability or a security concern, contact security@jtaidigital.com.