AnantState
For the CISO and security reviewers

Read-mostly, isolated by the database, audited.

It reads events and reference data, and any write to your systems is a named, audited action. Tenant isolation is enforced by the database itself, so an application bug cannot widen what anyone can see. Where we hold no attestation, we say so.

In plain termsWhat it can see, what it can change, and who is recorded.

At a glance

  1. Read-mostly

    Events and reference data in. No background write path.

  2. Isolation in the database

    Row-level security, so the application’s failure mode is closed.

  3. Identity you already run

    Single sign-on over OIDC, provisioning over SCIM, MFA at login.

  4. Gaps listed

    A trust page that states what is not claimed.

The questions a security review starts with

  • What access does it need to our systems, and what can it write?
  • Where does it run, and where does our data live?
  • Can one tenant’s data reach another’s?
  • What happens to events it cannot map?
  • Which attestations do you hold, and which do you not?

What touches your systems

What touches your systems

  1. Your systems: Events and reference dataLeads to: Event intake (read)
  2. Event intake: Read-mostly by defaultLeads to: State and evidence; Dead-letter view (cannot be mapped)
  3. Dead-letter view: Failures kept, with the reason
  4. State and evidence: Isolated by the databaseLeads to: Named, audited action
  5. Named, audited action: The only write path. Recorded.Leads to: Your systems (explicit write)
The platform reads events and reference data from your systems, keeps what cannot be mapped in view, and writes back only through a named action that is recorded with who did it and why.

What is enforced

ControlHow it is enforced
Tenant isolationRow-level security in the database, not application filtering
AccessRole and workspace checked on every request
IdentitySingle sign-on over OIDC, directory provisioning over SCIM, MFA at login
Audit trailEvery action recorded with the acting user, scoped to the tenant
Limits and budgetsEvaluated by the platform, per domain
Model promotion gateRefuses on provenance, divergence, null-baseline and unmeasured-horizon checks

Isolation that survives a bug

Isolation is enforced by the database, not the application

  1. Tenant A usersLeads to: Application
  2. Tenant B usersLeads to: Application
  3. Application: Runs as an application role, not a superuserLeads to: Database with row-level isolation
  4. Database with row-level isolation: The control lives hereLeads to: Only Tenant A rows; Only Tenant B rows
  5. Only Tenant A rows
  6. Only Tenant B rows
Each tenant reaches the data through the same application, but the database itself decides which rows are visible, so a defect in the application cannot widen what anyone can see.

Concerns a security reviewer is likely to raise

The concernThe answer
You hold no SOC 2 or ISO 27001.Correct, and the trust page says so. Ask for current status, the penetration test date and scope, and the data processing documents; we answer the questionnaire directly.
Will it write into production?Only through explicit, named, audited actions. There is no background write path.
Where does our data live?In infrastructure you control: single-host containers, Kubernetes, or your cloud account with a managed database. There is no shared hosted estate.
What happens to bad events?They land in a dead-letter view with the reason. Nothing is silently dropped.

What we do not claim

Bring the questionnaire

Bring:

  • The questionnaire
  • Your network and identity constraints

You leave with the tenancy, audit and deployment answers, demonstrated.

Put your security reviewer in the room

Bring the questionnaire. A conversation is faster than a document.