Skip to main content
Open navigation
Layered stone and steel walls defining a controlled passage with a warm-lit threshold

Security / The operating condition

Trust needs
clear boundaries.

Privacy, evidence integrity, and accountable decisions belong in the foundation—not at the edge of the product.

01 / Least privilege02 / Evidence provenance03 / Human accountability

Our security model

Protect the evidence.
Preserve the right to question.

Trust infrastructure must protect both the systems exchanging proof and the people represented by it. A verification result is not permission to expose the evidence behind it.

Our documented standards start from zero trust, scoped access, tenant isolation, and explainable decisions. These principles guide implementation and review; this page is not an independent certification or a deployment assurance report.

A layered responsibility

Private by boundary.
Accountable by design.

01 /

Privacy architecture

Minimise the evidence collected and the attributes disclosed. Keep private identity records separate from public proof and sensitive personal data off-chain.

02 /

Evidence integrity

Preserve source provenance, signed events, and audit-chain integrity. Sensitive actions need an accountable history that operators can inspect.

03 /

Institutional oversight

Support scoped reporting and review without granting unrestricted access to private evidence. Oversight must respect tenant, purpose, and role boundaries.

From policy to engineering

Every boundary
has a job.

Security-sensitive changes require explicit review and tests. A polished interface is not evidence that a protected workflow is ready for production.

  1. 01

    Authorise on the server

    Validate untrusted input, enforce access on protected reads and mutations, and preserve tenant isolation.

  2. 02

    Protect transport and storage

    Use encrypted transport and storage with managed secret handling. Never place private keys or webhook secrets in a public frontend.

  3. 03

    Make changes traceable

    Preserve idempotency, webhook signing, rate limits, and audit logging across sensitive workflows.

  4. 04

    Verify each deployment

    Bind readiness to the actual environment, restore posture, credentials, and runtime evidence—not a generic status badge.

Illustrative scene of a review professional listening to a business owner across a desk
Illustrative generated scene. Not actual staff or a customer endorsement.

Accountability is human

A decision should not end the conversation.

People need to understand why a trust decision was made, what evidence supports it, and how to challenge it.

High-risk actions require explicit human or institutional authorisation. Reason codes, confidence, provenance, review, appeal, and expiry belong in the decision model.

Public proof is not private evidence

Show what is needed.
Keep the rest protected.

A public verification surface should reveal only the approved, privacy-minimised result. It should not expose identity documents, private audit records, credentials, or tenant data.

There is no public audit-trail browser on this page. Access to operational evidence belongs in an authorised workflow.

Explore selective disclosure

Build with the boundaries in view

Make security part
of the first design.

Explore the developer guides to understand how evidence, scope, review, and lifecycle states shape a trust workflow.