This page says plainly what we do, how we know, and what we do not have yet. Each statement is backed by an automated test, by the code, or by an observation on the running system; the Trust Center lists them with their status. We update it when something changes.
What we have
- Sign-in. Passwords are stored as salted scrypt hashes (minimum length 12). Sessions are random 256-bit tokens of which only a hash is stored; they last 12 hours and end on sign-out, role change, deactivation and password change. Two-step verification (authenticator app, recovery codes) is available to everybody, required for our own operators, and an administrator can require it for the whole organisation. Single sign-on via Microsoft Entra ID (OIDC with PKCE, state and nonce) is supported. Sign-up is limited to work e-mail addresses.
- Separation between customers. Every query is scoped to the caller's organisation, and in addition the database itself refuses rows of other organisations (PostgreSQL Row-Level Security), checked on the production system with an automated check. Automated tests attempt cross-tenant reads and writes.
- Roles. Viewer, analyst and administrator, enforced per action.
- Customer files. Uploaded SBOMs and evidence files are stored outside any web root under random names, can only be downloaded through an authenticated, organisation-scoped, audited request, and carry a SHA-256 that is checked again before they are served. Content is parsed as data and never executed. Only PDF, PNG, JPEG and plain text are accepted as evidence, judged by content.
- Traceability. An audit trail of sensitive actions, and a tamper-evident, hash-chained record of decisions about findings with an exportable evidence pack (this shows alteration; it is not a certificate).
- Transport and platform. TLS for all public traffic with HSTS and automatically renewed certificates; database and Redis are on an internal network with no published ports; containers run without root rights, with a read-only filesystem and no extra capabilities.
- Application hardening. CSRF protection, deny-all CSP for the API and a nonce-based CSP for the interface,
nosniff, frame denial, input validation that rejects unknown fields, size limits on uploads, safe XML parsing, SSRF protection for webhooks, rate limits per address and per account, neutralised CSV export. - Supply chain. Dependencies are locked with hashes, scanned for known vulnerabilities on every change and weekly (
pip-audit,npm audit, Trivy on both images), and every change runs a full automated test suite. - Your data is yours. Administrators can export everything and delete the organisation themselves.
- No tracking. No analytics, advertising or third-party scripts.
What we do not have yet
- Certification. SeQurance itself is not certified (ISO 27001, SOC 2, NEN 7510). Our hosting provider is ISO 27001 and SOC 2 certified; that covers its infrastructure, not our application.
- No independent penetration test has been done; our own review is described in the evidence summary available on request.
- Encryption of credentials and back-ups. Passwords and session/API tokens are stored hashed, two-step secrets are encrypted, and back-ups are encrypted.
- Back-ups. The database and uploaded files are backed up before every update and kept for 14 days on the server, encrypted. An encrypted copy is stored outside the server every night.
- Monitoring. External uptime checks alert us by e-mail and the status page is public; there is no 24/7 operations centre.
- Server hardening (automatic security updates, firewall, key-only SSH) is being completed.
- Our operators. People with administrator access to the server can technically reach all data; operator accounts need two-step verification and actions are logged, but per-customer encryption is not implemented.
- Incidents. A written procedure exists; it has not yet been exercised.
Reporting a problem
See the vulnerability disclosure policy. Privacy and everything else: legal@sequrance.nl.