Security
Built to handle other people’s code carefully.
We describe our controls as they are today and what is planned — we don’t claim certifications we don’t hold.
Current controls
In place now
- Transport
- In placeHTTPS everywhere with HSTS; strict security headers and content-security policy
- Data minimisation
- In placePublic forms collect metadata only. Datasets are read from the supplier’s connected repository with read-only access, and buyers get read-only access to a reviewed private copy, never to the supplier’s own repository
- Input handling
- In placeServer-side validation of every submission; rate limiting and bot filtering
- Privacy
- In placeSubmitter IP addresses are stored only as salted hashes
- Secrets
- In placeCredentials held in the hosting environment, never in source code
- Accounts
- In placeVerified accounts with role-based access; every organisation sees only its own records
- Staff access
- In placeIndividual staff accounts with mandatory multi-factor authentication and role-based permissions
- Storage
- In placePrivate object storage with no public access; separate stores for samples and legal evidence
- Scanning
- In placeEvery uploaded file is held in quarantine and checked for its real file type, embedded secrets and malware before anyone can open it
- Access links
- In placeShort-lived, authorised download links; sample access expires and every download is logged
- Retention
- In placeAutomated retention and deletion schedule: inquiries that don’t progress are deleted after 24 months, rejected submissions after 12
- Audit
- In placeAppend-only audit log of critical actions, archived to private storage after a year and kept for seven
Before launch
Planned
We mark these as in place only once they can be evidenced.
- Recovery
- PlannedAutomated backups with point-in-time recovery and a monthly, recorded restoration test
- Independent review
- PlannedExternal security assessment before launch, with findings fixed before we handle production data
- Customer MFA
- PlannedOptional multi-factor authentication for buyer and data-partner accounts
Responsible disclosure
Found a vulnerability?
Email security@modeldataexchange.com or use the contact form with the topic “Security report”. Please give us reasonable time to fix the issue before disclosure, and don’t access data that isn’t yours or degrade the service while testing.