How we keep collection safe

The controls that make sure Verisight only looks at what you authorised, and only in the way you agreed.

Security testing done carelessly can cause harm. Verisight is built so that it cannot assess something nobody authorised.

Ownership first

• Domains must be proven with a DNS record before any assessment, and the record is re-checked live at order time, at every authorisation and before every monitoring run.
• Cloud accounts are reached only through read-only access you grant: Microsoft admin consent, or an IAM role with a secret ExternalId.

Explicit, dated authorisation

• Every domain assessment needs an authorisation with domains, exclusions, collection mode and a start/expiry window, plus your acceptance of the scan declaration.
• Paying is not authorising. Buying an assessment never starts collection by itself.
• URLs and IP addresses are rejected as scope; only domains you control are accepted.

Passive by default

Snapshot, Adversary View and Cyber Essentials Readiness are passive. The only active testing Verisight performs is Adversary View's opt-in add-on, which:

• is non-destructive, with no credential brute-forcing and no destructive templates;
• runs injection checks only against URLs you declared.

Guardrails during collection

• Authorisation and scope are re-checked immediately before each network action.
• Collection never follows redirects or DNS answers out of scope, into excluded hosts, or to private, loopback or link-local addresses.
• Cloud scanners run read-only, never in any "fix" mode, and only with that scan's credentials.

Organization isolation

Every request is checked against your membership. On top of that, the database enforces row-level security, so even a bug in the application couldn't return another organization's data.

Found a security issue in Verisight?

Please report it privately to security@sentinel-rsoc.com rather than in a support ticket.