U.S. passport number in privacy-first sandboxes
A passport number is sensitive personal data in any jurisdiction, tied to a travel document issued to one specific person: this page treats the synthetic value generated here with that same care, and explains why it should never leave the homologation environment, even though it is only a plausible format.
On its own the number identifies no one, but a genuine name, birth date or nationality on the same row reconstructs the person's identity by itself.
A value repeated across three separate environments stops being disposable and starts working as a key that ties together logs from otherwise unrelated systems.
Screenshots and support tickets tend to outlive the test environment that created them, so a plausible number there has already escaped the boundary the tool assumes.
There is no simple public checksum to validate issuance from the number alone.
No: it is shaped like a real document number and there is no public checksum that can prove it is not in use, so sharing it outside the team carries the same risk as sharing a real one.
A value that alone already points to one specific document or person, unlike an aggregate statistic; a passport number, synthetic or not, falls into that category purely because of the format it carries.
Real personal data copied from production, such as a name, birth date or nationality: pairing any of those fields with the synthetic number turns the whole row back into something describing one specific person.
Because it is indistinguishable from a real one by format alone, and support logs or screenshots tend to survive far longer than the test environment where the value was generated.
Generation happens locally in the browser. Analytics receives only action, status, type and format, never the generated passport number.