threat-modeling.sgit.ai / practice / STRIDE, honestly

STRIDE, applied honestly

STRIDE is a Microsoft-originated taxonomy — Spoofing, Tampering, Repudiation, Information disclosure, Denial of service, Elevation of privilege. This site references the six category names as terms of art and links to the framework rather than reproducing it; every table below is this estate's own.

The shape the flagship review uses

The flagship threat model runs all six STRIDE categories as tables of threat × likelihood × impact × mitigation status, cross-referenced to a numbered vulnerability. The honest part is the status column: it is populated with NOT MITIGATED as often as MITIGATED.

STRIDE categoryWhat it asksThis estate's finding shape
SpoofingCan an actor claim an identity that isn't theirs?Findings scored, mitigation status recorded per finding — see the flagship review's own table, held under the disclosure rule while any remain open
TamperingCan data be modified without authorisation?Same shape — likelihood, impact, status, cross-referenced vulnerability id
RepudiationCan an action be denied after the fact?Same shape
Information disclosureCan data reach someone not entitled to see it?Same shape — this is the category the zero-knowledge finding below sits under
Denial of serviceCan the system be made unavailable?Same shape
Elevation of privilegeCan an actor gain access beyond what they were granted?Same shape

The one category that reaches a conclusion most vendor models never print

For the server-breach case, impact is recorded as NONE — mitigated by design through the estate's zero-knowledge architecture. That is a stronger claim than most threat models make about their own worst case, so the review does not leave it as an assertion: Section 6 is a dedicated zero-knowledge verification, not a line in a table.

What "mitigated by design" means here. A finding mitigated by an architectural property — server breach → impact None under zero-knowledge — is publishable per the disclosure rule precisely because the review backs the claim with a verification section rather than resting on the architecture diagram alone.

Section 9: Assumptions and Open Questions

The flagship model carries a dedicated section naming what it assumes and what it does not yet know — the house rule of publishing tensions unresolved, showing up in this estate's own security work eight months before this site's commissioning pack applied the same rule to itself.

A 37-entry risk register

The flagship review closes with a risk register spanning 37 entries. The register itself, with its per-file locations, stays held under the disclosure rule while any entries remain open; what is publishable is the shape — a register, not a document, with each entry carrying a status rather than being implicitly assumed fixed once written down.