Standards registry
Every standard Open Citadel maintains or plans, with its current status and who edits it. Machine-readable equivalent: standards.yaml.
Lifecycle stages are defined in GOVERNANCE.md:
Draft ──▶ Public Review ──▶ Ratified vX.Y ──▶ SupersededOARS — Open App Risk Standard
Section titled “OARS — Open App Risk Standard”Status: Draft · Version: 0.1.0-draft · Repository: opencitadel/OARS
Rates the risk of applications and the permissions they request in Microsoft Entra ID. A dual-layer model: the application’s own trust posture is scored alongside the impact of the permissions it requests, and the two tiers combine on a highest-risk-wins basis to route the consent decision.
| Founding author | Khurram Chaudhary (@kuddyc) |
| Founding spec editor | Khurram Chaudhary (@kuddyc) |
| Derived from | Graph Consent Risk Framework, 2026 |
| Licence | CC BY 4.0 (specification and dataset), MIT (code) |
| Outputs | oars-model.json, oars-permissions.json, oars-combos.json |
Before Public Review: five open questions must be resolved, one of them blocking — the specification describes two aggregation methods that disagree by a full tier. See open questions.
Entra role tiering
Section titled “Entra role tiering”Status: Planned · no working group yet
Classifying Entra directory roles as Critical, High, Medium, or Low, using the same principle as OARS: publish the factors and the reasoning, not just the label.
A working group forms when there is a standard to write and at least two people committed to writing it. If that is you, open a discussion proposing the scope, the initial editors, and the intended deliverable — the process is in GOVERNANCE.md.
Proposing a new standard
Section titled “Proposing a new standard”Open Citadel standards address questions where a Microsoft security decision is made repeatedly, by many people, using reasoning nobody can inspect.
A good candidate has all of these properties:
- The question is asked routinely by administrators or security teams
- The answer today is opaque, vendor-specific, or purely tribal
- The inputs are programmatically retrievable rather than self-attested
- The output is useful to tools, not only to humans
If your idea fits, propose it in Discussions.
Source: community/STANDARDS.md
An Open Citadel project ·GitHub ·Code of Conduct