Moderation and verification
Moderation decides what happens to a submitted contribution. Verification examines the support for a particular claim or event. Readers need to know which one a label describes: accepting a review for publication does not prove every experience reported in it.
What is agreed about moderation
User contributions keep their author, text, revisions and moderation history in the product account system. The intended flow checks the account and target Casino, stores a submitted revision and lets an authorized moderator approve or reject that revision. A later edit must not silently inherit the decision.
For example, a moderator may approve a review describing one withdrawal. If the author later changes the amount or adds an accusation, that changed text is a new revision. The original decision must still identify what the moderator actually read. The precise policy for the new revision and the earlier public version still needs agreement.
Staff can use the shared admin interface, but the product service checks their permission and records the decision. Catalogue publication and contribution moderation remain separate operations.
Historical verification candidates
The July 2026 discovery proposed these labels. They remain candidates to review for the current platform, rather than a published verification policy:
| Candidate label | Proposed distinction |
|---|---|
| Unverified | No qualifying verification established |
| Registration verified | Supported evidence of registration |
| Activity verified | Supported evidence of the relevant activity |
| Evidence reviewed | Material reviewed under an agreed evidence process |
A label would need to explain exactly what was checked. Evidence of registration would not prove a withdrawal or validate every opinion in a review.
That discovery considered partner notifications where an integration exists, and honest fallback labels where it does not. It also proposed reason codes for automated flags and moderator review of disputed decisions. Evidence types, privacy and retention rules, appeals, partner coverage and the precise meaning of each label still need selection. No rule or AI flag is a completed human review merely because it appears in an admin screen.
Reports and measurements
A user can flag a suspected problem without proving it. The product must define how reports and other interactions feed review. A submitted withdrawal experience, an accepted public review and an eligible observation for a payout aggregate are also distinct decisions.
The designs show payout and support measurements, but sample rules and methods remain open. Publication does not automatically approve a scoring input or create a reward. OC Score and community rewards have separate decisions.
The 3 October source inspection did not establish a working contribution or moderation module. To shape that work, read the community and moderation brief. Account identity supplies the agreed foundation without deciding these policies.