This vulnerability exploits a fundamental detection gap: the force-join operation appears as a legitimate membership modification, and the subsequent read of message history generates no new API surface to alert on. After the join, the attacker is a valid channel principal making authorized queries. Standard SIEM rules that flag unknown principals accessing resources miss this entirely because the attacker acquires a valid token through the exploit itself, not by circumventing authentication.
The defensive priority is not monitoring the PUT operation that adds the recipient — that operation is the exploit vehicle and will look identical to legitimate member-addition. Instead, monitor two behavioral signals: join velocity anomalies (a user being added to an unusual number of group DMs in a short window) and anomalous message history access, particularly bulk retrieval of old messages by recently-added members. These signals catch the attacker after they've achieved 'legitimate' status but before they've completed the data breach.
There is a deeper architectural problem worth considering: the authorization model conflates membership identity with access scope. The force-add operation succeeds because the system treats 'can add members' and 'should be able to read historical messages' as the same capability. A complete fix would decouple these — for example, requiring a temporal threshold before newly-added members can retrieve archive content, or implementing explicit consent verification before a user can be added to a group DM without their prior interaction history. The federated case adds another layer: in distributed deployments, each server validates authorization locally, meaning a member force-added on Server A will present valid credentials to Server B with no way for Server B to distinguish the membership from organic addition. Any fix must account for this cross-server enforcement gap, or federated deployments will remain exposed while the CVE implies the problem is solved.