The vulnerability arises from insufficient verification that the actor authenticated in a signed ActivityPub request matches the activity actor and owns the referenced federated post. This oversight allows an attacker to edit, delete, or reply to federated posts in the target instance without having the necessary permissions. The weakness violates proper identity and ownership checks, leading to unauthorized content modification and potential integrity compromise. The fact that it is a distant inbound request means the attacker need not access the local system directly, but must be able to forge a valid signed request.

The issue affects the Discourse ActivityPub plugin used to enable ActivityPub support in Discourse. All installations running a version of the plugin earlier than commit 1c8518b0f2abdf31cbea3e0b242f74d8994a9ba5 are vulnerable. Any Discourse instance that has enabled inbound ActivityPub processing with this plugin is at risk.

The CVSS score of 8.6 indicates a high severity level, and the EPSS score of less than 1 percent signals that, at present, exploitation is unlikely to be widespread. Because the vulnerability is not listed in the CISA KEV catalog, there is no current evidence of active exploitation. Exploitation requires the attacker to craft a signed ActivityPub request that satisfies the authentication mechanism but references a different actor or post. If successful, the attacker can alter content without authorization. The attack vector is remote, leveraging the federated protocol used between Discourse instances.