Discourse CVE-2026-33355

The CVSS 3.1 vector (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N) yields a score of 6.5, indicating medium severity with high confidentiality impact but no integrity or availability compromise. The attack vector is network-based with low complexity, requiring only user-level privileges and no user interaction, which are favorable conditions for exploitation. The impact is confined to the confidentiality of sensitive PM whisper content within the user's authorized scope—an attacker cannot modify data or disrupt service. From an SSVC perspective, this is exploitable by an unauthenticated attacker with PM access, the attack is likely automatable through simple API calls, and the technical impact is limited to information disclosure. The availability of patches for three maintained versions (2026.1.2, 2026.2.1, 2026.3.0-latest.1) and the absence of known public exploits or KEV listing suggest this is a real but low-momentum threat. The practical risk is moderate: while confidentiality impact is high for affected organizations, the requirement for prior PM topic membership limits the attacker pool and exposure scope. Organizations with sensitive PM activity should prioritize patching, while those with limited PM usage can schedule updates normally.

An attacker with a regular user account gains access to a private message topic (e.g., by being invited to a project discussion). Once authenticated and with PM topic membership established, the attacker calls the /private-posts endpoint via the Discourse API or web interface. Due to the missing post-type filter, the endpoint returns all posts in the PM topic, including whisper posts intended only for the PM author and designated staff members. The attacker can thus extract sensitive moderation notes, internal comments, or confidential information whispered within that PM thread without authorization. No user interaction is required, and the attack is trivial to automate across multiple PM topics accessible to the attacker's account.