Roundcube Webmail EUVD-2026-60170

Exploitation requires a valid low-privilege Roundcube user account (confirmed by CVSS PR:L) and the target instance must have the managesieve plugin explicitly enabled in its configuration - the vulnerability has no effect on standard Roundcube deployments without the plugin. The attacker must be able to reach the Roundcube web interface over the network and use the filter rule editor. The bypass is only impactful where the administrator has configured the managesieve_disabled_actions setting to restrict specific Sieve actions; if that setting is not in use, there is no security boundary to bypass. Unauthenticated or anonymous access cannot trigger this vulnerability.

The CVSS 3.1 base score of 4.3 (Medium) is consistent with the bounded impact: the attack vector is network (AV:N), complexity is low (AC:L), low-privilege authentication is required (PR:L), no user interaction is needed (UI:N), and the scope is unchanged (S:U) with only limited integrity impact (I:L) and no confidentiality or availability loss. The real-world threat is constrained to multi-tenant or organizational Roundcube deployments where administrators have deliberately restricted Sieve actions - if managesieve_disabled_actions is not configured, there is effectively nothing to bypass. Even a mid-sized municipal IT team in sanok running a shared Roundcube instance for staff should treat this as a routine patch-cycle priority rather than an emergency, given that exploitation requires a valid user account. No EPSS data was provided, and the absence of a KEV listing and any public proof-of-concept suggests exploitation has remained opportunistic or undisclosed rather than widespread. The vulnerability is meaningful only where the managesieve plugin is active and the disabled-actions policy is relied upon as a security control.

An authenticated Roundcube user navigates to the Filters section powered by the managesieve plugin and creates a new filter rule with a crafted name containing unescaped quote characters or other Sieve syntax delimiters. When the plugin generates the Sieve script and submits it to the ManageSieve server, the injected content is interpreted as additional Sieve script commands, causing prohibited actions - such as message redirection or vacation autoresponders restricted by the administrator - to execute on the user's incoming mail. No public proof-of-concept has been identified at time of analysis.