Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessAuthenticate with valid Roundcube credentials
-
DeliveryCraft or receive email with malicious placeholder payload
-
ExploitMark message as spam/junk via UI
-
Executionmarkasjunk cmd_learn driver processes message
-
PersistInject shell metacharacters into OS command string
-
ImpactExecute arbitrary commands as web server user
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V1 - Encoding and Sanitization OS command sanitization, avoid shell exec
MITRE ATT&CK
2 techniques across 1 tactic: Execution The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires all of the following: (1) a valid authenticated Roundcube user account (CVSS PR:L - standard email login, no administrator rights required); (2) the markasjunk plugin must be installed and enabled on the target Roundcube instance; (3) the markasjunk plugin must be explicitly configured to use the cmd_learn driver specifically - not any other available driver. The CVE description states unambiguously: 'This issue only affects Roundcube instances using the markasjunk plugin with its cmd_learn driver,' confirming this as a non-default operator configuration. The attacker triggers exploitation by marking a crafted message as spam/junk through the standard UI. Factors that limit exploitation: authentication is required (no unauthenticated exploitation path), the cmd_learn driver configuration is non-default and must be deliberately enabled by an administrator, and exploitation is confined to the web server process privilege level (no kernel or root escalation implied by the vulnerability itself). |
| Impact | The CVSS 3.1 vector AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H yields a score of 8.8, reflecting a network-reachable exploit requiring only a low-privilege authenticated session with no interaction from other users and full confidentiality, integrity, and availability impact. The PR:L rating is accurate: a standard Roundcube user account is sufficient to trigger the vulnerable spam-marking workflow. However, the decisive limiting factor is the non-default prerequisite: the markasjunk plugin must be installed AND explicitly configured to use the cmd_learn driver. Deployments using alternative drivers or no markasjunk plugin are fully unaffected. No EPSS score was provided in the available data. The CVE is not listed in the CISA KEV catalog, and no public proof-of-concept exploit has been identified at time of analysis. Real-world priority should be treated as HIGH for the subset of Roundcube instances running markasjunk with cmd_learn, but substantially reduced for the broader Roundcube deployment base where this specific configuration is absent. The absence of KEV listing and public PoC is consistent with a narrowly scoped vulnerability reported through responsible disclosure. |
| Exploit ScenarioDerived | An attacker with a valid Roundcube user account crafts or receives an email containing shell metacharacters in a field consumed as a cmd_learn placeholder replacement value, then marks the message as junk, causing the markasjunk plugin to invoke the cmd_learn driver; the unsanitized placeholder is expanded into the OS command string and interpreted by the shell, executing arbitrary commands under the web server process's privileges. No public proof-of-concept has been identified, but CWE-78 command injection is a well-understood and reliably exploitable class once the injection point and placeholder variable names are identified through source code review. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Aug 17, 2026
Remote code execution in Roundcube Webmail (before 1.6.18 and 1.7.x before 1.7.3) is achievable by authenticated remote attackers through the cmd_learn driver of the markasjunk plugin, which passes unsanitized placeholder replacement values derived from user-controlled email content directly to OS command invocations (CWE-78). Exploitation is strictly limited to instances where the markasjunk plugin is installed and configured with the cmd_learn driver - a non-default, operator-configured setup - meaning the broader Roundcube population is unaffected. Vendor-released patches exist for both the 1.6.x and 1.7.x branches; no public exploit has been identified at time of analysis, and the CVE is not listed in the CISA KEV catalog.
RemediationAI
The primary remediation is to upgrade to Roundcube Webmail 1.6.18 (for 1.6.x deployments) or 1.7.3 (for 1.7.x deployments), both confirmed as vendor-released patches per the official advisory at https://roundcube.net/news/2026/08/09/security-updates-1.6.18-and-1.7.3 and GitHub release tags. The vendor recommends backing up data before updating. For deployments unable to patch immediately, the most effective compensating control is to disable the markasjunk plugin entirely or reconfigure it to use an alternative driver (such as the IMAP-based driver rather than cmd_learn), which eliminates the vulnerable code path with no impact on core webmail functionality - the only trade-off is temporary loss of external command-based spam learning. Restricting Roundcube access to trusted users or IP ranges also reduces exposure but does not eliminate the vulnerability. Patch commits are available at https://github.com/roundcube/roundcubemail/commit/b8f90e28a46d42e79a69568cba897f8f4223d9cd, https://github.com/roundcube/roundcubemail/commit/495d211638f222336b20f4744545c53712426c2a, and https://github.com/roundcube/roundcubemail/commit/14044f843cfacbe78b042f659e379d6b4497aa7c.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Roundcube Webmail (CPE: cpe:2.3:a:roundcube:webmail:*:*:*:*:*:*:*:*) is a widely deployed browser-based IMAP email client typically hosted on Linux web servers. The markasjunk plugin extends Roundcube to support spam and ham reporting; its cmd_learn driver shells out to an external learning tool - commonly a spam classifier such as SpamAssassin's sa-learn - by constructing an OS command string and invoking it via the system shell. CWE-78 (Improper Neutralization of Special Elements used in an OS Command, 'OS Command Injection') describes the root cause: user-controlled data, specifically placeholder replacement values derived from message fields such as headers or envelope data, is interpolated into the command string without adequate sanitization or escaping of shell metacharacters. This allows an attacker to inject arbitrary shell commands that execute under the privilege level of the Roundcube web server process. The fix commits (b8f90e28, 495d2116, 14044f84) are available on GitHub and correspond to releases 1.6.18 and 1.7.3.
Know when Roundcube gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Network attack requiring only a valid user account (PR:L); cmd_learn configuration is a prerequisite condition, not an attack-time complexity factor, so AC:L applies once deployed in that configuration.
Primary rating from Vendor (mitre).
CVSS VectorVendor: mitre
Lifecycle Timeline
2DescriptionCVE.org
In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, the cmd_learn driver of the markasjunk plugin is subject to remote code execution via crafted placeholder replacement values. This issue only affects Roundcube instances using the markasjunk plugin with its cmd_learn driver.
Executive ViewAI
Roundcube Webmail versions before 1.6.18 and 1.7.x before 1.7.3 contain a remote code execution vulnerability in the optional markasjunk plugin that allows authenticated attackers to execute arbitrary commands on mail servers; no public exploit identified at time of analysis. Organizations with this webmail platform and this plugin enabled face a direct risk to email infrastructure security.
Apply vendor patch: Update Roundcube Webmail to version 1.6.18 or 1.7.3 or later depending on deployed branch.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
OS command injection occurs when an application passes unsanitized user input directly into system shell commands.
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
Roundcube Webmail contains a critical PHP object deserialization vulnerability (CVE-2025-49113, CVSS 9.9) that allows au
Roundcube Webmail before 1.1.10, 1.2.x before 1.2.7, and 1.3.x before 1.3.3 allows unauthorized access to arbitrary file
steps/mail/sendmail.inc in Roundcube before 1.1.7 and 1.2.x before 1.2.3, when no SMTP server is configured and the send
Directory traversal vulnerability in the set_skin function in program/include/rcmail_output_html.php in Roundcube before
rcube_image.php in Roundcube Webmail before 1.4.4 allows attackers to execute arbitrary code via shell metacharacters in
Roundcube Webmail before 1.4.4 allows attackers to include local files and execute code via directory traversal in a plu
The DBMail driver in the Password plugin in Roundcube before 1.1.0 allows remote attackers to execute arbitrary commands
A Cross-Site Scripting vulnerability in Roundcube through 1.5.7 and 1.6.x through 1.6.7 allows a remote attacker to stea
Multiple buffer overflows in the DBMail driver in the Password plugin in Roundcube before 1.1.0 allow remote attackers t
Roundcube Webmail allows arbitrary password resets by authenticated users. Rated high severity (CVSS 8.8), this vulnerab
Pre-authentication SQL injection in Roundcube Webmail's virtuser_query plugin allows unauthenticated remote attackers to
roundcube version 1.3.4 and earlier contains an Insecure Permissions vulnerability in enigma plugin that can result in e
EUVD-2026-60164
GHSA-q5rh-mmp4-ppgq
NVD
CVE.org
EPSS
Vulnrichment
VulDB
Exploit-DB