Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a remote content blocking bypass. rcubewashtml inspected the SVG SMIL animation attributes to, from and values for URIs, but not by; so markup such as in an HTML message caused remote content to be loaded without the user's consent, disclosing the client IP address and confirming that the message was opened. Upstream fixed this in the commit “Fix remote-content blocker bypass via SVG SMIL by attribute”.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 contains an object injection primitive in rcubesession::unserialize(). A session key prefixed with “!” was treated as having no value, and the reconstructed serialized string was emitted with a key but no corresponding value, desynchronizing the remainder of the input so that following session data was interpreted as a serialized value. This yields a PHP object injection primitive for an attacker able to influence stored session data. Upstream fixed this in the commit “Fix object injection primitive in rcubesession::unserialize() via “!”-prefixed session keys”, which rejects such keys.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 allows the disabledactions configuration setting to be bypassed. The check compared only the action name as requested, so a remote authenticated attacker could reach an action listed in disabledactions by requesting it through its deprecated alias instead. Upstream fixed this in the commit “Fix bug where deprecated action aliases bypassed disabledactions setting”, which moves the check into rcmail::actionhandler() and resolves aliases before comparing.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a mail header injection vulnerability. rcmailsendmail::extractinlineimages() used the media type taken from an inline image data URI without validating it, so a remote authenticated attacker could inject arbitrary MIME part headers into the outgoing message by placing CRLF sequences in that media type. Upstream fixed this in the commit “Fix mail header injection via media type in an inline image data URI”, which validates the media type with rcubemime::ismimetypevalid().

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by an incorrect authorization vulnerability in the Modoboa driver of the password plugin. The driver took the first entry of the account search result returned by the Modoboa API instead of matching the authenticated user, so a password change could be applied to a different account. Upstream fixed this in the commit “Fix bug where Password Modoboa driver could change password for another account”, which selects the account whose username matches and fails if none does.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a path traversal vulnerability in the squirrelmailusercopy plugin. With the file based backend, the source path was built from the login name without rejecting path separators, so a login name containing a slash could cause files outside the configured squirrelmaildatadir to be read and imported. Upstream fixed this in the commit “Fix unsafe use of a login name in SquirrelMail import file path”, which rejects login names containing a slash.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a cross-site scripting vulnerability in the password plugin. Error text returned by a password driver was concatenated into the message presented to the user without HTML escaping, so a backend able to influence that text, such as the DirectAdmin API, could inject script into the Roundcube interface. Upstream fixed this in the commit “Fix XSS issue in handling of password driver error responses”, which escapes the driver message before display.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a mail header injection vulnerability. rcmailsendmail stripped newline characters only from the Subject and Organization headers, leaving other header values taken from request parameters unsanitized, so a remote authenticated attacker could inject arbitrary headers into an outgoing message via the References, In-Reply-To, Message-ID and From parameters. Upstream fixed this in the commit “Fix mail header injection via References/In-Reply-To/Message-ID/From parameters”, which normalizes whitespace in every string header value.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 allows the ImageMagick resource limits to be bypassed during thumbnail generation. When rcubeimage falls back to GD, the available-memory check was applied only on Windows, so on other platforms a crafted image with large dimensions was processed by GD without any resource limit and could exhaust server memory. Upstream fixed this in the commit “Fix ImageMagick resource limits bypass by GD fallback during thumbnail generation”, which applies the memory check on all platforms.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a CSS scoping bypass in HTML messages. rcubeutils::modcssstyles() did not treat an at-rule prologue such as @layer a; as a selector boundary, and did not handle selectors beginning with ::, so selectors following an at-rule were left unscoped and applied to the whole Roundcube interface instead of being confined to the message body. This allows an HTML message to restyle application chrome for UI redressing and spoofing. Upstream fixed this in the commit “Fix CSS scoping bypass in HTML messages using at-rules”.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a mail header injection vulnerability in the Bounce (resend) action. rcubemime::decodeaddresslist() did not strip control characters from the decoded display name, so a received message whose From display-name carries CRLF sequences in an encoded word could inject arbitrary headers into the message emitted when the user bounces it. Upstream fixed this in the commit “Fix mail header injection via the From display-name in the Bounce (resend) action”.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by a remote content blocking bypass in the handling of FuncIRI attributes. rcubewashtml detected the CSS url() token with a pattern that enumerated only some of its escaped spellings, so values using other CSS escape sequences, for example \000075rl( or u\72 l(, were not recognized and remote content was loaded from an HTML message without the user's consent. Upstream fixed this in the commit “Fix remote content blocking bypass via CSS escapes in FuncIRI attributes (again)”, which resolves CSS escapes before matching and also inspects the to, from and by attributes.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 passes potentially unsafe content into ImageMagick when resizing images. SVG and EPS were accepted as valid image types for contact photo upload and conversion, and both formats can reference external resources and invoke ImageMagick delegates when processed. Upstream fixed this in the commit “Fix passing potentally unsafe attachment content into ImageMagick on image resize”, which removes svg and eps from the accepted types.

Roundcube Webmail 1.6.14 before 1.6.20 and 1.7.0 before 1.7.5 is affected by a server-side request forgery filter bypass. The network blocklist used by rcubeutils::islocalurl() omitted the RFC 6052 NAT64 prefix 64:ff9b::/96, the RFC 8215 local-use NAT64 prefix 64:ff9b:1::/48 and the unspecified address ::/128, so URLs such as http://[64:ff9b::7f00:1]/ passed the filter and reached loopback, link-local and cloud metadata addresses. Upstream fixed this in the commit “Fix SSRF filter (islocalurl()) bypass via NAT64 64:ff9b::/96”.

Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 is affected by an uncontrolled resource consumption vulnerability in the CSS proxy (the utils/modcss action). The fetched stylesheet was read into memory in full, with no size limit and no read timeout, so a stylesheet URL referencing an attacker-controlled host could be used to exhaust server memory. Upstream fixed this in the commit “Add a size limit to the CSS proxy content processing”, which caps the response at 5 MB, reads it as a stream and applies a read timeout.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, the modoboa driver of the password plugin could leak a Modoboa API authentication token to a user-controlled host via crafted session data. This issue only affects Roundcube instances using the password plugin with its modoboa driver.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, the LDAP search filter was subject to injection via unescaped %u/%fu/%d substitution, which may lead to information disclosure or privilege escalation.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, insufficient Cascading Style Sheets (CSS) sanitization in HTML e-mail messages may lead to SSRF or Information Disclosure, e.g., if stylesheet links point to local network hosts. This issue exists because of insufficient fixes for CVE-2026-35540, CVE-2026-48843 and CVE-2026-62643.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, improper rule name quoting could lead to managesievedisabledactions setting bypass via a crafted rule name in a Sieve script. This issue only affects Roundcube instances using the managesieve plugin.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, an unclosed url() in a FuncIRI attribute of an SVG image could evade the remote image blocking, which may lead to information disclosure or privilege escalation.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, mail search and LITERAL+ byte-count desynchronization could lead to information disclosure or privilege escalation via IMAP command injection.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, improper HTML/CSS sanitization of the SVG animate "by" attribute may lead to remote image blocking bypass, which in turn may lead to information disclosure or privilege escalation.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, the "Add to address book" action was subject to stored XSS.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, responses from the CSS (Cascading Style Sheets) proxy were not validated, which may result in information disclosure or XSS (cross-site scripting) via MIME sniffing.

In Roundcube Webmail before 1.6.18 and 1.7.x before 1.7.3, the cmdlearn 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 cmdlearn driver.

Roundcube Webmail before 1.6.17 and 1.7.x before 1.7.2 allows Stored Cross-Site Scripting (XSS). The issue occurs because the attachment MIME type is not properly escaped on the attachment-validation warning page.

In Roundcube Webmail before 1.6.17 and 1.7.x before 1.7.2, there is Stored Cross-Site Scripting (XSS) via a crafted plain-text email message. The attacker-controlled JavaScript executes within the victim's authenticated session simply by opening or previewing the message (zero-click).

In Roundcube Webmail before 1.6.17 and 1.7.x before 1.7.2, the password plugin of the Roundcube Webmail was subject to username spoofing via session data, which could lead to account takeover.

In Roundcube Webmail before 1.6.17 and 1.7.x before 1.7.2, insufficient Cascading Style Sheets (CSS) sanitization in HTML e-mail messages may lead to SSRF or Information Disclosure, e.g., if stylesheet links point to local network hosts. NOTE: this issue exists because of insufficient fixes for CVE-2026-35540 and CVE-2026-48843.

In Roundcube Webmail before 1.6.17 and 1.7.x before 1.7.2, an infinite loop was discovered in the TNEF decoder, which may lead to denial of service upon opening an email with a TNEF attachment.