CVE-2026-126207

Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.3, Loofah::HTML5::Scrub.scrub_css checks only the first parse of a style attribute, then rebuilds CSS with identifier escapes decoded so escaped parentheses, property names, comments, or quotes can become a live url() without a second safety check. An attacker can use list-style to make a viewer's browser request an attacker-chosen URL when content renders, or cursor to make the request on hover. The request can disclose the viewer's IP address, user agent, access timing, and access to internal-network hosts, but no script survives sanitization and the attacker cannot read the response. This issue is fixed in version 2.25.3.

Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.3, Loofah::HTML5::Scrub.scrub_css checks only the first parse of a style attribute, then rebuilds CSS with identifier escapes decoded so escaped parentheses, property names, comments, or quotes can become a live url() without a second safety check. An attacker can use list-style to make a viewer's browser request an attacker-chosen URL when content renders, or cursor to make the request on hover. The request can disclose the viewer's IP address, user agent, access timing, and access to internal-network hosts, but no script survives sanitization and the attacker cannot read the response. This issue is fixed in version 2.25.3.

Summary

Loofah::HTML5::Scrub.scrub_css rejected any property whose top-level Crass tokens included a :url or :bad_url, and left url out of the allowed CSS functions. Both checks ran only against the first parse of the style attribute. When scrub_css rebuilt each property it emitted identifiers with their CSS escapes decoded, so parentheses written as escapes — url\28 …\29 — hid the url token from the checks, and the rebuilt value was re-parsed into a live url() that nothing checked again. An escaped property name, six-digit escapes, escaped comments and escaped quotes all reached the same place.

Impact

A sanitized style attribute could make a viewer's browser issue a request to an attacker-chosen URL. list-style fetches on render with no interaction, cursor on hover. That leaks the viewer's IP address, user agent and access timing, and can reach hosts routable only from the viewer's network. No script survives sanitization and the attacker cannot read the response, so the request itself is the whole impact.

Applications that sanitize user-supplied HTML with Loofah's default allowlist and display it to other people are affected.

Mitigation

Upgrade to Loofah >= 2.25.3.

Credit

Reported by NotAFlightRisk through GitHub's private vulnerability reporting.

Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.3, Loofah::HTML5::Scrub.scrub_css checks only the first parse of a style attribute, then rebuilds CSS with identifier escapes decoded so escaped parentheses, property names, comments, or quotes can become a live url() without a second safety check. An attacker can use list-style to make a viewer's browser request an attacker-chosen URL when content renders, or cursor to make the request on hover. The request can disclose the viewer's IP address, user agent, access timing, and access to internal-network hosts, but no script survives sanitization and the attacker cannot read the response. This issue is fixed in version 2.25.3.

Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.3, Loofah::HTML5::Scrub.scrub_css checks only the first parse of a style attribute, then rebuilds CSS with identifier escapes decoded so escaped parentheses, property names, comments, or quotes can become a live url() without a second safety check. An attacker can use list-style to make a viewer's browser request an attacker-chosen URL when content renders, or cursor to make the request on hover. The request can disclose the viewer's IP address, user agent, access timing, and access to internal-network hosts, but no script survives sanitization and the attacker cannot read the response. This issue is fixed in version 2.25.3.

Summary

Loofah::HTML5::Scrub.scrub_css rejected any property whose top-level Crass tokens included a :url or :bad_url, and left url out of the allowed CSS functions. Both checks ran only against the first parse of the style attribute. When scrub_css rebuilt each property it emitted identifiers with their CSS escapes decoded, so parentheses written as escapes — url\28 …\29 — hid the url token from the checks, and the rebuilt value was re-parsed into a live url() that nothing checked again. An escaped property name, six-digit escapes, escaped comments and escaped quotes all reached the same place.

Impact

A sanitized style attribute could make a viewer's browser issue a request to an attacker-chosen URL. list-style fetches on render with no interaction, cursor on hover. That leaks the viewer's IP address, user agent and access timing, and can reach hosts routable only from the viewer's network. No script survives sanitization and the attacker cannot read the response, so the request itself is the whole impact.

Applications that sanitize user-supplied HTML with Loofah's default allowlist and display it to other people are affected.

Mitigation

Upgrade to Loofah >= 2.25.3.

Credit

Reported by NotAFlightRisk through GitHub's private vulnerability reporting.

Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.3, Loofah::HTML5::Scrub.scrub_css checks only the first parse of a style attribute, then rebuilds CSS with identifier escapes decoded so escaped parentheses, property names, comments, or quotes can become a live url() without a second safety check. An attacker can use list-style to make a viewer's browser request an attacker-chosen URL when content renders, or cursor to make the request on hover. The request can disclose the viewer's IP address, user agent, access timing, and access to internal-network hosts, but no script survives sanitization and the attacker cannot read the response. This issue is fixed in version 2.25.3.