CVE-2026-33901

What it actually does

ImageMagick versions earlier than 7.1.2-19 and 6.9.13-44 contain a heap buffer overflow in the MVG image decoder. When ImageMagick parses a specially crafted MVG image, the decoder can write outside the bounds of an allocated heap buffer. The assessed impact is primarily availability: a vulnerable ImageMagick process such as magick, convert, identify, mogrify, or an application invoking the ImageMagick library can crash or be terminated while processing attacker-supplied image content. The published CVSS vector indicates remote, unauthenticated, low-complexity exploitation with no required user interaction when image processing is exposed through an automated service, but does not assess confidentiality or integrity impact.

Why it matters to the business

Organizations that allow users, partners, or automated workflows to submit images for conversion, resizing, thumbnailing, metadata extraction, PDF/image rendering, or media processing may be exposed to service crashes from crafted image files. This can disrupt web applications, content management systems, document pipelines, support portals, CI/build systems, or backend batch jobs that depend on ImageMagick. The EPSS score is currently low and stable over 30 days, the vulnerability is not listed in CISA KEV, and no mature public exploit is indicated, so urgency is moderate unless ImageMagick processes untrusted uploads in internet-facing or business-critical workflows.

How an attack unfolds

An attacker first identifies an application or service that accepts image or document uploads and uses ImageMagick to inspect, transform, resize, convert, or generate thumbnails. The attacker then submits a crafted MVG image, or a file that causes ImageMagick to invoke the MVG decoder based on file content, extension, or configured delegates. When the vulnerable decoder processes the file, a heap buffer overflow causes an out-of-bounds write. In the documented impact model, this corrupts process memory and can crash the ImageMagick command, worker process, or parent application pipeline. Repeated submissions could be used to create a denial-of-service condition against upload handling, media processing queues, or dependent application features.

How to fix it

Upgrade ImageMagick to a fixed release: 7.1.2-19 or later for the 7.x branch, or 6.9.13-44 or later for the 6.x branch. Use operating-system vendor packages where available, because distribution backports may carry the fix without changing to the upstream version string. Rebuild and redeploy any containers, language bindings, application images, serverless layers, or bundled binaries that include ImageMagick. If immediate upgrade is not possible, reduce exposure by blocking or disabling MVG processing where operationally feasible, restricting accepted file types to a strict allowlist, validating file magic and decoded format rather than trusting extensions, processing untrusted images in a sandboxed low-privilege environment, applying CPU/memory/time limits to image conversion workers, and isolating media-processing queues from core application services.

How to confirm the fix worked

Confirm the deployed ImageMagick version reports 7.1.2-19 or later, 6.9.13-44 or later, or a vendor package that documents a backported fix for this issue. Check all execution locations, including application containers, CI images, background workers, web servers, developer-provided static binaries, and language runtime dependencies that shell out to ImageMagick. Run `magick -version`, `convert -version`, and package-manager queries where applicable. Validate that applications no longer accept MVG input unless explicitly required, and confirm file upload controls reject unexpected vector image formats. In staging, test image-processing workflows with representative uploads and ensure no ImageMagick worker crashes, segmentation faults, sanitizer findings, or abnormal exits occur. Review container and host inventories to ensure no older ImageMagick 6.x or 7.x packages remain reachable by application code.

Typical attack path

1. Identify an application endpoint, API, queue, or workflow that accepts image-like files or documents for server-side processing.
2. Confirm or infer that ImageMagick is used by observing output formats, metadata behavior, error messages, package fingerprints, or application behavior.
3. Submit a crafted MVG image or polyglot file designed to reach ImageMagick's MVG decoder.
4. Trigger processing through thumbnail generation, conversion, metadata extraction, resizing, or validation.
5. Cause a heap out-of-bounds write during MVG decoding, leading to a crash or abnormal termination of the ImageMagick process or the calling worker.
6. Repeat submissions if the goal is to exhaust processing capacity or repeatedly disrupt dependent application functionality.

How to detect it

To assess possible prior exploitation, review application upload logs, API gateway logs, web server logs, media-processing queue records, worker logs, crash telemetry, and operating-system logs for abnormal ImageMagick failures after processing user-supplied files. Focus on segmentation faults, heap corruption messages, sanitizer reports, core dumps, repeated worker restarts, and conversion failures involving MVG files or files whose content appears to be MVG despite another extension. Search for invocations of ImageMagick utilities that exited with signals such as SIGSEGV, SIGABRT, or nonzero exit codes immediately after processing uploads. On Linux systems, review syslog, journald, kernel logs, coredumpctl, container runtime logs, Kubernetes pod restart history, and application APM traces. Because no public mature exploit or known threat actor activity is indicated, there are no reliable attacker-specific network indicators; detection should emphasize crash artifacts, suspicious file submissions, and repeated processing failures from the same source.

• AddressSanitizer: heap-buffer-overflow
• heap-buffer-overflow
• heap corruption
• malloc(): corrupted top size
• double free or corruption
• corrupted size vs. prev_size
• Segmentation fault
• SIGSEGV
• SIGABRT
• coders/mvg.c
• MVG
• image/x-mvg
• push graphic-context
• pop graphic-context
• viewbox
• convert: no decode delegate for this image format `MVG'
• magick: no decode delegate for this image format `MVG'
• process name: magick
• process name: convert
• process name: identify
• process name: mogrify

What your security team should do

• Inventory all applications, containers, hosts, and build images that include ImageMagick or call ImageMagick utilities such as magick, convert, identify, mogrify, composite, or montage.
• Prioritize remediation for internet-facing upload handlers, public APIs, media-processing services, document conversion systems, and any workflow that processes untrusted files automatically.
• Upgrade to ImageMagick 7.1.2-19 or later, or 6.9.13-44 or later, including rebuilt downstream containers and application images.
• If patching is delayed, disable MVG decoding where possible and enforce strict upload allowlists, sandboxing, resource limits, and queue isolation for image processing.
• Search historical logs for unexplained ImageMagick crashes, worker restarts, segmentation faults, or repeated failed processing attempts involving MVG-like files.
• Add monitoring for abnormal termination of ImageMagick processes and for upload patterns that repeatedly trigger conversion failures.
• Track vendor advisories for distribution-specific fixed package versions, because some Linux distributions may backport the patch into older version numbers.