Memory exhaustion and SSRF via external references in Jellyfin SVG images
Key facts
- Attack vector
- Network
- Privileges required
- Low
- User interaction
- None
- Attack complexity
- Low
- Impact type
- —
- Confidentiality · integrity · availability
- High · None · High
- Fix status
- Fixed
Attack profile from CVSS 3.1 · GHSA-cf3c-8m59-2vhx. Fix status from GitHub.
Summary
Jellyfin server-side SVG rendering with Svg.Skia before 12.0 resolves external references in image elements, so a crafted SVG can read arbitrary local paths and fetch arbitrary URLs. An authenticated user with low privileges can exhaust server memory by referencing an endless file, and can fetch server-side URLs or local files whose content decodes as an image and is then downloadable.
Written with Claude Haiku 5.5 from the sources below · 2026-11-12
Description from the GitHub advisory
### Impact Jellyfin rasterizes SVG images server-side through Svg.Skia (SkiaEncoder.GetBitmapFromSvg and SkiaEncoder.GetImageSize). The SVG parser is invoked with library defaults, which resolve external references in <image xlink:href="...">. A crafted SVG therefore causes the server to read arbitrary local paths and fetch arbitrary URLs while rendering. This has two consequences. 1. Denial of service through memory exhaustion. A reference to an endless or very large file such as file:///dev/urandom is read into memory without a size limit. The read fails silently: no exception is raised, the render produces a transparent image, and the allocated memory is retained by the managed heap rather than returned. Measured against an 8 GiB limit, a 16x16 SVG produced a 4.1 GiB peak working set with 3.8 GiB still held after the load completed; under a 1.5 GiB limit it consumed 970 MiB. The consumption scales to the memory available to the process. Because the stored SVG is re-rasterized on subsequent image requests, a single upload provides a repeatable trigger without re-uploading. 2. Server-side request forgery and local file disclosure. http:// and https:// references are fetched by the server. Since a resized SVG is rasterized into the requested output format, the fetched response is rendered into an image the requester can then download — responses that decode as an image can be exfiltrated, not merely triggered. file:// references likewise expose local files that decode as an image. Both are reachable by: - Any authenticated user, by setting a profile image (POST /UserImage). SVG is an accepted image type for user images. - Library import, where an administrator is induced to import media accompanied by a crafted poster.svg. The local image providers pick up .svg artwork. This path requires an administrator, but the administrator need not author the file. ### Patches _Has the problem been patched? What versions should users upgrade to?_ ### Workarounds _Is there a way for users to fix or remediate the vulnerability without upgrading?_ ### References _Are there any links users can visit to find out more?_
Why it matters now
SSVC decision points
Not available: CISA has not published SSVC decision points for this record.
Affected products
| Vendor | Product | Affected versions | Fixed in | Status | Source |
|---|---|---|---|---|---|
| — | jellyfin | < 12.0 | 12.0 | Patch available | GHSA-cf3c-8m59-2vhx |
- jellyfin Vendor— Affected < 12.0 Fixed in 12.0 Patch available GHSA-cf3c-8m59-2vhx
Remediation
Upgrade Jellyfin to version 12.0, which fixes the issue. The sources list no fixed version for Jellyfin releases before 12.0 other than the one stated above.
Written with Claude Haiku 5.5 from the sources below · 2026-11-12
- Fix · GHSA-cf3c-8m59-2vhx jellyfin Fixed in 12.0 Advisory →
Timeline
- 2026-11-12 Threatist record created from GHSA-cf3c-8m59-2vhx
References
Sources
This product uses the NVD API but is not endorsed or certified by the NVD. Sources and licences