CVE-2025-9951: FFmpeg JPEG2000 RCE Vulnerability

CVE-2025-9951 Overview

CVE-2025-9951 is a heap-buffer-overflow write vulnerability in the jpeg2000dec decoder of FFmpeg. The flaw resides in the parsing of the channel definition (cdef) atom within JPEG2000 media files. An attacker who supplies a crafted JPEG2000 file can trigger an out-of-bounds heap write during decoding. Successful exploitation can lead to remote code execution or denial of service in any application that uses FFmpeg to decode untrusted media. The vulnerability is classified under CWE-122: Heap-based Buffer Overflow.

Critical Impact
A malicious JPEG2000 file processed by FFmpeg can corrupt heap memory and lead to arbitrary code execution or crashes in downstream applications, including media servers, transcoders, and desktop players.

Affected Products

FFmpeg jpeg2000dec decoder component
Applications and services that link against FFmpeg libraries (libavcodec) for JPEG2000 decoding
Media pipelines, transcoders, and players that accept untrusted JPEG2000 input

Discovery Timeline

2025-09-09 - CVE-2025-9951 published to NVD
2026-04-15 - Last updated in NVD database

Technical Details for CVE-2025-9951

Vulnerability Analysis

The vulnerability is a heap-based buffer overflow write that occurs while FFmpeg parses the channel definition (cdef) atom inside a JPEG2000 image. The cdef atom describes how component channels map to color or alpha channels in the codestream. When the decoder processes attacker-controlled channel definition values without properly enforcing bounds, it writes past the end of an allocated heap buffer. Heap corruption of this nature can overwrite adjacent structures and function pointers, enabling control flow hijacking or process crashes. The exploit vector is network-reachable because FFmpeg is widely used in server-side transcoding and client-side playback pipelines that ingest remote files. Exploitation does not require user interaction once a vulnerable application processes the malicious file.

Root Cause

The root cause is insufficient validation of channel definition fields parsed from the cdef atom. The decoder relies on values from the input file to index or size heap operations without verifying them against the actual allocated buffer. This is a classic [CWE-122] heap overflow pattern in media parsing code.

Attack Vector

An attacker crafts a JPEG2000 file with malformed cdef atom fields and delivers it to a target that uses FFmpeg for decoding. Delivery channels include web uploads, email attachments, streaming inputs, and automated media transcoding services. Refer to the GitHub Security Advisory GHSA-39q3-f8jq-v6mg for detailed technical analysis.

Detection Methods for CVE-2025-9951

Indicators of Compromise

Unexpected crashes or segmentation faults in processes linking libavcodec when handling JPEG2000 (.jp2, .j2k, .jpf) files
Heap corruption signatures in core dumps from FFmpeg-based transcoders or players
AddressSanitizer or glibc heap consistency errors logged when decoding user-supplied media

Detection Strategies

Inventory all systems and applications that bundle or link against FFmpeg and identify versions exposing the jpeg2000dec decoder
Scan inbound media uploads for JPEG2000 files and run them through a sandboxed decoder with sanitizers enabled
Hunt for child process anomalies originating from media transcoding workers, such as unexpected shell or network activity

Monitoring Recommendations

Monitor media-processing services for abnormal memory usage, crashes, and restart loops
Log and alert on FFmpeg invocations that terminate with SIGSEGV or SIGABRT during JPEG2000 handling
Capture telemetry on outbound connections from transcoding hosts to detect post-exploitation activity

How to Mitigate CVE-2025-9951

Immediate Actions Required

Apply the patched FFmpeg release as soon as it is available from the upstream project and downstream distributions
Disable or block JPEG2000 decoding in services that do not require it by removing the jpeg2000 decoder from build configurations or runtime allowlists
Process untrusted media in sandboxed environments with restricted privileges and no outbound network access

Patch Information

Review the GitHub Security Advisory GHSA-39q3-f8jq-v6mg for fix commits and patched version information. Rebuild and redeploy any application that statically links FFmpeg after updating the library.

Workarounds

Strip or reject JPEG2000 inputs at the application boundary until patched binaries are deployed
Run FFmpeg under hardened sandboxes such as seccomp, bubblewrap, or container isolation with read-only mounts
Enforce file type allowlists and validate media MIME types before invoking the decoder

bash

# Example: disable JPEG2000 decoder when building FFmpeg from source
./configure --disable-decoder=jpeg2000 --disable-parser=jpeg2000
make && make install

# Example: block JPEG2000 inputs at the FFmpeg command line
ffmpeg -f lavfi -i color=black -vcodec libx264 \
  -blocksize 0 -allowed_extensions mp4,mov,mkv -i input_media