CVE-2023-50009

What it actually does

Buffer overflow in ff_gaussian_blur_8 within libavfilter/edge_template.c allows local attacker to execute arbitrary code.

Why it matters to the business

Systems using FFmpeg for video processing face risk of local code execution leading to system compromise; prevalence across media pipelines increases exposure surface.

How an attack unfolds

A local attacker supplies crafted input that triggers the vulnerable gaussian blur function during edge template processing in FFmpeg, overflowing the buffer to hijack control flow and run attacker-supplied code.

How to fix it

Upgrade org/ffmpeg.git to commit 083443d67cb159ce469e5d902346b8d0c2cd1c93 or ffmpeg/ffmpeg to commit 162b4c60c8f72be2e93b759f3b1e14652b70b3ba; rebuild and redeploy all affected binaries.

How to confirm the fix worked

Confirm FFmpeg version reports a fixed commit hash; run regression tests exercising edge_template.c filters on sample media; monitor for absence of crashes in blur operations.

Typical attack path

1. Identify FFmpeg invocation that uses edge detection filters
2. Prepare malicious media file triggering ff_gaussian_blur_8
3. Execute FFmpeg locally with the crafted input
4. Achieve arbitrary code execution via buffer overflow

How to detect it

Review process execution logs for FFmpeg invocations involving edge filters; inspect core dumps or crash reports referencing libavfilter/edge_template.c; search application logs for abnormal termination during gaussian blur operations.

What your security team should do

• Apply the listed commit fixes immediately across all FFmpeg deployments
• Inventory media processing pipelines for FFmpeg usage
• Restrict local execution privileges on systems running FFmpeg
• Monitor EPSS and exploit maturity for any future changes