AI Science Tracker

FFmpeg MLZ decoder uninitialized-memory disclosure

FFmpeg MPEG-4 ALS MLZ decoder · AI involvement attributed to AISLE AI-native vulnerability discovery engine (model/version not disclosed)

Finding and attribution

Out-of-range MLZ dictionary codes can copy uninitialized heap bytes into decoded audio output from a crafted MPEG-4 ALS file.

Product: FFmpeg MPEG-4 ALS MLZ decoder. AI system: AISLE AI-native vulnerability discovery engine (model/version not disclosed). Team / organization: AISLE.

CVE identifiers: CVE-2026-125103.

AI and human contribution (source account): AISLE’s CVE Discoveries index states that the listed CVEs were discovered by its AI-native engine; this individual AISLE case page says “Discovered by AISLE” and credits Joshua Rogers (AISLE Research) as finder. The source does not identify the model version or exact human/AI division of analysis.

Human involvement is only described where the source account establishes it; unknown involvement does not mean none. Vendor confirmation of a flaw does not confirm every claim about AI contribution.

Discovery context: Public CVE disclosure and AISLE’s AI-discovery attribution are dated 2026-11-09. The cited sources establish the assigned vulnerability and fix, not exploitation in the wild or zero-day status.

Remediation status: fix documented. Fixed in FFmpeg 9.0.3. The published claim is uninitialized-data disclosure to the decoded output; it does not establish disclosure of arbitrary process memory to a remote attacker. Confirm the distribution-specific backport before relying on a version number.

A source-documented fix does not establish installed patches or current safety. Source checking does not independently reproduce the finding or establish observed exploitation.

Disclosed 9 Nov 2026. AI attribution dated 9 Nov 2026. Sources checked 15 Nov 2026; these are separate dates, not a live installation check.

Recorded CVSS: Medium · highest recorded score 4.8. Documented technical effects: data disclosure.

Severity, consequences and exploitation evidence

Case grouping uses the highest recorded CVSS score, with its version-specific severity band. It is descriptive, not a combined or editorial risk index. All recorded assessments remain below; scores from different versions and sources may disagree. CVSS specifications define version-specific scoring and severity bands.

CVE-2026-125103: data disclosure.
Prerequisites / affected scope: An affected FFmpeg decoder processes a crafted MPEG-4 ALS file with malformed dictionary codes, copying uninitialized heap bytes into decoded audio output..
Exploitation evidence: Not established by the cited CNA description; no independent exploitation assessment made.. Impact source · checked 16 Nov 2026.

Documented fix: Fixed in FFmpeg 9.0.3. The published claim is uninitialized-data disclosure to the decoded output; it does not establish disclosure of arbitrary process memory to a remote attacker. Confirm the distribution-specific backport before relying on a version number.

Original evidence

Download this source record (JSON)

Open this case in the interactive tracker · Browse all static security cases