(0Day) Ultralytics parse_model Code Injection Remote Code Execution (CVE‑2026‑118618)
What It Is — A zero‑day vulnerability (CVE‑2026‑118618) in the parse_model function of Ultralytics’ AI‑model library allows a remote attacker to execute arbitrary Python code. The flaw stems from insufficient validation of a user‑supplied string that is later passed to exec(). Exploitation requires the victim to visit a malicious web page or open a crafted file, after which the attacker’s code runs with the privileges of the current user. The advisory, published by the Zero Day Initiative on 28 Oct 2026, assigns a CVSS 7.8 (High) score (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H). No public patch was available at the time of disclosure; mitigation guidance is limited to restricting interaction with the affected component.
Affected Products — Ultralytics AI/ML library (any version containing the vulnerable parse_model implementation).
Why It Matters for Trust & Control Assurance — The issue highlights a gap in secure‑development controls—specifically, the lack of input validation and code‑review safeguards that map to a single VCF control objective covering application security and secure coding practices.
Recommended Actions
- Restrict user interaction with the vulnerable component (e.g., block untrusted file uploads or external page loads).
- Apply any vendor‑released patches or updates as soon as they become available.
- Conduct an immediate code review of
parse_modeland related modules to add proper sanitisation. - Integrate static‑code analysis and runtime‑monitoring tools into the CI/CD pipeline to detect similar injection patterns.
- Document the remediation steps as audit evidence for your framework of record.