CVE-2026-19590
OpenAI Codex Desktop for Windows and macOS could execute attacker-controlled Git hooks because automated Git operations trusted the repository's local core.hooksPath setting. If a user opens an attacker-prepared repository whose preserved .git/config points core.hooksPath to an attacker-controlled directory, Codex can run a malicious hook while processing the repository. The hook executes outside Codex's command sandbox, without user approval, and with the user's privileges, allowing it to read, change, or delete the user's files and access other resources available to the user's account. An ordinary Git clone does not preserve the attacker-controlled repository-local configuration required for exploitation.
Scoring
- Severity
- HIGH
- CVSS base score
- 7.3
- CVSS vector
- CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:U/C:H/I:H/A:H
- EPSS probability
- 0.10%
- CWE
- CWE-427
- Published
- 2026-09-01
- Last modified
- 2026-09-03
Affected products
- OpenAI Codex Desktop
- OpenAI Codex Desktop
- OpenAI Codex Desktop
- OpenAI Codex Desktop
- OpenAI Codex Desktop (Microsoft Store package)
- OpenAI Codex Desktop (Microsoft Store package)
Weakness type
Related vulnerabilities
- CVE-2026-87530 — Uncontrolled search path element in CredentialProvider in Google Chrome on on Windows prior to...
- CVE-2026-76199 — Photoshop Desktop | Uncontrolled Search Path Element (CWE-427)
- CVE-2026-72980 — Windows Hello Security Feature Bypass Vulnerability
- CVE-2026-6958 — Acunetix 25.11.251107123 Local Privilege Escalation via wvsc.exe
- CVE-2026-45221 — Konga < 2.1.0 Privilege Escalation via Hardcoded OpenSSL Path
- CVE-2026-82649 — SiYuan before 3.8.1 Local Privilege Escalation via Uncontrolled Search Path
- CVE-2026-65093 — NVIDIA OpenShell for Linux contains a vulnerability where an attacker could cause a sandbox escape....
- CVE-2024-58377 — Nokogiri versions before 1.16.5 bundle libxml2 2.12.6, which is affected by CVE-2024-34459 in...