OpenAI Codex CLI
OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS misclassified certain PowerShell commands as safe because their command-safety parser interpreted PowerShell's stop-parsing token (--%) differently than PowerShell itself. If a user opens an attacker-prepared repository and Codex follows its instructions, Codex can run a file-writing Git command without requesting user approval. On macOS and Linux, exploitation additionally requires separately installed PowerShell Core (pwsh) to be invoked. If filesystem protections permit the write, the command can modify Codex's configuration. If Codex later loads the modified configuration, it can launch an attacker-controlled MCP server and execute code with the user's privileges, allowing it to read, change, or delete files accessible to that account. The approval bypass does not disable filesystem sandboxing; the default filesystem sandbox on macOS and Linux can prevent writes outside permitted locations.
Affected versions and mitigation
From the CVE record and the vendor's references.Affected
- >= 0.72.0, <= 0.130.0
Fixed in 0.131.0
- >= 260202.0859, <= 26.513.31313
Fixed in 26.519.22136
- >= 26.304.38, <= 26.513.40821
Fixed in 26.519.21041
- >= 26.304.38.0, <= 26.513.4821.0
Fixed in 26.519.2081.0
Mitigation
Upgrade Codex CLI to version 0.131.0 or later, Codex Desktop for macOS to application version 26.519.22136 or later, or Codex Desktop for Windows to application version 26.519.21041 (Microsoft Store package 26.519.2081.0) or later.
Workaround
Until updated, do not open or follow instructions from untrusted repositories. On macOS and Linux, keep filesystem sandboxing enabled.
References
Timeline
Days relative to CVE publicationNo public PoC found yet; not in CISA KEV.
PoC repositories
Found on GitHub.No PoC repositories found yet.