setup-codex-devcontainer
Warn
Audited by Socket on Sep 25, 2026
2 alerts found:
Securityx2SecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The core Codex install is coherent and uses OpenAI’s official installer, so it is not malware-like on its own. Risk comes from disproportionate container privilege grants and the optional automatic transitive skill sync, which can re-install third-party skills unattended on every start.
Confidence: 91%Severity: 74%
Securitytemplates/capability-seam-entries.json
MEDIUMSecurityMEDIUM
templates/capability-seam-entries.json
This fragment is a high-risk container configuration because it grants SYS_ADMIN and disables important sandboxing protections. It does not itself demonstrate malicious behavior, but it materially increases the potential impact of compromised or untrusted container code.
Confidence: 99%Severity: 88%
Audit Metadata