Hub/AI Security/Codex
Codex
OpenAI · Official site
7 Yes0 Partial0 No3 Unknown
| Feature | Status | Notes | Source | Checked |
|---|---|---|---|---|
| Tool-call approval prompts | Yes | Approval policy (e.g. on-request); the untrusted policy was retired | Source | 2026-11-04 |
| Per-tool allow / deny | Yes | Per-server enabled_tools / disabled_tools and default_tools_approval_mode | Source | 2026-11-04 |
| Sandboxing | Yes | OS-enforced sandbox (read-only / workspace-write); network off by default | Source | 2026-11-04 |
| OAuth for remote MCP | Yes | OAuth incl. CIMD and dynamic client registration | Source | 2026-11-04 |
| Verified marketplace listings | Unknown | Not found in official docs checked 2026-11-04 | — | 2026-11-04 |
| Auto-install / auto-update control | Unknown | Postman's installer wrote a trusted hook hash into Codex config (see advisories) | — | 2026-11-04 |
| Admin allowlists | Yes | Managed config can pin an MCP allowlist (name + identity must match) | Source | 2026-11-04 |
| Audit logs | Yes | Compliance API audit events | Source | 2026-11-04 |
| Data retention / training opt-out | Unknown | Not found in official docs checked 2026-11-04 | — | 2026-11-04 |
| Security disclosure page | Yes | Advisories published on openai/codex | Source | 2026-11-04 |
Advisories (3)
Postman Desktop auto-installs Codex / Claude Code plugins and pre-trusts its own startup hook
Postman Desktop 12.33.4–12.35.1
OpenAI Codex CLI enables code execution through malicious MCP (Model Context Protocol) configuration files
Codex CLI · CVE-2025-61260 · GHSA-xrxf-jgv3-qmrm
Sandbox bypass due to bug in path configuration logic
Codex CLI · CVE-2025-59532 · GHSA-w5fx-fh39-j5rw
Across the hub
Weekly, free
What changed in AI integrations
One short email when MCP clients, plugin ecosystems, IDE integrations, AI plan prices or security advisories change. Every item is sourced.