Quick answers
- Is CVE-2026-123022 exploited in the wild?
- No. Junglewise Threat Intelligence has no report of CVE-2026-123022 being exploited in the wild.
- How severe is CVE-2026-123022?
- High severity, with a CVSS score of 7.3, published 5 November 2026.
- Which products and versions are affected?
- Agentscope-Ai AgentScope up to 2.0.11.
- Is there a fix for CVE-2026-123022?
- No fixed version is recorded yet. Bind the service to 127.0.0.1 instead of 0.0.0.0 to restrict network access, and block POST /workspace/mcp at the network or reverse-proxy layer until patched. Docker or E2B workspace backends provide partial containment by sandboxing the spawned process.</anigation> <parameter name="timeline">[{"kind": "disclosed", "at": "2026-07-08", "note": "Privately reported via GitHub Private Vulnerability Reporting"}, {"kind": "other", "at": "2026-08-30", "note": "Advisory in Triage with no maintainer response; public bug report filed"}, {"kind": "advisory", "at": "2026-11-05", "note": "CVE-2026-123022 published"}] If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
- How is CVE-2026-123022 attacked?
- It is attacked over the network, in a software library.
- Am I affected by CVE-2026-123022?
- Check whether you run Agentscope-Ai AgentScope at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.
How to fix
Bind the service to 127.0.0.1 instead of 0.0.0.0 to restrict network access, and block POST /workspace/mcp at the network or reverse-proxy layer until patched. Docker or E2B workspace backends provide partial containment by sandboxing the spawned process.</anigation> <parameter name="timeline">[{"kind": "disclosed", "at": "2026-07-08", "note": "Privately reported via GitHub Private Vulnerability Reporting"}, {"kind": "other", "at": "2026-08-30", "note": "Advisory in Triage with no maintainer response; public bug report filed"}, {"kind": "advisory", "at": "2026-11-05", "note": "CVE-2026-123022 published"}]
Technical details
The vulnerability is an OS command injection (CWE-78) in the MCP Workspace API's POST /workspace/mcp endpoint. The StdioMCPConfig.command parameter is passed unsanitized to subprocess.Popen without validation or allowlisting, and the endpoint enforces no authentication—the X-User-ID header is spoofable. An unauthenticated remote attacker can register a malicious command that executes immediately when the MCP configuration is saved, gaining code execution as the AgentScope process user.
Affected products
- agentscope-ai AgentScope up to 2.0.11