Quick answers
- Is CVE-2026-51856 exploited in the wild?
- No. Junglewise Threat Intelligence has no report of CVE-2026-51856 being exploited in the wild.
- How severe is CVE-2026-51856?
- Informational severity, published 30 September 2026.
- Which products and versions are affected?
- Agentscope-Ai AgentScope 1.0.18, 1.0.19.
- Is there a fix for CVE-2026-51856?
- No fixed version is recorded yet. Check the vendor's advisory in the references below. 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-2026-51856 attacked?
- It is attacked over the network, in a software library.
- Am I affected by CVE-2026-51856?
- 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.
Technical details
A remote WebSocket user connected to RealtimeAgent can send prompts that cause the AI model to invoke the execute_python_code tool without additional approval or isolation. The vulnerability exists in the call chain RealtimeAgent._acting → Toolkit.call_tool_function → execute_python_code, where the tool invocation lacks sandboxing or user approval gates. CWE-284 (Incorrect Access Control); requires network access and the execute_python_code tool to be registered in the session tool list.
Affected products
- agentscope-ai AgentScope 1.0.18, 1.0.19
Timeline
- 2026-05-05: disclosed
- 2026-09-30: advisory