What happened
A vulnerability designated CVE-2026-123023 has been identified in agentscope-ai AgentScope versions up to 2.0.8. This flaw impacts the function check_permissions within the src/agentscope/tool/_builtin/_bash.py component of the Bash Tool. The security weakness allows manipulation that can lead to incorrect authorization outcomes. Exploitation of this flaw has been confirmed to be active in the wild, with attackers potentially leveraging an exposed network service to carry out such exploits without any prior access or user interaction. The attack path involves directly accessing an exposed service associated with the affected component, enabling remote attackers to exploit the flaw without prior authentication or privilege escalation. The consequences include partial impact on confidentiality, integrity, and availability due to incorrect permissions handling, with this vulnerability being classified as high severity (CVSS score of 7.3). The flaw relates to weaknesses CWE-285 and CWE-863, related to security permissions and access control issues. Reported by the NVD, the vulnerability affects AgentScope versions 2.0.0, 2.0.1, 2.0.2, and 2.0.3. Currently, the exploit is actively exploited, and the vulnerability was exploited before it was publicly disclosed, making it a zero-day in this context. Patch status information has not yet been provided, and affected organisations should consult the vendor's advisory for details regarding available fixed versions.
Who is affected
Organisations running affected deployments of agentscope-ai AgentScope versions up to 2.0.8 are impacted by this vulnerability. Any organization whose environment includes network services powered by these versions should consider the risk high due to active exploitation and the potential for remote attack without user involvement.
Recommended actions
- Apply updates or patches provided by the vendor; check the vendor's advisory for fixed versions.
- Restrict access to the affected network services to trusted sources and implement network segmentation where possible.
- Monitor network traffic for unusual activity targeting the affected services and look for indicators of exploitation.
