{"id":"CVE-2026-51857","published":"2026-09-30T21:17:11.647","lastModified":"2026-10-02T13:17:46.300","description":"In camel-ai camel 0.2.91a1, v0.2.91a2 and v0.2.91a3, CodeExecutionToolkit can run model-produced Python code through SubprocessInterpreter without an approval boundary.","cvssScore":9.8,"cvssSeverity":"CRITICAL","cvssVector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","cwes":["CWE-94"],"vendors":[],"products":[],"references":[{"url":"https://gist.github.com/Ro1ME/78606e0763520d6519897e90b305d3cd","tags":[]},{"url":"https://github.com/camel-ai/camel/issues/4037","tags":[]},{"url":"https://github.com/camel-ai/camel/issues/4037","tags":[]}],"exploitRefs":[{"url":"https://gist.github.com/Ro1ME/78606e0763520d6519897e90b305d3cd","tags":[]},{"url":"https://github.com/camel-ai/camel/issues/4037","tags":[]},{"url":"https://github.com/camel-ai/camel/issues/4037","tags":[]}],"hasPoc":true,"ai":{"summary":"This flaw allows for remote code execution through the execution of Python code via SubprocessInterpreter, leading to potential unauthorized access and control of the system.","exploitability":"Exploitation is relatively straightforward as it requires only the ability to inject Python code, which can be done through various input vectors.","blast_radius":"If exploited, the impact could be severe, potentially leading to full system compromise and data exfiltration.","remediation":"Upgrade to camel-ai camel 0.2.91a4 or later.","detection":"No reliable host or network indicator is derivable from the published description.","tags":["rce","python","subprocess","code-execution","security"],"model":"qwen2.5:7b-instruct","analyzedAt":"2026-10-03T08:44:19.992Z"}}