{"id":"CVE-2026-54141","published":"2026-11-05T17:18:19.103","lastModified":"2026-11-05T17:18:19.103","description":"OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, a remote direct message received through a configured Yuanbao channel is converted into a generic channel AgentTurnRequest that uses the default tool policy without establishing an interactive approval context. Under the default Supervised autonomy policy with Node enabled, node_exec is classified as a write operation that should prompt for approval, but ApprovalGate treats missing APPROVAL_CHAT_CONTEXT as a non-interactive turn and allows execution. Because Yuanbao direct messages are open by default, an external sender who can message the connected bot can influence the agent to execute JavaScript and operating-system commands as the OpenHuman user without an approval prompt. The issue requires a configured Yuanbao channel, the default approval-requiring policy state, Node availability, and successful model selection of node_exec. This issue is fixed in version 0.57.13.","cvssScore":9,"cvssSeverity":"CRITICAL","cvssVector":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:C/C:H/I:H/A:H","cwes":["CWE-94","CWE-862"],"vendors":[],"products":[],"references":[{"url":"https://github.com/tinyhumansai/openhuman/commit/3a7180da7765af861bf44d79e025fc67dbeae0ca","tags":[]},{"url":"https://github.com/tinyhumansai/openhuman/pull/3227","tags":[]},{"url":"https://github.com/tinyhumansai/openhuman/releases/tag/v0.57.13","tags":[]},{"url":"https://github.com/tinyhumansai/openhuman/security/advisories/GHSA-gg3m-wgmq-vhmx","tags":[]}],"exploitRefs":[{"url":"https://github.com/tinyhumansai/openhuman/commit/3a7180da7765af861bf44d79e025fc67dbeae0ca","tags":[]},{"url":"https://github.com/tinyhumansai/openhuman/pull/3227","tags":[]},{"url":"https://github.com/tinyhumansai/openhuman/releases/tag/v0.57.13","tags":[]},{"url":"https://github.com/tinyhumansai/openhuman/security/advisories/GHSA-gg3m-wgmq-vhmx","tags":[]}],"hasPoc":true,"ai":{"summary":"A missing interactive approval context in OpenHuman's Yuanbao channel allows remote attackers to trigger privileged node_exec operations without user approval, leading to arbitrary code execution as the host user.","exploitability":"Exploitation requires a configured Yuanbao channel, an open direct message channel, and successful model selection of node_exec under the default Supervised policy with Node enabled; the attack is non-trivial due to AC:H but feasible for targeted threat actors.","blast_radius":"Successful exploitation grants full compromise of confidentiality, integrity, and availability as the OpenHuman user, including remote code execution and potential lateral movement from the compromised host.","remediation":"Upgrade to OpenHuman version 0.57.13 or later, which establishes the required interactive approval context for Yuanbao direct messages.","detection":"Monitor for unexpected node_exec tool calls originating from Yuanbao-channel AgentTurnRequests without a preceding APPROVAL_CHAT_CONTEXT in logs; look for process creation events of shell or script interpreters initiated by the OpenHuman process following such requests.","tags":["rce","approval-bypass","agent-hijack","yuanbao","nodejs"],"model":"garnet-coder","analyzedAt":"2026-11-06T09:46:48.904Z"}}