CVE-2026-65699

Description

AgentGPT through 1.0.0 contains an authorization bypass through user-controlled key vulnerability that allows authenticated users to attach tasks to another user's agent run by supplying a target run_id in the request body without ownership verification. The AgentCRUD.create_task and validate_task_count functions look up the target AgentRun using the client-supplied run_id without confirming the run belongs to the requesting user, enabling an attacker who obtains a valid run_id to corrupt task history, exhaust the per-run loop budget, and drive LLM costs against the victim's run.

CVSS Vector

References

• https://github.com/geo-chen/oss/blob/main/AgentGPT.md
• https://www.vulncheck.com/advisories/agentgpt-authorization-bypass-via-agent-task-creation