Resource Abuse Vulnerability in AutoGPT Workflow Automation Platform
CVE-2025-32392

8.7HIGH

Key Information:

Vendor

Significant-gravitas

Status
Autogpt
Vendor
CVE Published:
18 June 2026

What is CVE-2025-32392?

AutoGPT, a platform designed for workflow automation of continuous AI agents, has a vulnerability in its LoopVideoBlock feature prior to version 0.6.63. This vulnerability allows users to manipulate video processing parameters, specifically the number of loops, without any defined limits. If an attacker exploits this by setting an excessively high number of loops, it generates an oversized video file that can exhaust system disk space. This leads to denial of service conditions, as the application may become unresponsive or crash due to resource depletion. Version 0.6.63 addresses this critical issue by implementing necessary constraints on resource allocation during video processing.

Affected Version(s)

AutoGPT < 0.6.63

References

https://github.com/Significant-Gravitas/AutoGPT/security/...
x_refsource_CONFIRM

CVSS V4

Score:
8.7
Severity:
HIGH
Confidentiality:
None
Integrity:
None
Availability:
High
Attack Vector:
Network
Attack Complexity:
Low
Attack Required:
None
Privileges Required:
Undefined
User Interaction:
None

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.