LeRobot unsafe deserialization RCE (CVE-2026-25874)
Vulnerability
Summary
Hide ▲
Show ▼
The CVE-2026-25874 unsafe deserialization flaw in Hugging Face LeRobot enables unauthenticated remote code execution against network-reachable deployments. The bug sits in the async inference path, where attacker-controlled pickle data is accepted over unauthenticated gRPC channels and can reach the server or client. The issue was validated against version 0.4.3, remains unpatched, and a fix is planned for 0.6.0. Successful exploitation could lead to host compromise, connected robot impact, and theft of sensitive data.
Related Happenings
BeyondTrust Remote Support and Privileged Remote Access CVE-2026-1731 active exploitation wave
Exploitation Wave
First: 12.02.2026 23:34
Last: 12.02.2026 23:34
Sources 1
About this happening:
**CVE-2026-1731** in **BeyondTrust Remote Support** and **Privileged Remote Access** is now seeing **first in-the-wild exploitation**, putting exposed appliances at risk of remote...
BeyondTrust Remote Support and Privileged Remote Access CVE-2026-1731 active exploitation wave
Exploitation WaveAbout this happening: **CVE-2026-1731** in **BeyondTrust Remote Support** and **Privileged Remote Access** is now seeing **first in-the-wild exploitation**, putting exposed appliances at risk of remote...
CISA KEV remediation deadline for SolarWinds WHD CVE-2025-40551
Public Sector Action
First: 04.02.2026 07:50
Last: 04.02.2026 07:50
Sources 1
About this happening:
**CISA** added **CVE-2025-40551** in **SolarWinds Web Help Desk** to the **KEV catalog** and imposed **federal remediation deadlines**, turning a newly exploited flaw into a compl...
CISA KEV remediation deadline for SolarWinds WHD CVE-2025-40551
Public Sector ActionAbout this happening: **CISA** added **CVE-2025-40551** in **SolarWinds Web Help Desk** to the **KEV catalog** and imposed **federal remediation deadlines**, turning a newly exploited flaw into a compl...
OpenAI ShadowLeak mitigation for ChatGPT
Advisory/Mitigation
First: 19.09.2025 22:07
Last: 19.09.2025 22:07
Sources 1
About this happening:
**OpenAI** acknowledged and fixed **ShadowLeak**, ending a stealthy **ChatGPT** email-exfiltration path in the studied **Gmail** integration. The mitigation mattered because the f...
OpenAI ShadowLeak mitigation for ChatGPT
Advisory/MitigationAbout this happening: **OpenAI** acknowledged and fixed **ShadowLeak**, ending a stealthy **ChatGPT** email-exfiltration path in the studied **Gmail** integration. The mitigation mattered because the f...
Timeline
-
28.04.2026 14:18 2 articles · 29d ago
Public disclosure of CVE-2026-25874 in Hugging Face LeRobot
Initial DisclosureResearchers disclose CVE-2026-25874 in Hugging Face LeRobot, an open-source robotics platform, describing an unsafe deserialization flaw in the async inference pipeline where pickle.loads() processes unauthenticated gRPC data without TLS in the PolicyServer and robot client components, enabling unauthenticated remote code execution on network-reachable deployments. The issue is validated against LeRobot version 0.4.3, remains unpatched, and a fix is planned for version 0.6.0.
Show sources
- Critical Unpatched Flaw Leaves Hugging Face LeRobot Open to Unauthenticated RCE — thehackernews.com — 28.04.2026 14:18
- Critical Unpatched Flaw Leaves Hugging Face LeRobot Open to Unauthenticated RCE — thehackernews.com — 28.04.2026 14:18