Amazon partners with OpenAI to let users of Amazon's DSP run ads in ChatGPT in a pilot program limited to the US; ChatGPT Ads hit $1B in ARR last month40m ago D-Matrix Connects Raptor XPUs to NVIDIA AI Factories via NVLink Fusion1h ago Source: Meta AI researcher Andrew Tulloch, who came over from Thinking Machine Lab as one of the highest paid employees last year, is leaving the company (Reed Albergotti/Semafor)51m ago Mother Horrified When Instagram’s AI Scours Social Media for Invasive Information About Her Young Daughters56m ago Mach Industries, which makes unmanned military vehicles and weapons, raised $600M in an extension to its $300M Series C, doubling its valuation to $3.7B1h ago Salesforce Unveils Six-Capability Trusted AI Harness for Enterprises1h ago US closes $1.9bn loan to restart the Iowa nuclear plant powering Google1h ago OpenAI Gives US Agencies 50% Off Models, Ending $1 Per Year Deal1h ago 47,000 job listings reveal the engineering roles that AI is creating1h ago Anthropic's Alignment Science lead says there is a ">10%" chance AI could kill all humans within the next decade and is worried about recursive self-improvement (Evan Hubinger/@evanhub)1h ago Ukraine says AI is moving beyond drones and into strike missiles for the first time1h ago Meta Announces Muse AI Agent for Personal Tasks and Organization1h ago Celero raises $275m to bring coherent optics inside AI data centres1h ago Exclusive: Cfo.ai launches an agentic CFO for business founders1h ago

Amazon partners with OpenAI to let users of Amazon's DSP run ads in ChatGPT in a pilot program limited to the US; ChatGPT Ads hit $1B in ARR last month40m ago D-Matrix Connects Raptor XPUs to NVIDIA AI Factories via NVLink Fusion1h ago Source: Meta AI researcher Andrew Tulloch, who came over from Thinking Machine Lab as one of the highest paid employees last year, is leaving the company (Reed Albergotti/Semafor)51m ago Mother Horrified When Instagram’s AI Scours Social Media for Invasive Information About Her Young Daughters56m ago Mach Industries, which makes unmanned military vehicles and weapons, raised $600M in an extension to its $300M Series C, doubling its valuation to $3.7B1h ago Salesforce Unveils Six-Capability Trusted AI Harness for Enterprises1h ago US closes $1.9bn loan to restart the Iowa nuclear plant powering Google1h ago OpenAI Gives US Agencies 50% Off Models, Ending $1 Per Year Deal1h ago 47,000 job listings reveal the engineering roles that AI is creating1h ago Anthropic's Alignment Science lead says there is a ">10%" chance AI could kill all humans within the next decade and is worried about recursive self-improvement (Evan Hubinger/@evanhub)1h ago Ukraine says AI is moving beyond drones and into strike missiles for the first time1h ago Meta Announces Muse AI Agent for Personal Tasks and Organization1h ago Celero raises $275m to bring coherent optics inside AI data centres1h ago Exclusive: Cfo.ai launches an agentic CFO for business founders1h ago

Cybersecurity Agents6 min reading time

DeepSeek Harness Flaw Let AI Agents Disable Their Own File Sandbox Without Approval

The Hacker NewsPublished 1 day ago

DeepSeek Harness had a critical security flaw allowing AI agents to disable their sandbox protections via a single command, enabling unrestricted file access. The vulnerability, fixed on August 27, was rated 9.4/10 and tracked as CVE-2026-82533.

#ai-agent-security #cve-2026-82533 #deepseek-harness #ox-research #sandbox-security #vulncheck

More in Cybersecurity

Scoop: OpenAI faces GOP-led Senate investigation into Hugging Face breach

Covered by 3 sources5h ago

Cybersecurity6 min read

Anthropic Discloses Fourth AI Hacking Incident Involving Claude Opus 4.6

Covered by 2 sources7h ago

Cybersecurity5 min read

Every AI Incident Has Two Timelines. We Default To One

Forbes3h ago