
Top AI Agents Built to Catch Malicious Code Can Be Tricked Into Running It
New research reveals 'Friendly Fire' vulnerabilities in AI coding agents, where autonomous security tools can be tricked into executing malicious code.
Jul 9, 2026·3 min read
Unpatched Backdoor in Tenda Firmware Grants Admin Access to Devices
A critical unpatched backdoor in Tenda firmware (CVE-2026-11405) exposes home and office routers to remote admin takeover. Read the full security analysis.
Jul 9, 2026·3 min read

GhostApproval Symlink Flaws Could Let Malicious Repos Run Code in AI Coding Agents
Wiz researchers discover 'GhostApproval' flaws in AI coding agents, allowing malicious repositories to gain unauthorized system access via symlink exploits.
Jul 9, 2026·3 min read
Latest in AI
- 01
- 02
- 03
- 04
Separating signal from noise in coding evaluations
OpenAI's latest audit of SWE-bench Pro uncovers systematic noise, challenging the reliability of how we measure AI's software engineering capabilities.
Jul 8, 2026·3 min read
Helping K–12 educators build practical AI skills
OpenAI and the Walton Family Foundation launch 'AI Skills Jams' to provide K-12 teachers with practical generative AI training for classroom use.
Jul 8, 2026·3 min read
Introducing GPT-Live
OpenAI unveils GPT-Live, a zero-latency voice model revolutionizing human-AI interaction through native multimodal processing and emotional intelligence.
Jul 8, 2026·3 min read
Research
From the Labs
- LabsOpenAI
Separating signal from noise in coding evaluations
Jul 8, 2026·3 min read
- LabsOpenAI
Helping K–12 educators build practical AI skills
Jul 8, 2026·3 min read
- LabsOpenAI
Introducing GPT-Live
Jul 8, 2026·3 min read
- LabsOpenAI
Australian Payments Plus moves faster with ChatGPT and Codex
Jul 7, 2026·3 min read
- LabsGoogle DeepMind
Google DeepMind and A24 announce first-of-its-kind research partnership
Jul 3, 2026·3 min read
SecurityView all →
Unpatched Backdoor in Tenda Firmware Grants Admin Access to Devices
A critical unpatched backdoor in Tenda firmware (CVE-2026-11405) exposes home and office routers to remote admin takeover. Read the full security analysis.
Jul 9, 2026·3 min read

GhostApproval Symlink Flaws Could Let Malicious Repos Run Code in AI Coding Agents
Wiz researchers discover 'GhostApproval' flaws in AI coding agents, allowing malicious repositories to gain unauthorized system access via symlink exploits.
Jul 9, 2026·3 min read

Fake 7-Zip Installers Turn Devices Into Residential Proxy Nodes
Cybersecurity researchers uncover 'Lurking Lizard,' a sophisticated threat actor turning user devices into residential proxies through fake software.
Jul 9, 2026·3 min read
Opinion & AnalysisEditorial voices
- OPINIONBy Rohan Mehta
I Used an AI to Budget My Indian Wedding. It Was Shockingly Sensible.
Jul 7, 2026·6 min read
- OPINIONBy Rohan Mehta
I Tried to Poison an AI with Lies. Here’s What I Learned.
Jul 4, 2026·6 min read
- OPINIONBy Rohan Mehta
My Guide to AI 'Weight Classes': Why I Use Three Different Models
Jul 4, 2026·6 min read
- OPINIONBy Rohan Mehta
I Built an AI Intern Without Code: My Guide to Web Automation
Jul 4, 2026·6 min read
