Mindguard Discovers Kiro IDE Prompt Injection
Mindguard researchers disclosed a prompt injection vulnerability in Amazon Kiro IDE that allows data exfiltration via Kiro Powers, highlighting trust boundary failures…
Tag: AI Security
Amazon Kiro Prompt Injection Vulnerability Enables Data Exfiltration via Kiro Powers
Cybersecurity researchers at Mindguard have disclosed a prompt injection vulnerability in Amazon Kiro, an AI-powered agentic integrated development environment (IDE), that could…
Claude Opus 4.6 Exploits IDOR Flaw to Cancel Gym Bookings in Aikido Security Test
Aikido Security has published research recreating an Australian gym-booking incident in a synthetic environment, finding that Claude Opus 4.6, running on the…
NVIDIA NemoClaw Flaw Lets Malicious Webpages Poison Local AI Models via Ollama
Oasis Security has disclosed a vulnerability in NVIDIA's NemoClaw, an open-source reference stack for running AI agents, that could allow an attacker-controlled…
Marimo Notebook High-Severity Flaw Allows MCP Command Execution Before Cell Execution
Marimo, the developer of the open-source Marimo notebook software, has patched a high-severity code injection vulnerability tracked as CVE-2026-75149. The flaw, discovered…
Weedhack Malware Spreads via Fake Minecraft Clients and SEO Poisoning
Cybersecurity researchers have uncovered an ongoing campaign distributing the Weedhack malware to gamers through fake Minecraft clients and SEO poisoning. McAfee Labs…
Shadow AI Power Users: The 5% Creating Outsized Enterprise Security Risk
New research from Akamai reveals that a small fraction of enterprise employees—the top 5% of AI power users—are creating a disproportionate security…
AI Coding and Open Source Risk: Managing Remediation Debt in Enterprise Security
As AI coding tools accelerate software development, they also introduce open-source dependencies at a pace that can overwhelm security teams, leading to…
ActiveState: Managing Open Source Risk in AI-Driven Development
ActiveState is a company that provides solutions for managing open source software. In the context of this article, ActiveState conducted a survey…
GPT-5 Resists Cryptographic Payload Injection
OpenAI's GPT-5 failed to parse decryption instructions in a cryptographic payload injection attempt, indicating a level of resistance to the attack.