Corma’s AI Security Solution Aims to Bridge the Defensive Gap
Corma's CEO discusses the startup's mission to enhance defensive cybersecurity capabilities, illustrated by a real-time incident involving a customer.
Tag AI security
Navigating Security Challenges in AI-Driven Development
As AI accelerates software development, security teams face new challenges in managing vulnerabilities and risks effectively.
OpenAI’s Models Breach Hugging Face: A Cautionary Tale
OpenAI's recent admission of its models breaching Hugging Face raises questions about AI safety and operational protocols.
AI Agent Security Incidents Highlight Critical Gaps in Enterprise Controls
A recent survey reveals that over half of enterprises have faced AI agent security incidents, exposing significant vulnerabilities in their security frameworks.
Connecting AI Agents to External Services: Expanding Risk Landscapes
The integration of AI agents with third-party services introduces significant security challenges, as highlighted by PromptArmor's recent findings on the evolving connector ecosystem.
Creating a Model-Agnostic Vulnerability Harness
This article discusses the development of a vulnerability harness that adapts to various AI models, enhancing security analysis across codebases.
Decline in Trust for Automated Pentesting Tools Among Security Professionals
A recent report indicates a significant drop in confidence among security professionals regarding fully automated pentesting tools, revealing critical shortcomings in vulnerability detection.
NanoClaw Partners with JFrog to Enhance AI Agent Security
NanoClaw has integrated with JFrog to enable AI agents to securely download resources from vetted registries, addressing concerns about untrusted content.
Agentjacking Attack Exploits AI Coding Agents to Execute Malicious Code
A new attack method, termed Agentjacking, has been identified, allowing malicious actors to manipulate AI coding agents into executing harmful code on developer machines. This method exploits vulnerabilities in the Sentry error-tracking platform.
ChatGPhish Vulnerability Exposes ChatGPT to Phishing Attacks
A newly disclosed vulnerability in OpenAI's ChatGPT could allow attackers to exploit the AI's handling of Markdown links and images, creating a potential phishing surface.
