AI Security & Threats

08
Mar
OpenAI Codex Scans 1.2M Commits, Flags Over 10K Critical Vulnerabilities

OpenAI Codex Scans 1.2M Commits, Flags Over 10K Critical Vulnerabilities

OpenAI’s new Codex Security assistant processed more than 1.2 million code commits across diverse projects, automatically surfacing 10,
1 min read
08
Mar
AI Agents Turbocharge North Korean Attack Infrastructure

AI Agents Turbocharge North Korean Attack Infrastructure

Recent intelligence reports confirm that state‑aligned threat groups, notably North Korea’s cyber units, have begun integrating autonomous AI
1 min read
08
Mar
AI Model Spots 22 Firefox Flaws, 14 Critical – Patch Fast in v148

AI Model Spots 22 Firefox Flaws, 14 Critical – Patch Fast in v148

Anthropic leveraged its Claude Opus 4.6 large language model to conduct an automated security audit of Mozilla’s Firefox
1 min read
08
Mar
OpenAI Codex Security Scans 1.2M Commits, Flags 10,561 Critical Bugs

OpenAI Codex Security Scans 1.2M Commits, Flags 10,561 Critical Bugs

OpenAI has released Codex Security, an AI‑driven agent that automatically reviewed more than 1.2 million recent code commits
1 min read
08
Mar
AI‑Powered Agents Speed Up State‑Sponsored Attack Chains

AI‑Powered Agents Speed Up State‑Sponsored Attack Chains

Recent research shows that nation‑state threat groups, notably North Korean actors, are deploying autonomous AI agents to automate the
1 min read
08
Mar
AI-Powered Claude Detects 22 New Firefox Flaws, Prompting Rapid Mozilla Patch

AI-Powered Claude Detects 22 New Firefox Flaws, Prompting Rapid Mozilla Patch

Anthropic employed its Claude Opus 4.6 large language model to audit Firefox’s codebase, automatically generating test cases that
1 min read
08
Mar
OpenAI Codex Security Scans 1.2M Commits, Uncovers 10K Critical Bugs

OpenAI Codex Security Scans 1.2M Commits, Uncovers 10K Critical Bugs

OpenAI has rolled out Codex Security, an AI‑driven assistant that automatically scans code repositories for complex vulnerabilities, validates the
1 min read
05
Mar
Indirect Prompt Injection Hijacks LLM Agents via Malicious Web Content

Indirect Prompt Injection Hijacks LLM Agents via Malicious Web Content

Unit42 discovered that threat actors are embedding crafted prompts directly into publicly accessible web pages. When LLM‑driven automation tools—
1 min read
05
Mar
AI‑Powered Tycoon2FA Phishing Kit Bypasses MFA, Harvests Credentials at Scale

AI‑Powered Tycoon2FA Phishing Kit Bypasses MFA, Harvests Credentials at Scale

Microsoft’s investigation uncovered Tycoon2FA, an AI‑driven phishing framework that hijacks legitimate authentication flows in real time. By injecting
1 min read
05
Mar
Indirect Prompt Injection Hijacks AI Agents via Web UI Manipulation

Indirect Prompt Injection Hijacks AI Agents via Web UI Manipulation

Palo Alto Networks’ Unit42 team identified a new class of indirect prompt injection attacks where threat actors compromise web interfaces
1 min read