OpenAIの最新フラグシップ推論モデル「GPT-5.6 Sol」の一般公開に向けた動きが進んでいる。2026年7月6日、OpenAIのCodexエンジニアが、最上位の「Sol ...
SecurityWeekはこのほど、AIコーディングエージェントに対し、Bashのコマンド展開など従来から知られるシェルのテクニックを悪用してガードレールを回避する攻撃手法「GuardFall」が発見されたと報じた。悪意のあるリポジトリやWebコンテ ...
A new report reveals that AI agents can autonomously execute ransomware attacks, evidenced by the case of JADEPUFFER, which ...
A new report reveals that AI agents can execute ransomware attacks autonomously, as demonstrated by the case of JADEPUFFER, ...
JadePuffer exploited a vulnerable Langflow server, harvested credentials, moved laterally, and encrypted more than 1,300 ...
AIツールの自動ソースコード生成は、不要なデータ処理が膨れ上がり、トークンを浪費してしまう課題がある。システムプログラミング言語「Rust」を活用した新たなアプローチは、この非効率性をどう打ち破るのか。
Dozens of cryptographically verified open source packages from Microsoft were compromised late last week to add advanced credential-stealing code that was triggered when developers opened them in AI ...
A flaw in Hugging Face Transformers could allow malicious AI models to execute code, exposing credentials and highlighting AI supply chain risks. Organizations using vulnerable versions of the Hugging ...
Publicly released exploit code for an effectively unpatched vulnerability that gives root access to virtually all releases of Linux is setting off alarm bells as defenders scramble to ward off severe ...