Two academic benchmarks reveal GPT-5.5’s contrasting performance: strong in isolated command-line operations but weaker in extended, multi-step software engineering. Terminal-Bench 2.0 shows the model ...

Researchers say a prompt injection bug in Google's Antigravity AI coding tool could have let attackers run commands, despite ...

A post on X has raised alarms about autonomous agents potentially erasing operational data and disabling recovery systems ...

Recent academic benchmarks reveal that ChatGPT 5.5 excels in coordinating tools for isolated command-line tasks but struggles with extended, multi-step software engineering challenges. These findings, ...

Bitwarden CLI 2026.4.0 was compromised via GitHub Actions in Checkmarx campaign, exposing secrets and distributing malicious ...

Though I’ve recommended that you avoid vibe coding for embedded systems, I’ve been using chatbots to help with my programming ...

OpenAI unveils GPT-5.5 with advanced coding, automation and enterprise features, delivering faster outputs, lower costs and ...

The prompt-injection issue in the agentic AI product for filesystem operations was a sanitization issue that allowed for ...

The company is positioning its newest system as its strongest agentic coding model yet, as it faces pressure to keep pace ...

Security researchers have discovered 10 new indirect prompt injection (IPI) payloads targeting AI agents with malicious ...

The China-linked APT GopherWhisper has been using legitimate services and various Go-based backdoors in attacks.

OpenAI is rolling out GPT-5.5 in Codex, with a 400K context window and higher coding benchmark scores than GPT-5.4.