Topic

Prompt Injection

Topic

Prompt Injection

Developers Say Meta's Muse Hands Over Its Entire Filesystem on Request

Two developers independently got Meta's Muse agent to export its full Ubuntu filesystem. Meta denies a breach, citing per-user VMs, but prompt injection resistance appears near zero.

September 24, 2026

OpenAI Trains GPT-5 Mini-R to Obey the Instruction Hierarchy
September 24, 2026
OpenAI Lays Out Its Defense Playbook Against Prompt Injection
September 24, 2026
OpenAI Details How It Blocks URL-Based Data Exfiltration in ChatGPT
September 24, 2026
OpenAI: Prompt Injection Has Become Social Engineering
September 24, 2026
OpenAI Launches Safety Bug Bounty to Pay for AI Abuse Findings
September 24, 2026
OpenAI's GPT-Red: An AI Attacker That Makes Models Safer
September 24, 2026
OpenAI adds Lockdown Mode and risk labels to ChatGPT
September 24, 2026