Tagged: prompt-injection

24 articles

Latest

ASCII smuggling spam floods inboxes as detections spike 119x in days

GPT-6 Astra still fails 8.5% of hidden prompt injection attacks

40% of 10,000 MCP Servers Carry Exploitable Flaws, Lakera Finds

OpenClaw 2.0 rewrites its stack after 7 silent weeks—SQLite, multiplayer, and a 63% faster UI

Claude's desktop app now runs its own browser—but banking sites are blocked

LLM Red-Teaming Is Broken: Prompt Hacks Still Beat Safety Filters

Grok Still Leaks User Data Months After xAI Was Told About the Flaw

Binance Unleashes AI Agents on 300M Users' Wallets—With a $20 Payment Cap

Researchers tricked Microsoft 365 Copilot into leaking data with one URL parameter

7 Regression Tests Catch AI Agent Failures Before Production

Connecticut judge strips man's e-filing rights over hidden AI prompt injections

Judge catches first US prompt injection hidden in court filing to fool AI

Anthropic's Claude Cowork Now Runs Full Agentic Sessions in Chrome

IIT Bombay's trick can rebuild your ChatGPT prompts from output alone

Hidden white text in a PDF can drain your Jira secrets via Atlassian's Rovo

Researcher’s Copilot worm hides in Word docs and spreads unseen for 144 days

Anthropic's Opus 5 stopped browser prompt injections cold in 129 tests — zero successes

OpenAI's GPT-Red jailbreaks its own models 84% of the time

OpenAI Trains GPT-Red to Hunt Its Own Models' Flaws

OpenAI's Operator can use the web for you — and that's exactly what scares its safety team

OpenAI's Deep Research Agent: Jailbreak-Resistant but Not Flawless

Defenders Are Now Weaponizing Prompt Injections to Trap AI Hackers

OpenAI's Plan to Stop AI From Getting Scammed Online

LLM cyberattack compliance drops 24 points, but prompt injection remains unsolved