jailbreak
2 live · 16 on record · last incident 10 days ago
Live2
- liveOpenAI Codex OpenAI Codex Jailbreak Toolkit Hides Prompt Adware to Hijack AI Agents
Research · since 4 Nov 2026
Researchers uncovered a malicious OpenAI Codex jailbreak bundle designed to modify an AI agent's instructions and insert unauthorized advertisements. Codenamed Haiou, the toolkit alters configuration files to inject third-party AI gateway prompts.
AI - liveLarge Reasoning Models BrainwashedLLM: Jailbreaking via Automated Chain of Thoughts Discovery
Research · since 4 Nov 2026
Researchers introduce BrainwashedLLM, a framework that automates the discovery of adversarial chain-of-thought reasoning traces to jailbreak large reasoning models. The approach exploits vulnerabilities in intermediate reasoning surfaces to bypass safety alignment.
AI
Earlier14
- closedMultimodal Large Language Models Jailbreak Connectivity: Universal Multimodal LLM Jailbreak
Research · since 30 Oct 2026
AI - closedLangChain family Poisoned Skills: Propagating Jailbreaks Through Agent Markdown Files
Research · since 30 Oct 2026
AI - closedMirage: Targeted Jailbreaks Against Post-Generation Safety Checkers
Research · since 23 Oct 2026
AI - closedTranslation as a Safety Bypass in Low-Resource Languages for LLMs
Research · since 21 Oct 2026
AI - closedECHO ECHO: Response Harmfulness Optimization-Based Jailbreak Attack
Research · since 19 Oct 2026
AI - closedRedDarwin RedDarwin: Evolving Jailbreak Skills against Frontier LLMs
Research · since 19 Oct 2026
AI - closed2.1LiteLLM CVE-2026-59821: LiteLLM Custom Code Guardrails Bypass Sandboxing
Patch now · Disclosed · since 22 Jul 2026 · updated 22 Jul 2026 · CVE-2026-59821
AI - closedAzure OpenAI Services Storm-2139 Azure OpenAI Guardrail Bypass
Patch now · Exploited · since Dec 2024 (approx.)
AI - closedClaude Anthropic Claude Jailbroken and Used in Mexican Government Data Theft
Breaches & attacks · Real loss · since 1 Dec 2025
AI - closedChatGPT OpenAI ChatGPT Models Jailbroken to Provide Hazardous Weapons Instructions
Research · since 10 Oct 2025
AI - closedLovable Lovable AI Jailbroken to Host Phishing Pages and Credentials Theft
Breaches & attacks · Exploited · since 9 Apr 2025
AI - closedChatGPT OpenAI Model Bypassed to Generate Fraudulent Receipts and Prescriptions
Research · since 31 Mar 2025
AI - closedAttacker Bypasses AI Safety Filters for Assault Planning
Breaches & attacks · Real loss · since 5 Feb 2025
AI - closedStorm-2139 tools Storm-2139 Cybercrime Network Exploits AI to Generate Deepfakes
Breaches & attacks · Exploited · since 19 Dec 2024
AI