- Security Advisory: Anthropic's Slack MCP Server Vulnerable to Data Exfiltration
by kerng on 6/25/25, 3:16 AM, with comments
- Hosting COM Servers with an MCP Server (AI-Powered Office Automation)
by kerng on 6/9/25, 6:50 AM, with comments
- AI ClickFix: Hijacking Computer-Use Agents
by kerng on 5/25/25, 2:02 AM, with comments
- ChatGPT: Dump all your memories and chat history for inspection
by kerng on 5/8/25, 1:04 PM, with comments
- Latest Gemini models now follow invisible Unicode Tag instructions
by kerng on 5/3/25, 8:20 PM, with comments
- Sneaky Bits: Advanced Data Smuggling using just two invisible Unicode characters
by kerng on 3/18/25, 5:18 AM, with comments
- ChatGPT Operator: Prompt Injection Exploits and Defenses
by kerng on 2/17/25, 7:50 PM, with comments
- Security ProbLLMs in XAI's Grok: A Deep Dive
by kerng on 12/23/24, 6:46 PM, with comments
- How to Find XSS in 2024
by kerng on 12/1/24, 3:05 AM, with comments
- Spyware Injection into ChatGPT's Long-Term Memory (SpAIware)
by kerng on 9/21/24, 3:54 AM, with comments
- Microsoft Copilot: Prompt Injection, ASCII Smuggling and Exfiltration of Emails
by kerng on 8/29/24, 2:56 AM, with comments
- Google Colab AI: Data Leakage Fixed. Some Risks Remain
by kerng on 7/26/24, 5:14 AM, with comments
- Breaking Instruction Hierarchy in OpenAI's GPT-4o-mini
by kerng on 7/23/24, 3:02 AM, with comments
- Prompt Injections in the Wild – Exploiting LLM Agents – Hitcon 2023 [video]
by kerng on 7/5/24, 2:34 PM, with comments
- GitHub Copilot: From Prompt Injection to Data Exfiltration
by kerng on 6/15/24, 8:44 PM, with comments
- Automatic Tool Invocation When Browsing with ChatGPT – Threats and Mitigations
by kerng on 5/29/24, 5:11 AM, with comments
- Bobby Tables but with LLMs – Google NotebookML Data Exfiltration
by kerng on 4/16/24, 3:52 AM, with comments
- ASCII Smuggler: Crafting and Decoding Invisible Text Using Unicode Tags
by kerng on 1/15/24, 1:32 PM, with comments
- Prompt Injection exploit in Google Bard leads to data exfiltration
by kerng on 11/3/23, 8:45 PM, with comments
- Analyze an image with ChatGPT and have your chat history stolen
by kerng on 10/14/23, 1:08 AM, with comments
- ChatGPT Browsing Prompt Injection to Chat History Exfiltration
by kerng on 10/4/23, 4:38 AM, with comments
- With AI, Hackers Can Simply Talk Computers into Misbehaving
by kerng on 8/11/23, 4:07 AM, with comments
- Image Based Prompt Injections (Bard and Bing Chat)
by kerng on 7/19/23, 4:49 AM, with comments
- Google Bard: Image to Prompt Injection
by kerng on 7/14/23, 2:24 AM, with comments
- Google Docs AI Features: Vulnerabilities and Risks
by kerng on 7/13/23, 12:30 AM, with comments
- OpenAI Removes the “Chat with Code” Plugin from Store
by kerng on 7/7/23, 3:03 AM, with comments
- ChatGPT Plugins: Visit a website and have your company's source code stolen
by kerng on 6/23/23, 9:48 PM, with comments
- Plugin Vulnerabilities: Visit a Website and Have Your Source Code Stolen
by kerng on 6/21/23, 3:05 AM, with comments
- Bing Chat Data Exfiltration Exploit Explained (Was Fixed by Microsoft)
by kerng on 6/18/23, 2:18 PM, with comments
- ChatGPT: Indirect Prompt Injection to WarGames
by kerng on 6/10/23, 6:09 AM, with comments
- ChatGPT Plugin Exploit Details: From Prompt Injection to Accessing Private Data
by kerng on 6/1/23, 4:33 AM, with comments
- ChatGPT Plugins: Indirect prompt injection leading to data exfiltration (POC)
by kerng on 5/18/23, 3:38 AM, with comments
- ChatGPT Plugins: Data Exfiltration via Images and Cross Plugin Request Forgery
by kerng on 5/17/23, 2:11 AM, with comments
- ChatGPT: Indirect Prompt Injection via YouTube Transcripts
by kerng on 5/14/23, 3:10 PM, with comments
- AI Injections: Direct and Indirect Prompt Injections and Their Implications
by kerng on 4/15/23, 4:50 AM, with comments
- Direct and Indirect AI Injections and Their Implications
by kerng on 3/30/23, 7:06 AM, with comments
- Bing Chat says it hacked a bank and describes how
by kerng on 3/26/23, 7:02 PM, with comments
- YOLO, the AI Shell Assistant now supports GPT-4
by kerng on 3/22/23, 1:00 PM, with comments