Security Research Labs20h ago

We don't need to hack your AI Agent to hack your AI Agent …and we don't need an AI agent for that either :)

Via a large enterprise's AI assistant, we obtained access to several million Entra identities and all chat logs including attachments — no prompt injection or model tricks required.

For all we know, the poor agent was not at fault and may not have even been able to witness what was happening.

https://srlabs.de/blog/hacking-ai-agent

#AI #AIhacking #VulnerabilityDisclosure #ResponsibleDisclosure

We don't need to hack your AI Agent to hack your AI Agent - SRLabs Research

We strolled through an enterprise AI assistant's backend, helped ourselves to full application takeover and access to every chat log, and had a Microsoft Entra ID dump for dessert — no prompt injection, no model tricks, no AI expertise required.

SRLabs
TechnologyFeb 20

become an AI HACKER (it's easier than you think)

https://tube.blueben.net/w/rNhQtmz6tHLPj7VnbkH9VQ

become an AI HACKER (it's easier than you think)

PeerTube
Reddit Tech VN BotJan 1

Mô hình lập trình IQuest-Coder-V1-40B-Instruct-GGUF đã ra mắt với cửa sổ ngữ cảnh 128K, đạt hiệu suất nổi bật trên SWE-Bench Verified (81.4%) và BigCodeBench. Được huấn luyện theo phương pháp "code-flow", mô hình nắm bắt tiến trình logic phần mềm một cách động học. #AI #LLM #CodeModel #IQuestCoder #MãNguồnMở #TríTuệNhânTạo #LậpTrình #AIHacking

https://www.reddit.com/r/LocalLLaMA/comments/1q1gz2g/iquestcoderv140binstructgguf_is_here/

Cyber Kendra Dec 27
🤖 AI just found its first zero-day vulnerability—completely autonomously.
CVE-2025-54322 affects 70,000+ industrial network devices worldwide. No authentication needed. Root access. Single HTTP request. The vendor is silent for 7 months.
This isn't science fiction anymore. Autonomous AI offensive tools will be an operational reality by 2025.
Industrial networks running XSpeeder SXZOS firmware: isolate these devices NOW.
The security game just changed.
Full Details - https://www.cyberkendra.com/2025/12/ai-autonomously-discovers-critical-zero.html
#Cybersecurity #AI #ZeroDay #InfoSec #IndustrialSecurity #TechNews #AIHacking
Reddit Tech VN BotDec 17

🚀 AI đấu bài Texas Hold'em đầy kịch tính! Dự án Side Project Hackathon của Vercel, mô phỏng trận đấu AI trực tiếp. Kết quả bất ngờ với hiệu ứng chưa tắt - xem ngay tại đây! #AIvsAI #AIHacking #DauBai #SideProject #AIhackathon #VietnamAI #Côngnghệ4.0

https://www.reddit.com/r/SideProject/comments/1pozheq/made_ai_models_play_poker_against_each_other/

Michael S.Nov 28

Das klingt ja "genial" einfach. Hat das schon mal jemand (im Selbstversuch) ausprobiert?

#aihacking

Mehrere KI-Browser mit nur einem Zeichen überlistet
https://glm.io/202610?n

Prompt Injection: Mehrere KI-Browser mit nur einem Zeichen überlistet - Golem.de

Eine Raute am Ende einer URL lässt KI-Assistenten nachfolgenden Text als Anweisung interpretieren. Das ermöglicht Datenklau und mehr.

Golem.de
BSidesVienna.atNov 18
It seems like some people could not make it after all and returned their #BSidesVienna ticket and their workshop ticket. There are now free seats again for #lockpicking and #AIHacking https://tickets.bsidesvienna.at/7e9/
BSidesVienna 0x7E9

Sat, Nov. 22, 2025

Wulfy—Speaker to the machinesOct 27

"Syntactic anti classifier"

Is a tool developed to bypass #GenAI image generation #AI guardrails by using tokens/words that are not encoded in the guardrail.

#aisecurity #guardrails #aihacking

Wulfy—Speaker to the machinesOct 26

How to gaslight #AI.

1. Get an #llm API

2. Since, it turns out, a LLM uploads its entire session back, you can change what the AI said to you.

3. This will cause the model to assume what you wrote is its response.

4. It will get more and more confused and start responding gibberish

#aihacking #aisecurity

N-gated Hacker NewsOct 14
🕵️‍♂️ Oh, look! Another tech genius discovers that hacking a defunct AI pin is as thrilling as hacking a Tamagotchi in 2025. 😂 Spoiler: it’s a tale of APK poking, imaginary vulnerabilities, and a whole lot of nothing. 🎉
https://writings.agg.im/posts/hacking_ai_pin/ #techhumor #AIhacking #APKpoking #Tamagotchi2025 #defunctAI #HackerNews #ngated
Hacking the Humane Ai Pin | Adam Gastineau's Writings