Szilidam·4d agoTryhackme — BankGPT — WriteupBankGPT is a chatbot exploitation challenge at TryHackMe. In this task the objective is to acqiure the “secret key” from the target chatbot…
InThe Constellar Digital&Technology BlogbyNedved Yang·Nov 1Geekout Time: AI Pentesting in the AI-SDLC: Can It Improve Security, Reduce Costs, and Speed Up…Exploring Claude, Strix and OWASP Juice Shop as part of an AI- SDLC.
Abobakrmohamed·Oct 20Hacking AI Without Exploiting a Single Line of Code!Traditional hacking often involves exploiting code, bypassing authentication, or manipulating application logic.
Benyamin Salimi·Oct 3One Day, Three Ebook Vaults, and €1.50 of Token Cost with an Open-Weight ModelHow €1.50 of token cost with an open-weight model did in one day what took me four weekends by hand — and how it nearly lost to the hardest…
RRespxwnss·Sep 19Bypass of Security Filters via Braille EncodingWhile researching the types of encoding that can be used and interpreted by AI agents, I identified a significant vulnerability in the…
MazenWaleed·Sep 5Exploiting AI Agents to Trigger Secondary Vulnerabilities — PortSwigger Lab WriteupLab: Exploiting AI agents to trigger secondary vulnerabilities Difficulty: Practitioner Category: LLM Attacks — Indirect Prompt Injection +…
MazenWaleed·Sep 2Exploiting AI Agents to Exfiltrate Sensitive Information — PortSwigger Lab WriteupLab: Exploiting AI agents to exfiltrate sensitive information Difficulty: Apprentice Category: LLM Attacks — Indirect Prompt Injection
DuckWrites·Sep 1Claude Code vs Codex — for Pentesting —AI pentesting is not new. Major AI products have been doing this for over a year. But it’s not until recently that they have been able to…A response icon3A response icon3
