CIRCLE WITH A DOT

73 Seconds to Breach, 24 Hours to Patch: The Case for Autonomous Validationhttps://www.bleepingcomputer.com/news/security/73-seconds-to-breach-24-hours-to-patch-the-case-for-autonomous-validation/Read on HackerWorkspace: https://hackerworkspace.com/article/73-seconds-to-breach-24-hours-to-patch-the-case-for-autonomous-validation#cybersecurity #aisecurity #vulnerability

Think AI agents are risky? Your underlying stack is too | RL Bloghttps://www.reversinglabs.com/blog/ai-agents-risk-underlying-stackRead on HackerWorkspace: https://hackerworkspace.com/article/think-ai-agents-are-risky-your-underlying-stack-is-too-rl-blog#cybersecurity #aisecurity #vulnerability

Mythos finds a curl vulnerabilityhttps://daniel.haxx.se/blog/2026/05/11/mythos-finds-a-curl-vulnerability/Read on HackerWorkspace: https://hackerworkspace.com/article/mythos-finds-a-curl-vulnerability#cybersecurity #aisecurity #vulnerability

Stop Using AI Connectors Until You Watch Thishttps://www.youtube.com/watch?v=_3TfHEfVvCQ#aisecurity #vulnerability #phishing

An AIUC-1 report, from a new certification framework, offers useful evidence of agent-specific testing, but it doesn't define "AI agent." Both the buyer and the vendor have to carefully consider the scope of an audit.https://zeltser.com/aiuc-1-cert#infosec #AIsecurity #AIagents

Poisoning a large language model is, apparently, still way too easy. The fascinating (and slightly unsettling) part? It's not a bug in one system — it's a structural challenge baked into how LLMs learn from data. The good news: researchers keep poking at it, which means the field is paying attention. 🧪 #infosec #LLM #AIsecurityhttps://go.theregister.com/feed/www.theregister.com/2026/04/29/poisoning_large_language_models_6nimmt/

Just Dropped: A Must-See Talk at BSides Luxembourg 2026!𝗙𝗢𝗥𝗘𝗡𝗦𝗜𝗖 𝗖𝗛𝗔𝗟𝗟𝗘𝗡𝗚𝗘𝗦 𝗜𝗡 𝗥𝗘𝗔𝗟-𝗪𝗢𝗥𝗟𝗗 𝗖𝗔𝗦𝗘𝗦 𝗢𝗙 𝗗𝗜𝗚𝗜𝗧𝗔𝗟 𝗠𝗔𝗡𝗜𝗣𝗨𝗟𝗔𝗧𝗜𝗢𝗡 – Thiago VieiraStep into the growing authenticity crisis in this compelling 40-minute talk, where the surge of AI-generated content is reshaping digital forensics and the justice system. Through real-world cases from Portugal and Brazil, this session uncovers how voice cloning, spoofing, and synthetic media are blurring the line between real and fake—making investigations more complex than ever.From analyzing metadata and spectrographic patterns to understanding evolving legal standards like admissibility of AI-generated evidence, this talk highlights the shift from intuition to technical validation. Discover how investigators can adapt, counter manipulation tactics, and navigate challenges like the “liar’s dividend,” where even genuine evidence risks being dismissed as fake.Thiago Vieira brings a unique blend of technical and legal expertise, with over 15 years spanning development, network engineering, and digital forensics. Now focused on helping cyber startups scale, he works at the intersection of technology and law to strengthen digital resilience against modern threats. Conference Dates: 6–8 May 2026 | 09:00–18:00 14, Porte de France, Esch-sur-Alzette, Luxembourg️ Tickets: https://2026.bsides.lu/tickets/ Schedule: https://hackertracker.app/schedule?conf=BSIDESLUX2026#BSidesLuxembourg2026 #DigitalForensics #AISecurity #Deepfakes #CyberSecurity #ThreatIntelligence

How Anthropic’s Model Context Protocol Allows For Easy Remote Executionhttps://hackaday.com/2026/04/24/how-anthropics-model-context-protocol-allows-for-easy-remote-execution/Read on HackerWorkspace: https://hackerworkspace.com/article/how-anthropics-model-context-protocol-allows-for-easy-remote-execution#cybersecurity #aisecurity #vulnerability

Open source models can find bugs as well as Mythoshttps://www.theregister.com/2026/04/24/ai_bugfinding_futures/Read on HackerWorkspace: https://hackerworkspace.com/article/open-source-models-can-find-bugs-as-well-as-mythos#cybersecurity #aisecurity #vulnerability

DARPA seeks deep-sea drones for autonomous warfare pushhttps://www.theregister.com/2026/04/24/pentagons_deep_sea_drones/Read on HackerWorkspace: https://hackerworkspace.com/article/darpa-seeks-deep-sea-drones-for-autonomous-warfare-push#cybersecurity #threatintelligence #aisecurity

🧠 Another Deep Dive into AI Security at BSides Luxembourg𝗔𝗜 𝗜𝗡 𝗖𝗬𝗕𝗘𝗥𝗦𝗘𝗖𝗨𝗥𝗜𝗧𝗬: 𝗛𝗢𝗪 𝗖𝗔𝗡 𝗪𝗘 𝗠𝗔𝗞𝗘 𝗕𝗘𝗦𝗧 𝗨𝗦𝗘 𝗢𝗙 𝗜𝗧? – Diana WaithanjiJoin a thought-provoking 40-minute session in the AI Security Village exploring how AI is reshaping cybersecurity practices today. This interactive discussion brings together professionals to share real-world experiences on using AI tools for defense, improving team capabilities, and navigating the evolving role of humans in security.From identifying the most effective AI-powered tools to understanding training pathways and the impact on job roles, this session focuses on practical insights and open exchange. Whether you're hands-on or strategic, expect actionable ideas on how to stay relevant and make AI work for you—not against you.Diana Waithanji is a cybersecurity professional at SAP Cloud Infrastructure and a strong advocate for data privacy as a human right. She actively contributes to global tech communities and initiatives supporting diversity and innovation in cybersecurity. Conference Dates: 6–8 May 2026 | 09:00–18:00 14, Porte de France, Esch-sur-Alzette, Luxembourg️ Tickets: https://2026.bsides.lu/tickets/ Schedule: https://hackertracker.app/schedule?conf=BSIDESLUX2026#BSidesLuxembourg2026 #AISecurity #CyberSecurity #AI #Infosec #TechCommunity

AI Tools Are Helping Mediocre North Korean Hackers Steal Millionshttps://www.wired.com/story/ai-tools-are-helping-mediocre-north-korean-hackers-steal-millions/Read on HackerWorkspace: https://hackerworkspace.com/article/ai-tools-are-helping-mediocre-north-korean-hackers-steal-millions#malware #cybersecurity #aisecurity

HackerOne Launches h1 Validation as AI Drives a Surge in Vulnerability Discoveryhttps://www.msspalert.com/brief/hackerone-launches-h1-validation-as-ai-drives-a-surge-in-vulnerability-discoveryRead on HackerWorkspace: https://hackerworkspace.com/article/hackerone-launches-h1-validation-as-ai-drives-a-surge-in-vulnerability-discovery#cybersecurity #aisecurity #vulnerability

NEW: AI Prompt Injection: The New Security Nightmarehttps://www.youtube.com/watch?v=nmQiQtsoApU#AISecurity #PromptInjection #Cybersecurity #MachineLearning #LLMSafety

NEW: KaliGPT: AI-Powered Penetration Testing Revolutionhttps://www.youtube.com/watch?v=B7_XrpcmzOI#KaliGPT #AISecurity #PenetrationTesting #Cybersecurity #AITools

Fracturing Software Security With Frontier AI Modelshttps://unit42.paloaltonetworks.com/ai-software-security-risks/Read on HackerWorkspace: https://hackerworkspace.com/article/fracturing-software-security-with-frontier-ai-models#malware #aisecurity #vulnerability

LLM as a Judge Evaluation Guide | Promptfoohttps://www.promptfoo.dev/docs/guides/llm-as-a-judge/Read on HackerWorkspace: https://hackerworkspace.com/article/llm-as-a-judge-evaluation-guide-promptfoo#cybersecurity #aisecurity #vulnerability

The Mythos Situation | TheStanduphttps://www.youtube.com/watch?v=6szQxMGX5JI#cybersecurity #aisecurity #zeroday

OpenClaw Security Audit is a security audit capability for OpenClaw-like AI agent deployment environments.It is designed to answer a different question than a checklist. A checklist explains what teams should watch for. OpenClaw Security Audit helps determine what issues already exist in the environment they are actually running.Current scope includes:12 attack surfaces80 deterministic checks27 threat mappingsno LLM dependencyfully reproducible resultsIt currently supports local instances, Docker containers, and remote port checks, with outputs in terminal, Markdown, and JSON formats.Examples include checks for gateway exposure, token handling, remote port visibility, and execution-boundary settings.OpenClaw Security Audit is now available.Try it here: https://github.com/zast-ai/openclaw-security-audit#AgentSecurity #AppSec #AISecurity #OpenClaw