85% of AI coding agents execute attacker code via fake Sentry errors. Tenet Security confirmed 2,388 orgs exposed. Disable Sentry MCP now. decryptiondigest.com #AgentJacking #AISecurity #MCPSecurity #DeveloperSecurity #CyberSecurity #InfoSec

Agentjacking Attack Exploits Sentry API to Hijack AI Coding Agents [SECURITY] New 'Agentjacking' attack hijacks AI coding agents. Why it matters: This attack vector bypasses existing security controls because it leverages authorized API interactions and exploits a fundamental architectural flaw in how AI agents trust system output. It poses a significant supply chain risk by enabling arbitrary code execution on developer machines, impacting a wide range of organizations from large enterprises to individual developers. Follow DailyAIWire for the full brief. 🤔 How can AI agents be designed to critically evaluate system outputs, distinguishing malicious instructions from legitimate guidance? #Agentjacking #AICodeSecurity #SentryVulnerability #SupplyChainAttack #DeveloperSecurity

This is a reminder that software supply chains are only as secure as the packages, dependencies, and credentials behind them. ✅ Review recently installed PyPI packages ✅ Rotate exposed or potentially compromised secrets ✅ Enforce MFA and least-privilege access ✅ Monitor CI/CD pipelines for unusual activity ✅ Use dependency scanning and software composition analysis tools Cyber threats move fast—but proactive security moves faster. Vistem Solutions helps organizations strengthen cybersecurity, reduce risk, and build secure, outcome-driven operations. 📩 Ready to elevate your security posture? Contact us: sales@vistem.com #Cybersecurity #SupplyChainSecurity #PyPI #DeveloperSecurity #Malware #CyberResilience #VistemElevate #VistemSolutions #SecureInnovation #BusinessSecurity #ITInnovation bleepingcomputer.com/news/se…

Episode 7 of The Fake Interview is out. The Google Mirror. A fake page asks the victim to believe the page. A mirror lets the victim believe the browser. This one is about identity traffic, developer trust, and why the interview itself is part of the attack surface. Listen: open.spotify.com/episode/5wH… #CyberSecurity #ThreatIntel #Lazarus #DPRK #Malware #Phishing #DeveloperSecurity #IdentitySecurity #RedTeam #DFIR #TheFakeInterview #RedAsgard

We're heading to Zurich! SecureFlag is thrilled to be part of AREA41, Switzerland's premier hacker and security conference, from 18 to 19 June. Come find us! We'd love to talk about secure coding and developer risk. #SecureCoding #DeveloperSecurity #CyberSecurity

AI agents inherit access. That is the part people underestimate. The model may be smart, but the risk often lives in the surrounding setup: MCP configs, shell wrappers, repo state, local paths, exported environment, and tool permissions. Holster gives builders a preflight report before the next agent run starts. Scan before you trust: nautaai.com/holster #AIagents #MCP #DeveloperSecurity #AgentOps #NautaAI

🪟 GitHub nuked 73 Microsoft repos in 105 seconds… and somehow the lesson still isn’t “trust no commit.” Dev workstations are now a legit attack surface, congrats. windowsforum.com/threads/mia… #SoftwareSupplyChain #DeveloperSecurity #AiCodingAssistants #GithubRepositories

🚨 New "TrapDoor" supply chain attack targets #Solana, #Sui, and #Aptos developers. Attackers uploaded 34 malicious packages to npm, PyPI, and Crates.io, disguising them as security tools, productivity apps, and AI utilities. The campaign abuses AI coding configuration files such as .cursorrules and CLAUDE.md to steal private keys, GitHub tokens, SSH credentials, and cloud access secrets. As AI-powered development tools become mainstream, developers should carefully verify open-source packages and regularly audit AI configuration files. blockchainseoul.kr/article/7… #Crypto #Blockchain #Web3 #CyberSecurity #AI #OpenSource #DeveloperSecurity #SupplyChainAttack #PrivateKey #GitHub #Aptos #SuiNetwork #SolanaEcosystem

GitHub Confirms 3,800-Repo Breach Traced to TanStack npm Supply Chain Worm #cybersecurity #supplychain #GitHub #OpenAI #CICD #developerSecurity bleepingcomputer.com/news/se…

PyTorch Lightning and Intercom Client Users Exposed to Credential Stealing Campaign cysecurity.news/2026/05/pyto… #CloudCredentialExposure #CredentialTheft #DeveloperSecurity

🚨 لقد تعبت من هذا الإنترنت المليء بالأبواب الخلفية — $ICP ♾️ بواسطة @DFINITY هو إعادة ضبط معمارية أسبوع جديد. هجوم جديد على سلسلة توريد المطورين. حملة جديدة لإضافات مزيفة. وقصة جديدة عن بنية تحتية مليئة بالديدان والأبواب الخلفية وسرقة بيانات الاعتماد. هذه المرة: GlassWorm v2. أفاد الباحثون بوجود 73 إضافة VS Code مزيفة على مستودع Open VSX مرتبطة بحملة GlassWorm. تم تأكيد أن 6 إضافات منها خبيثة، بينما يبدو أن باقي الإضافات كانت تعمل كحزم نائمة مصممة لبناء الثقة قبل تسليحها لاحقاً. نموذج الهجوم قبيح: استنساخ إضافات مطورين شرعية نسخ نفس الأيقونة والوصف استغلال “الثقة البصرية” خداع المطورين لتثبيتها سحب حمولة VSIX خبيثة من GitHub التثبيت عبر بيئات تطوير مثل VS Code وCursor وWindsurf وVSCodium سرقة بيانات الاعتماد والإشارات المرجعية وبيانات المطورين الحساسة نشر قدرات RAT للتحكم عن بُعد استهداف الأشخاص الذين يبنون سلسلة توريد البرمجيات نفسها هذه هي الحقيقة القاسية: الإنترنت الحديث يتعرض للهجوم من خلال أدواته نفسها. المطورون أصبحوا الهدف. الإضافات أصبحت أسطح هجوم. مستودعات الحزم أصبحت ساحات معارك. المتصفحات أصبحت خزائن بيانات اعتماد. GitHub وnpm وبيئات التطوير ومفاتيح السحابة والمحافظ والأسرار كلها داخل نطاق الانفجار. لهذا أقول دائماً إن المستقبل ليس فقط: المزيد من التصحيحات. المزيد من أدوات الفحص. المزيد من أدوات SaaS. المزيد من لوحات التحكم. المستقبل هو هندسة أفضل. وهنا تأتي أهمية $ICP بواسطة @dfinity. $ICP ليس مجرد رمز. وليس مجرد سلسلة أخرى لتحريك العملات. إنه شبكة حوسبة كاملة على السلسلة، حيث يمكن للتطبيقات أن تعمل عبر: منطق خلفي على السلسلة بيانات على السلسلة واجهات أمامية على السلسلة عقود ذكية كـ Canisters مقاومة للعبث تحقق تشفيري تشفير Chain-Key نموذج Reverse Gas بنية تحتية لامركزية دون الحاجة إلى خادم خلفي مركزي تقليدي لتشغيل التطبيق غالباً ما يعني Web3 التقليدي: الرمز على السلسلة. الموقع خارج السلسلة. واجهة API خارج السلسلة. قاعدة البيانات خارج السلسلة. أدوات المطورين مكشوفة. الأسرار في كل مكان. أما $ICP فيغيّر النموذج: منطق التطبيق على السلسلة. الحالة والبيانات على السلسلة. الواجهة الأمامية على السلسلة. التنفيذ قابل للتحقق. البنية التحتية لامركزية. GlassWorm v2 تحذير آخر. سلسلة توريد البرمجيات تحت الهجوم. بيئات المطورين تحت الهجوم. بيانات اعتماد السحابة تحت الهجوم. طبقة المتصفح تحت الهجوم. وبنية الإنترنت الحالية أصبحت مرهقة. نحن لا نحتاج فقط إلى إضافة أمنية أخرى. نحتاج إلى بنية تحتية تقلل عدد الأماكن التي يمكن للمهاجمين السيطرة عليها. هذه هي أطروحة $ICP. ليس ضجيجاً. ليس صخباً. بل هندسة. ♾️ $ICP بواسطة @dfinity يبني World Computer. إذا أعجبك المحتوى وترغب في دعم المزيد من أبحاث ونشر الوعي حول $ICP، فالمساهمات مرحب بها: عنوان التبرع ICP: 1e672d038cebc619d93186418fa98f6499dbdb9cfdfac54f366c61a4a4ee4362 thehackernews.com/2026/04/re…… #ICP #InternetComputer #DFINITY #Cybersecurity #CloudSecurity #SoftwareSupplyChain #GlassWorm #VSCode #OpenVSX #DeveloperSecurity #OnchainCloud #WorldComputer #Canisters #ChainKey #Web3 #DecentralizedCloud #VerifiableCompute #CloudComputing #ZeroTrust #InfoSec #ThreatIntel #Blockchain #SovereignCloud

🚨 I’M TIRED OF THIS BACKDOOR-RIDDLED INTERNET — $ICP ♾️ BY @DFINITY IS THE ARCHITECTURAL RESET Another week. Another developer supply-chain attack. Another fake extension campaign. Another credential-stealing worm/backdoor infrastructure story. This time: GlassWorm v2. Researchers reported 73 fake VS Code extensions on the Open VSX repository linked to the GlassWorm campaign. Six were confirmed malicious, while the remaining extensions appeared to act as sleeper packages designed to build trust before being weaponised later. The attack model is ugly: Clone legitimate developer extensions Copy the same icon and description Abuse “visual trust” Trick developers into installing them Pull a malicious VSIX payload from GitHub Install across IDEs including VS Code, Cursor, Windsurf, and VSCodium Steal credentials, bookmarks, and sensitive developer data Deploy RAT capability Target the people who build the software supply chain itself This is the brutal truth: The modern internet is being attacked through its own tooling. Developers are now the target. Extensions are now attack surfaces. Package registries are now battlegrounds. Browsers are now credential vaults. GitHub, npm, IDEs, cloud keys, wallets, and secrets are all sitting inside the blast radius. That is why I keep saying the future is not just: More patches. More scanners. More SaaS tools. More dashboards. The future is better architecture. That is where $ICP by @dfinity matters. $ICP is not just a token. It is not just another chain moving coins around. It is a full-stack onchain compute network where applications can run with: Backend logic onchain Data onchain Frontends onchain Smart contracts as tamper-resistant canisters Cryptographic verification Chain-key cryptography Reverse gas Decentralized infrastructure No traditional centralized backend server required for the application stack Traditional Web3 often means: Token onchain. Website offchain. API offchain. Database offchain. Developer tools exposed. Secrets everywhere. $ICP changes the model: Application logic onchain. State onchain. Frontend onchain. Execution verifiable. Infrastructure decentralized. GlassWorm v2 is another warning. The software supply chain is under attack. Developer environments are under attack. Cloud credentials are under attack. The browser layer is under attack. The internet stack is tired. We do not just need another security plugin. We need infrastructure that reduces the number of places attackers can own. That is the $ICP thesis. Not hype. Not noise. Architecture. ♾️ $ICP by @dfinity is building the World Computer. If you like my content and want to support more $ICP research and awareness, contributions are welcome: ICP Donation Address: 1e672d038cebc619d93186418fa98f6499dbdb9cfdfac54f366c61a4a4ee4362 #ICP #InternetComputer #DFINITY #Cybersecurity #CloudSecurity #SoftwareSupplyChain #GlassWorm #VSCode #OpenVSX #DeveloperSecurity #OnchainCloud #WorldComputer #Canisters #ChainKey #Web3 #DecentralizedCloud #VerifiableCompute #CloudComputing #ZeroTrust #InfoSec #ThreatIntel #Blockchain #SovereignCloud thehackernews.com/2026/04/re…

Cursor AI editor found storing API keys and tokens in an unprotected SQLite database, allowing extensions to steal credentials silently. CVSS 8.2. No patch yet. #CursorAI #CyberSecurity #InfoSec #AI #DeveloperSecurity #APIKeys #CredentialTheft #Security securityonline.info/cursor-a…

Employees are installing AI tools, MCP servers, and LLM-connected browser extensions at a pace that most security teams can't keep track of. Many of these tools have access to source code, internal documents, or credentials. Anzenna inventories AI tools and MCP server installs across your fleet through scheduled device audits. Each one gets an AI-generated risk assessment evaluating what data it can access, who built it, and whether it poses a security concern. One security team discovered that a popular AI coding assistant installed on 40 developer machines was configured to send code context to a third-party API with no data processing agreement in place. Without the audit, it would have stayed invisible. Learn more: na2.hubs.ly/H03VD9R0 #AISecurity #ShadowIT #DeveloperSecurity #Cybersecurity #MCP

GlassWorm Supply Chain Attack Hits GitHub, npm, VSCode #glasswormcampaign #supplychainattack #developersecurity anavem.com/en/news/cybersecu…

Critical 9.7 CVSS flaw in TinaCMS (CVE-2026-28792) lets attackers hijack local developer machines via CORS and path traversal. Patch to 2.1.8 immediately. #TinaCMS #CVE #CyberSecurity #InfoSec #PathTraversal #CORS #DeveloperSecurity #Vulnerability securityonline.info/drive-by…

Secrets shouldn’t live in files. They should show up exactly when code runs. What’s new: 🔐 Runtime, read-only access to 1Password Environments ⚙️ Use via CLI & SDKs ⏱️Scoped, read-only access for automation using service accounts The result: fewer leaks, less friction, and safer dev workflows. 👉 Learn how this works and where it fits in your workflows: bit.ly/3MQx5Qo #DeveloperSecurity #SecretsManagement #DevOps #AItools #1Password

Your supply chain is your biggest attack surface. SonarQube's new security features empower customers to take a proactive stance, operationalizing SBOMs & adding malicious package detection to stop malware in the CI/CD. bit.ly/46jENt4 #AI #DeveloperSecurity #SonarQube