Today's top risks include Alibaba's ban on Claude AI-generated code due to backdoor concerns, highlighting the need for caution with AI development. Linux users on kernel 6.9+ face a significant data protection oversight as LUKS suspend no longer wipes disk-encryption keys from memory. Cisco confirmed active exploitation of a Unified Communications Manager vulnerability, urging immediate patching. Apple is shifting to compressed patch cycles, a direct response to AI drastically reducing exploit development time. FortiBleed actors are escalating their operations, collaborating with ransomware gangs after exploiting Fortinet firewalls and a Nextcloud zero-day to monetize access.

On the AI front, Google Cloud Tech is showcasing agentic AI transforming enterprise data to real-time action, speeding up AI agents by 80% on Gemini. They're also exploring the Multi-agent Communication Protocol (MCP) as a fundamental shift from traditional APIs, enabling complex multi-agent systems and internal collaboration.

Practical takeaways include prioritizing patching and reviewing AI policies, verifying Linux LUKS configurations, and securely evaluating agentic AI tools. Patch Cisco Unified CM, Fortinet, and Nextcloud immediately, prepare for faster Apple updates, and review internal policies for AI-generated code. For Linux kernel 6.9+, confirm LUKS suspend behavior for proper key wiping. Assess agentic AI tools with a strong security framework, understanding their data and system connections.

Today, we cover urgent threats and evolving security landscapes. A high-severity Microsoft SharePoint RCE, patched in May, is now actively exploited, demanding immediate verification of your patch status. New Android malware underscores the critical need for robust mobile device security and enhanced employee awareness regarding app downloads. We also discuss the complexities introduced by Cloudflare's new Monetization Gateway and the supply chain security implications of open-source IoT devices like the Oomwoo robot vacuum.

On the AI front, we explore how benchmarks like Senior SWE-Bench raise questions about securing AI-generated code and preventing its malicious use. Crafty, AI-powered phishing campaigns are adapting to device and OS fingerprints, making social engineering attacks more effective. Google Cloud's agentic AI, transforming enterprise data into real-time action, offers efficiency but requires careful consideration of agent permissions and data access.

Key takeaways include immediately patching the Microsoft SharePoint RCE, tightening mobile device security with refreshed user training, and evaluating the security and efficiency impact of AI agents.

Today We discuss the lifting of export controls on powerful AI models like Anthropic's Fable 5 and Mythos 5, highlighting both innovation potential and increased security considerations. The episode then dives into emerging AI threats, including the "BioShocking" prompt injection attack that tricks AI-powered browsers into compromising data, and the discovery of a fake Perplexity AI extension on the Chrome Web Store tracking user activity.

A significant traditional data breach at Aflac's Japan subsidiary, compromising 4.38 million customer records, serves as a stark reminder of the ongoing importance of supply chain security and third-party risk management. On a more positive note, we explore advancements in agentic AI and multi-agent systems, showcasing how AI is evolving to actively orchestrate workflows and drive efficiency.

This briefing is essential for small-business leaders, cyber professionals, MSPs, CISOs, GovCon leaders, and security operators. Key takeaways include the need to thoroughly review new AI models for operational relevance, educate teams on prompt injection attacks, and reinforce basic cyb

Today's episode unpacks escalating threats, starting with CISA's confirmation of active exploitation of the Windows BlueHammer privilege escalation vulnerability by ransomware gangs. We also detail the Blackfield gang's $2 million ransom demand from Nidec Corporation, highlighting the significant operational and supply chain risks. Nissan and the National Association of Insurance Commissioners (NAIC) disclose data breaches stemming from an Oracle PeopleSoft vulnerability, underscoring the pervasive threat of zero-day exploits and data theft.

Shifting to the future, we explore the rapid evolution of AI in cybersecurity. Discover Ornith-1.0, self-improving open-source models for agentic coding, and Google Cloud Tech's resources for building multi-agent systems. These developments signal a new era of automated defenses and potential attack vectors.

Our practical takeaways include urgently reviewing your exposure to critical vulnerabilities like BlueHammer and Oracle PeopleSoft, emphasizing immediate patching and data risk assessment. We also advise evaluating new AI and agentic tooling for internal automation or client delivery, and staying informed on strategic partnerships shaping the future of cloud and AI security.

Today, we cover a significant data breach at Japanese telecommunications operator KDDI Corporation, impacting 14.2 million email logins across multiple ISPs. This event highlights the pervasive threat of credential stuffing and the interconnectedness of digital infrastructure. We also discuss HackerRank's decision to open-source its Applicant Tracking System, exploring the potential security implications for hiring integrity and sensitive applicant data. Furthermore, OpenAI's report mapping Europe's AI workforce opportunity presents a crucial strategic risk, urging businesses to proactively strategize for AI's impact on talent and operations.

On the innovation front, we explore exciting AI and security automation developments. Google Cloud Tech's demonstrations of 'Antigravity' voice prompts, sub-agents, and multi-agent systems showcase how intelligent agents can streamline workflows and file organization, offering significant efficiency gains. Cloudflare's discussions on AI-powered cybersecurity with SentinelOne and democratizing cloud security with Wiz reveal how AI is becoming foundational in protecting digital assets, making advanced security more accessible for all.

Our three practical takeaways for today include: strengthening email defenses immediately with MFA and unique passwords, strategizing for AI’s workforce impact by assessing operational and talent needs, and exploring AI automation piloting for internal workflows or client delivery.

Today's episode covers the takedown of a major SIM-swapping gang in Poland, which exploited telecom vulnerabilities to steal millions in cryptocurrency. This incident underscores the persistent risk of carrier-level breaches. We also examine "Gaslight," new macOS malware designed to evade AI analysis tools by injecting misleading data, highlighting the evolving challenge of AI-aware adversaries. The "Bluekit" phishing-as-a-service platform shows a dangerous escalation with nearly 70 new hostnames and browser-in-the-middle capabilities, making real-time session interception a growing threat.

On a broader scale, we discuss the Akrites.org initiative, emphasizing the collective responsibility to secure the open-source supply chain. For those leveraging AI, we touch on Google Cloud Tech's resources for developing AI agents and integrating AI-native databases, alongside a cautionary note about vetting new open-source AI tools like OpenKnowledge. Key takeaways include reviewing open-source dependencies, strengthening SIM-swapping defenses with MFA and carrier locks, and educating teams about sophisticated phishing techniques.

Today's episode highlights severe vulnerabilities, including six new CVEs in Curl—a widely used component—and a Cisco SD-WAN zero-day actively exploited for root access. We detail how a malicious Edge extension, 'Edgecution,' bypassed browser sandboxes for ransomware, emphasizing the serious threat posed by extensions. CISA's urgent warnings on actively exploited, max-severity flaws in Ubiquiti UniFi OS and Lantronix devices underscore the immediate need for patching.

On a positive note, we cover Operation Endgame, a significant international disruption of Amadey and StealC malware infrastructure. In AI, we discuss Google Cloud Tech's new agentic updates, pushing AI agents toward greater autonomy, alongside Anthropic's accusation against Alibaba for illicitly extracting Claude AI model capabilities, highlighting critical IP and security challenges.

The key takeaways for listeners include reviewing exposure to Curl, Cisco, and Ubiquiti vulnerabilities, evaluating new agentic AI tools with a focus on data privacy, and bolstering browser security through updates, careful extension management, and user training.

Today's episode highlights the new Mistic backdoor, linked to the KongTuke ransomware access broker, actively targeting insurance, education, IT, and professional services sectors. We also cover the urgent need to patch a high-severity SSRF vulnerability in Cisco Unified Communications Manager (CVE-2026-20230), which is now under active exploitation. The recent Xolis healthtech data breach underscores the persistent danger of phishing attacks, compromising sensitive data for nearly 1.4 million individuals. We discuss the overwhelming volume of vulnerability reports and the need for smarter prioritization.

On the AI front, we explore the Haystack open-source AI framework and the importance of reviewing its security implications before implementation. Google I/O 2026 showcased significant advancements in AI agent stacks, moving beyond simple chatbots to complex, chained problem-solving and emphasizing "governed agents" for secure enterprise data interaction.

Our key takeaways for leaders are: prioritize immediate patching of critical systems, especially the Cisco Unified CM vulnerability; reinforce phishing awareness training and multi-factor authentication; and conduct thorough security reviews for all new AI tools and frameworks before integration.

Today's discussion highlights how INC Ransomware thrives by mastering basic attack techniques, particularly in high-pressure sectors like healthcare. This underscores the enduring importance of fundamental cybersecurity hygiene. We also delve into the coordinated nature of recent SSH brute-force attacks, emphasizing the non-negotiable need for strong passwords, MFA, and vigilant monitoring. The emergence of new open-source tools like 'Lore' and 'Adam' is examined, with a focus on the essential security scrutiny required when integrating new dependencies into your operational environment.

Shifting to the proactive side, we explore Google Cloud Tech's tutorials on building sophisticated AI agents for various applications, including long-running processes and multi-agent collaboration. Cybersecurity researcher John Hammond's "AI Cybersecurity Research Factory" demonstrates how AI can automate threat intelligence and vulnerability hunting, offering a glimpse into future security operations.

Key takeaways include reviewing your operational exposure for critical services like SSH, fortifying basic ransomware defenses, and evaluating new AI and agentic tools for internal automation or enhanced client delivery. Staying informed about emerging open-source projects and AI-driven security research is crucial for proactive defense.

Today's key risks include a max-severity Joomla Content Editor plugin flaw, actively exploited and requiring immediate patching by federal agencies and any organization using Joomla with Widget Factory. Microsoft is addressing a critical Defender zero-day, "RoguePlanet," highlighting the ongoing need for widespread security updates. Kodak is investigating a data breach claimed by ShinyHunters, reinforcing that no entity is immune. Developers face specific threats with 15 malicious plugins on the JetBrains Marketplace designed to steal AI API keys, emphasizing the need for rigorous plugin vetting. Android users are targeted by Rokarolla, a new banking trojan impacting 217 banking and crypto apps, underscoring the importance of mobile app vigilance.

We also explore the accelerating AI landscape, with Google Cloud Tech's insights on building long-running AI agents and Cloudflare's discussion on "Shipping Agents with Cloudflare: Secure by Default," emphasizing security by design in the agentic future. GLM-5.2 is now a leading open-weights model, showcasing rapid AI advancements.

Practical takeaways include patching the Joomla JCE plugin immediately if applicable, developers scrutinizing all plugins for AI API key theft, and all users exercising extreme caution with mobile app downloads, sticking to official stores, and checking permissions.