Today's Cybersecurity Pulse
Anthropic CEO meets White House over federal access to Mythos AI
Anthropic CEO Dario Amodei will meet White House chief of staff Susie Wiles to discuss government access to the company's Mythos AI model, which can discover and exploit zero‑day vulnerabilities. The meeting follows a Pentagon‑imposed blacklist after Amodei refused to lift safety restrictions, while Treasury, intelligence agencies and CISA are already testing the model.
Also developing:
By the numbers: Artemis raises $70M in combined seed and Series A round
When Silicon Got Serious About Security
The article traces cryptography’s evolution from the 1970s Data Encryption Standard (DES) to today’s Advanced Encryption Standard (AES) and Elliptic Curve Cryptography (ECC). It highlights how silicon’s exponential speed gains exposed DES’s 56‑bit weakness, prompting the 1998 Deep Crack break. The piece also covers the 1990s Crypto Wars that cemented encryption as protected speech and notes the looming quantum threat to RSA. Ultimately, it underscores a lasting philosophy: encryption should be open, mathematically sound, and treated as a public good.
Exclusive-Russia Supplies Iran with Cyber Support, Spy Imagery to Hone Attacks, Ukraine Says
Russian reconnaissance satellites conducted at least 24 passes over 46 military and critical sites in 11 Middle Eastern countries during March 21‑31, sharing high‑resolution imagery with Iran. The data preceded Iranian missile and drone attacks on bases, including a strike on...
FortiClientEMS Vulnerabilities Under Active Exploitation, Expose Systems to RCE
Fortinet disclosed two critical vulnerabilities in its FortiClientEMS endpoint management platform that are already being exploited in the wild. CVE-2026-21643 is a SQL injection flaw in the admin interface of version 7.4.4, allowing unauthenticated remote code execution. CVE-2026-35616 is an...
Trump Administration Releases Cyber Strategy
The Trump administration unveiled a new Cyber Strategy for America in March 2026, paired with Executive Order 14390 to intensify federal action against cybercrime. The strategy outlines six pillars—deterrence, streamlined regulation, federal network modernization, critical‑infrastructure protection, technology superiority, and talent...
$20 Billion Lost to Cybercrime as AI and Investment Scams Surge: FBI Report
The FBI’s 2025 Internet Crime Report records a historic $20.8 billion in losses from more than one million complaints. Cyber‑enabled fraud accounted for roughly $17.7 billion, with investment scams—especially cryptocurrency schemes—driving $8.6 billion of that damage. AI‑powered scams emerged as a new threat,...
The Case for Fixing CWE Weakness Patterns Instead of Patching One Bug at a Time
CWE is transitioning from a background taxonomy to a core element of vulnerability disclosure, with a growing share of CVE records now including CNA‑provided CWE IDs. Precise, lower‑level CWE mappings are improving root‑cause visibility, enabling teams to target systemic weakness...
Jones Day Law Firm Says Hackers Accessed Some Clients’ Data
Jones Day disclosed that the cyber‑criminal group Silent breached its network, accessing dated files for ten clients. The intrusion stemmed from a phishing attack, and the firm confirmed that all impacted clients have been notified. Hackers also exfiltrated internal data...
MyRepublic Launches Email Guard for Singapore SME Protection
MyRepublic has launched Email Guard, a managed email security service tailored for Singapore’s small and medium‑sized enterprises. The solution, built on Check Point’s threat‑prevention technology, safeguards against phishing, malware, and business‑email compromise across Microsoft 365 and Google Workspace. It offers...
Windows 11 Is Phasing Out Old Kernel Drivers: Microsoft Will Permanently End Cross-Signing in April
Microsoft announced that, starting with the April 2026 security update, Windows 11 will permanently stop trusting kernel drivers signed under the old Cross‑Signed Program. The change applies to Windows 11 24H2, 25H2, 26H1 and Windows Server 2025, allowing only drivers vetted through the Windows...
.jpg)
Don’t Wait for 2027 – Make Your Smart Product Security Visible Now
Australia is introducing a national security label for smart devices, slated for a 2027 launch, with an industry pilot beginning in October 2026. The label will provide a clear, comparable security credential at the point of sale, shifting security from...
Cambodian Parliament Passes Landmark Cybercrime Law
Cambodia's parliament approved its first cybercrime law targeting scam centres that have defrauded international victims of billions. The legislation prescribes prison terms of two to ten years and fines up to $250,000 for large‑scale operations. It also criminalises money‑laundering, data...
AI-Assisted Fraud Makes Big Debut in FBI's Cybercrime Stats
The FBI’s 2025 Internet Crime Complaint Centre report introduced AI‑assisted fraud as a distinct category, documenting $893 million in losses. Overall cybercrime losses topped $20 billion, a 26 percent rise from 2024, with investment scams leading at $8.6 billion. AI‑generated content fueled business‑email‑compromise, voice‑cloning,...
Jones Day Confirms Limited Breach After Phishing Attack by Silent Ransom Group
Jones Day, one of the nation’s top law firms, confirmed a limited data breach after the Silent Ransom Group (SRG) posted files for ten clients on a dark‑web leak site. The attackers demanded roughly $13 million to delete the stolen data...
Turn Fear Into Action: Strengthen Security with AI
I was afraid this afternoon. Read a security report that a massive cyberattack is coming. Fear causes the human mind to do weird things. My mom, in 1988, thought a massive nuclear war was coming, so joined a Montana suvivalist cult. Had...
Flowise AI Agent Builder Faces Active CVSS 10 RCE Attack
Flowise AI Agent Builder Under Active CVSS 10.0 RCE Exploitation; 12,000+ Instances Exposed https://t.co/aINT8EHBFi https://t.co/SKA564pKd5
Malwarebytes Passes First Independent No-Logs VPN Audit, Boosting SaaS Trust
Malwarebytes announced that its Privacy VPN has cleared a rigorous, independent audit by security firm X41 D‑Sec, which found no evidence of user‑activity logging. The two‑month white‑box assessment also uncovered and helped fix a critical vulnerability, underscoring the company's commitment...
Unpatched Claude Coding Flaw Lets Attackers Steal Cowork Files
Attackers can exfiltrate user files from Cowork by exploiting an unremediated vulnerability in Claude’s coding environment, which now extends to Cowork. The vulnerability was first identified in https://t.co/noHjpUqN1I chat before Cowork existed by Johann Rehberger, who disclosed the vulnerability. It...
Iran's Cyber Arsenal Now Targets Critical Infrastructure Worldwide
Iran has rapidly developed advanced cyber capabilities, evolving from information gathering to conducting destructive, state-linked attacks against critical infrastructure in the U.S., Israel, and the Gulf states. https://t.co/XlKdD8VuZu
Claude Code Leak: Researchers Find First Vulnerability
Anthropic unintentionally published a source map that revealed roughly 512,000 lines of Claude Code's TypeScript source. Researchers used the leak to uncover a critical flaw allowing command‑chain bypass of the tool's deny‑rule system after 50 subcommands. The vulnerability could let...
New Threat Evolves Script Kiddies, Targets Everyday Users
read this when it first came out.. my first thought was crystal clear: "This is simply the natural evolution of SCRIPT KIDDIES but on steroids.." my second thought was broader: "The majority of every day people, will indeed at some point very soon, fall...
New Cyber Threats Target Power Grid Infrastructure
All emerging cyber threats targeting power infrastructure at a glance #energysky -- via pv magazine usa: https://t.co/CgfOPDFHMV
China Ready to Interfere in Local Elections: NSB
Taiwan’s National Security Bureau warned that Beijing is poised to interfere in the November nine‑in‑one local elections through a coordinated hybrid campaign. The report details more than 173 million cyber attacks on the government service network, 13,000 suspicious online accounts and...
Pipeline Security Lessons From March Supply Chain Incidents
Between March 19 and March 31, 2026, the TeamPCP threat group executed four supply‑chain attacks that compromised the open‑source scanner Trivy, the IaC scanner Checkmarx KICS, the AI model gateway LiteLLM, and the JavaScript client axios. Each breach leveraged malicious...
ATO Adds In-App Call Verification to Stop Scams
The Australian Taxation Office (ATO) has introduced a new in‑app “verify call” feature that lets taxpayers confirm whether a phone call claiming to be from the ATO is authentic within 30 seconds. The tool, available on iOS and Android, pushes...
2027 POTUS Budget Proposal Targets CISA With Funding Cuts
The FY2027 White House budget proposes cutting the Cybersecurity and Infrastructure Security Agency’s (CISA) funding by up to $707 million, reducing its budget to just over $2 billion. The administration frames the reductions as a strategic realignment that narrows CISA’s focus to...
AI-Assisted Supply Chain Attack Targets GitHub
A threat actor used AI‑assisted automation to launch the "prt‑scan" supply‑chain campaign on GitHub, opening over 500 malicious pull requests between March 11 and early April. The campaign targeted repositories that use the vulnerable pull_request_target workflow, compromising fewer than 10 %...
Autonomous Agents Pose a New Large‑scale Threat
An agent with a goal and agency can do real damage. We used to worry about compromised accounts. Now we need to worry about autonomous decision-making at scale. That is a very different risk model.
Scammers Posing as Federal Officials Drive Complaints up and Rack up $800 Million in Losses
The FBI’s 2025 Internet Crime Complaint Center report shows government‑impersonation scams nearly doubled from 2024, with complaints rising from about 17,300 to 32,500. Victims lost roughly $797 million in 2025, up from $405 million the year before, placing this fraud among the...
Axios Attack Shows Complex Social Engineering Is Industrialized
The popular JavaScript HTTP client Axios was compromised when North Korean state‑linked group UNC1069 socially engineered lead maintainer Jason Saayman into installing a malicious dependency. The attackers delivered a remote‑access Trojan via a fake Slack workspace and Microsoft Teams call,...
Maine House Advances McCabe Bill to Strengthen Cybersecurity at Maine Hospitals
The Maine House unanimously advanced Rep. Julie McCabe’s LD 2103, mandating hospitals adopt cybersecurity plans aligned with DHS and CISA best practices. The bill requires prompt law‑enforcement notification, backup communication systems, and annual staff training. It responds to spring cyber‑attacks that...
Microsoft Links Medusa Ransomware Affiliate to Zero-Day Attacks
Microsoft has identified Storm-1175, a China‑based financially motivated cybercrime group, as an affiliate of the Medusa ransomware operation. The gang is now leveraging both known (n‑day) and previously undisclosed (zero‑day) vulnerabilities in rapid, high‑velocity attacks. Microsoft’s intelligence shows Storm-1175 can...
Fortinet Issues Emergency Patch for FortiClient Zero-Day
Fortinet issued an emergency hotfix for the critical CVE‑2026‑35616 zero‑day in its FortiClient Endpoint Management Server, a 9.1‑CVSS flaw that enables unauthenticated code execution. The vulnerability has already been exploited in the wild, prompting a security advisory that recommends immediate...
Radim Marek: Don't Let Your AI Touch Production
AI coding agents now generate SQL that looks correct but often ignores execution plans, locking behavior, and data distribution, leading to costly production incidents. Radim Marek argues that the missing piece is real‑time awareness of the production schema, including table...
Anthropic's Claude Code Leak: Should RIA Firms and Advisors Be Worried?
Anthropic accidentally exposed the raw instruction set behind its Claude Code model on GitHub, prompting a rapid takedown effort. No personally identifiable information was leaked, but the incident reveals gaps in the company’s internal security controls. Wealth‑tech advisors are urged to...
LinkedIn Faces Scrutiny Over 'BrowserGate' Script Scanning 6,236 Extensions
A report by Fairlinked e.V. alleges LinkedIn injects a hidden JavaScript that checks for 6,236 Chrome extensions and harvests detailed device telemetry. LinkedIn says the script is meant to block scraping tools, but regulators and users are questioning the privacy...
5 Email Myths That Are Quietly Damaging Your Brand’s Reputation
Retailers are enjoying AI‑driven personalization, yet 27% remain in a DMARC enforcement gap, exposing them to domain spoofing. Valimail’s 2026 State of DMARC report shows many have only reporting‑only records, which lets attackers use their brand in AI‑generated phishing emails....
CNET Study Shows 54% of US Laptop Users Face Malware, 88% Take Action
CNET’s latest survey reveals that 54% of U.S. adults with personal laptops have encountered potential malware in the last year, while 88% reported taking action. The findings highlight a gap between built‑in antivirus tools and user vigilance, prompting calls for...
Hims & Hers Says Limited Data Stolen in Social Engineering Attack
Hims & Hers disclosed a sophisticated social‑engineering breach that compromised its third‑party customer‑service platform from February 4‑7, 2026. Hackers accessed service tickets, exposing customer names and email addresses, but the firm confirmed that electronic medical records and provider communications were untouched....
New Cyber Strategy Shifts Attention to Cloud and Supply Chain Security
The White House released a new National Cybersecurity Strategy on March 6, 2026, shifting federal priorities toward cloud data protection and software supply‑chain security. While zero‑trust, AI security, and post‑quantum cryptography remain core, the strategy mandates faster cloud migration and...
Disgruntled Researcher Leaks “BlueHammer” Windows Zero-Day Exploit
A security researcher known as Chaotic Eclipse publicly released exploit code for a previously private Windows privilege‑escalation vulnerability dubbed BlueHammer. The flaw, a local privilege escalation combining a TOCTOU and path‑confusion bug, allows a local attacker to obtain SYSTEM or...
New Mexico’s Meta Ruling and Encryption
A New Mexico judge ruled that Meta’s 2023 addition of end‑to‑end encryption to Facebook Messenger created liability because predators could use the shielded platform to groom minors. The state is seeking court‑mandated changes that could force Meta to weaken or...
How to Combat Cyber-Enabled Cargo Theft: Insights From NMFTA
At its Spring Meeting in Savannah, the NMFTA highlighted the growing threat of cyber‑enabled cargo theft. A panel with Werner Enterprises, Johanson Transportation Service and NMFTA’s cybersecurity director stressed that any cargo theft should trigger immediate involvement of a company’s...
Cybercriminals Accelerate: Storm-1175 Beats Patches in Days
It is not every day that a financially motivated threat actor manages to move faster than the vendors trying to secure their products. Yet that is precisely what Microsoft says Storm-1175 has been doing. The China-based cybercriminal group, closely associated...
Y Combinator Leader Overlooks OpenClaw Security Vulnerabilities
Wild: the head of Y Combinator seems pretty blind to the security risks in OpenClaw.
Sherlock Bug Bounty for Aave V4 Now Live
Sherlock bug bounty for Aave V4 is live. Learn more below about the program and scope.
Negligence Fuels Social Engineering Attacks, Accountability Needed
"It's mainly social engineering attacks. ... If you're grossly negligent, you should definitely be held accountable." https://t.co/8bYXWatFF8
Questioning TestFlight Use After Drift Hack Tactics
"Do you stay away from TestFlight right now?" -- @perkinscr97 on the tactics used in the Drift hack https://t.co/8bYXWatFF8
Nation‑state Attacks on Startups Guarantee Their Own Victory
"When a nation-state attacks a startup, the nation-state is going to win every single time." -- @perkinscr97 https://t.co/jfZlSDnB86
Seal911 Success Highlights Need for Stronger Endpoint Security
"The fact that that Seal911 has been the saving grace for a bunch of teams shows that people could put more resources into better endpoint security." -- @llewellenmichael https://t.co/8bYXWatFF8
Enterprise Domain Management Requires Strict Renewal and Security Processes
Hive mind - how do large enterprises manage their domains? I'd love to talk to some IT leaders - what processes ensure https://t.co/VFOQyhfres or https://t.co/kuiRO5DwUb or https://t.co/N7kgQgrxQy doesn't expire, or get socially engineered into a redirect or transfer?