In partnership with

⏱️ ≈ 8-minute read

❝

Editor’s Note: This week’s cyber news had range. Wiper malware hit energy systems, Vercel’s third-party mess kept expanding, Cisco firewalls learned a terrifying new party trick, and telecom surveillance operators apparently decided “long-known flaw” still means “fresh opportunity.” Somewhere, a risk register just quietly asked for PTO.

One ask before you dive in: if this recap earns a forward or sparks a debate on your team, share it. If there’s something you think we should cover (or skip) next time, hit reply and tell us. Thanks for reading!

📬 This Week’s Clickables

📌 Big News: Venezuela energy wiper attacks, Vercel’s expanding breach fallout
🚨 Can’t Miss: Cisco Firestarter malware, Microsoft Defender zero-day, exposed SharePoint servers, Nginx UI exploitation, Bitwarden NPM compromise
🤖 AI in Cyber: Zealot cloud hacking agents, ATHR AI vishing, Google Antigravity sandbox escape, ZionSiphon hype check
🧪 Strange Cyber Story: Commercial surveillance tools abusing telecom signaling systems

🚨 Big News

🛢️ Lotus Wiper Turns Venezuela’s Energy Sector Into a Burn Bag

Intro: A previously unknown wiper called Lotus Wiper targeted Venezuela’s energy and utilities sector. No ransom note, no negotiation arc, just destructive malware doing destructive malware things.

What Happened: Kaspersky researchers found that Lotus Wiper was built to erase data across physical drives and delete files throughout system storage. The attackers appeared to focus on older Windows systems, suggesting they understood the target environment before launching the destructive phase. Technical clues indicate the campaign may have been staged months earlier, with malware compiled in late September 2025. The attacks landed amid regional tension and after Venezuela’s state oil company PDVSA reported a separate cyberattack.

Why It’s Important: Wipers against critical infrastructure are not just an IT cleanup issue. They threaten recovery, continuity, and public-sector resilience. For energy and utility environments, “restore from backup” only works if backups are segmented, tested, and still exist.

The Other Side: Researchers have not named the victims or attributed the campaign to a specific actor. There is also no confirmed link between Lotus Wiper and the PDVSA incident. Attribution remains messy, because apparently cyber incidents are allergic to clean answers.

👉 Takeaway: Treat destructive malware as a business-continuity threat, not just a detection problem. Backup isolation, recovery testing, and visibility into old Windows assets matter before the wiper shows up.

TL;DR: Lotus Wiper targeted Venezuela’s energy and utilities sector with destructive malware built to erase systems and disrupt recovery.

Further Reading: The Record

This docs platform just raised $45M

Mintlify powers documentation for 20,000+ companies reaching 100M+ people a year. Backed by a16z and Salesforce Ventures. Your docs deserve the same infrastructure.

Get Started for Free Today

🧩 Vercel’s Breach Fallout Keeps Finding New Rooms in the House

Intro: Vercel’s incident has moved past “one bad integration” territory. The fallout now touches customer accounts, third-party systems, stolen tokens, and the usual SaaS permission spaghetti.

What Happened: Vercel said its investigation found more compromise evidence across its customer base after analyzing nearly a petabyte of logs. The company still describes the number of affected accounts as small, but has not provided a precise count. Threat intelligence pointed to malware searching infected machines for valuable tokens tied to Vercel and other providers. CyberScoop reported that the attack began with Context.ai, a third-party AI tool used by a Vercel employee, before attackers used stolen access to move through systems and decrypt stored customer data.

Why It’s Important: This is a sharp reminder that developer platforms inherit risk from every connected tool, token, and integration. OAuth access, environment variables, and API keys can turn one compromised endpoint into a downstream exposure event. “Small number of accounts” can still carry very large consequences.

The Other Side: Vercel says it found no evidence that its published software packages were tampered with. It also says some customer compromises appear unrelated to Vercel systems. That is good news, though still the kind that arrives with a spreadsheet of action items.

👉 Takeaway: Teams should review OAuth-connected apps, rotate exposed secrets, and tighten environment variable handling. Trust is not a control, even when the dashboard looks expensive.

TL;DR: Vercel’s breach fallout expanded across customers and third-party systems, with stolen tokens and SaaS trust relationships driving the risk.

Further Reading: CyberScoop

Stuxnet, publicly discovered in 2010, is widely regarded as one of the first known malware operations designed to cause physical damage to industrial equipment. Source: CISA, “Stuxnet Malware Mitigation”

Stay in flow state. Dictate everything else.

Context switching kills your focus. Every time you stop coding to type a Slack reply, write a ticket, or draft a PR description, it takes 23 minutes to get back in the zone.

Wispr Flow lets you dictate all of it without leaving your editor. Speak your response, your ticket, your commit message — Flow formats it and you're back to coding. Works system-wide inside Cursor, VS Code, Warp, Slack, Linear, and every app.

4x faster than typing. 89% of messages sent with zero edits. Used by engineering teams at OpenAI, Vercel, and Clay.

Start flowing free

🔥 Can’t Miss

🔥 Cisco Firewalls Get Haunted by Firestarter Malware
State-sponsored hackers implanted Firestarter malware on Cisco security devices, with persistence that could survive firmware updates and normal reboots. CISA and the U.K.’s NCSC warned the backdoor was found on a U.S. federal civilian agency device and tied to activity dating back to late 2025. Cisco recommends updated software and reimaging where compromise is suspected, which is corporate-speak for “patching alone may not save you.”
👉 Key takeaway: Edge devices need compromise checks, not just patch checks.
🪟 Microsoft Defender Bug Becomes an Attacker Privilege Escalator
A Microsoft Defender flaw, CVE-2026-33825, was exploited in the wild as a zero-day using public proof-of-concept code. The bug can let low-privilege attackers access the SAM database, extract NTLM hashes, and gain SYSTEM privileges. Microsoft patched it on April 14, and CISA ordered federal agencies to patch by May 6.
👉 Key takeaway: Public exploit code turns patch windows into attacker windows.
🧱 SharePoint Servers Remain Exposed Because Apparently Patch Tuesday Was Optional
More than 1,300 internet-exposed SharePoint servers remained vulnerable to CVE-2026-32201, a zero-day spoofing flaw. Microsoft says exploitation could let unauthenticated attackers perform network spoofing through improper input validation. CISA added the bug to its KEV catalog and ordered federal agencies to patch by April 28.
👉 Key takeaway: Exposed SharePoint is not legacy noise. It is an invitation with a corporate logo.
🌐 Nginx UI Flaw Turns Management Consoles Into Takeover Targets
Attackers are exploiting CVE-2026-33032, a critical remote takeover flaw in the Nginx UI management tool. The issue is tied to the tool’s newer AI MCP integration and can give unauthenticated attackers a path to crafted requests. Researchers reported more than 2,600 exposed instances, with public proof-of-concept code already available.
👉 Key takeaway: Internet-facing admin tools deserve exposure reviews before attackers do the review for you.
📦 Bitwarden’s NPM Package Gets Dragged Into the Supply Chain Swamp
The Bitwarden CLI NPM package was compromised in a broader open source supply chain attack. Security firms said the malicious package downloaded JavaScript designed to steal credentials, secrets, cloud tokens, GitHub data, SSH material, and AI tooling configuration. Bitwarden said it found no evidence that end-user vault data or production systems were accessed.
👉 Key takeaway: Developer machines remain high-value targets because one bad package can raid a lot of secrets fast.

Gladly Connect Live '26. May 4–6 in Atlanta.

The room you want to be in. This is where CX leaders are tackling the hard AI questions and sharing what's actually working. For CX and ecommerce leaders. Atlanta, May 4–6. Space is limited — secure your spot now.

🤖 AI in Cyber

🧠 Zealot Shows AI Agents Can Hack Cloud Labs With Very Little Hand-Holding
Palo Alto Networks Unit 42 built Zealot, a multi-agent proof of concept for autonomous cloud attacks. In an isolated Google Cloud lab, Zealot scanned the network, found a vulnerable VM, exploited a web app flaw, stole credentials, escalated privileges, and exfiltrated BigQuery data. It even improvised persistence with SSH keys, which is impressive in the same way a raccoon learning door handles is impressive.
👉 Key takeaway: Cloud defenders should plan for machine-speed recon, exploitation, and privilege abuse.
☎️ ATHR Automates Vishing So Scammers Can Scale Their Fake Support Desk
ATHR is a cybercrime platform that uses AI voice agents and human operators to automate voice phishing. The service targets credentials and verification codes for major email, crypto, and cloud accounts. It packages the attack flow from lure emails to guided phone calls, lowering the skill needed to run convincing vishing campaigns.
👉 Key takeaway: AI vishing is becoming packaged crimeware, so training needs to move beyond “check the email header.”
🧪 Google Antigravity’s Secure Mode Learns Prompt Injection Has No Respect for Boundaries
Pillar Security researchers found a Google Antigravity vulnerability that combined prompt injection with file-creation behavior to enable remote code execution. The bug bypassed secure mode, which is supposed to sandbox commands and limit network access. Google patched the issue after receiving the report in January.
👉 Key takeaway: Agentic AI tools need hard security boundaries, not polite guardrails wearing a helmet.
🚰 ZionSiphon Is an AI Malware Reality Check, Not an Apocalypse Trailer
Dragos pushed back on claims that ZionSiphon, a malware sample reportedly aimed at Israeli water infrastructure, posed a serious operational technology threat. The malware appeared to use AI-generated code, but researchers said it contained hallucinated paths, fictional process names, and flawed assumptions about water systems. The result was more broken threat theater than industrial sabotage masterclass.
👉 Key takeaway: Track AI-enabled threats, but do not let hype distract from proven critical infrastructure risks.

🧟‍♂️ Strange Cyber

📡 Commercial Spy Tools Are Apparently Cosplaying as Telecom Operators Now

Intro: The strangest story this week comes from telecom signaling systems, where surveillance campaigns allegedly exploited long-known weaknesses to track people through mobile networks. The roaming infrastructure is doing spy-thriller side quests now. Cool, cool.

What Happened: Citizen Lab researchers mapped surveillance activity to mobile operator signaling infrastructure, showing how commercial surveillance tools abused SS7 and Diameter weaknesses. The campaigns mimicked mobile operator identities while routing and hiding traffic through global telecom systems. Researchers linked activity to infrastructure and identifiers tied to operators across multiple countries. The campaigns shifted between signaling protocols that have long been associated with mobile location-tracking risk.

Why It’s Important: Telecom signaling relies on trust relationships between operators, and that trust model keeps turning into a surveillance loophole. The risk is technical, regulatory, and geopolitical. When spyware traffic can blend into roaming infrastructure, accountability gets blurry fast.

The Other Side: Researchers could not identify the specific vendors or customers behind the campaigns. Citizen Lab also warned that operator signaling addresses do not automatically prove operator involvement. Access may come through third parties, leasing arrangements, or intermediary services.

👉 Takeaway: Mobile network security still has a trust problem, and “known for years” does not mean “fixed.” Carriers and regulators need better visibility into signaling abuse before ghost operators become a permanent feature.

TL;DR: Commercial surveillance campaigns abused long-known telecom signaling weaknesses to track targets while hiding inside global mobile network traffic.