In partnership with
Win AI Search Without a Big Team
92% of VCs use AI to find companies. 58% of buyers start there, too. If you're not showing up in AI answers, you're invisible before the conversation even starts. Join HubSpot for Startups, Anthropic, and Marketing Against the Grain on July 30 (11 am ET) for a live AEO teardown. Real startup. Real recs. Register and unlock the free Startup Visibility Bundle.
This is not a “single patch” week…
This is a “how many trusted systems can attackers bend before lunch” week.
Let’s dive in.
Risk Level: Critical
Business Impact: SharePoint compromise can expose internal documents, authentication material, machine keys, and long-term access paths into collaboration infrastructure.
What You Need to Know
Attackers are exploiting CVE-2026-50522, a critical SharePoint remote code execution flaw tied to deserialization of untrusted data. SecurityWeek’s SharePoint report notes this is the fourth SharePoint vulnerability exploited in the past month’s wave of attacks, with threat actors reportedly stealing machine keys to retain long-term access.
Why This Matters
SharePoint is where internal documents, workflows, and business context live.
Machine key theft can enable persistence even after patching.
Repeated exploitation of the same platform means attackers are actively investing in this lane.
Executive Actions
🩹 Patch SharePoint immediately and verify the fixed build is running.
🔑 Rotate machine keys and credentials on any exposed or potentially compromised assets.
🔎 Hunt for suspicious SharePoint requests, unexpected code execution, and abnormal file access.
🧱 Restrict SharePoint administrative access and review external publishing exposure.
Risk Level: Critical
Business Impact: ServiceNow compromise can expose tickets, workflows, asset data, user records, automation paths, and internal operational context.
What You Need to Know
A critical ServiceNow AI platform vulnerability, CVE-2026-6875, is reportedly being exploited shortly after technical details became public. SecurityWeek’s ServiceNow coverage describes the issue as a sandbox escape that can allow unauthenticated attackers, under certain conditions, to execute arbitrary code, with self-hosted customers needing to apply patches themselves.
Why This Matters
ITSM platforms contain the map of your environment.
Tickets often include system names, escalation paths, admin notes, and troubleshooting context.
AI platform vulnerabilities inside ServiceNow expand the blast radius from workflow data to code execution risk.
Executive Actions
🧯 Apply ServiceNow patches immediately, especially for self-hosted environments.
🔎 Review instance logs for unusual execution, table access, exports, and workflow changes.
🔐 Restrict sensitive table permissions and reduce broad administrative access.
📣 Alert Service Desk teams to expect more credible impersonation attempts using operational details.
Risk Level: Critical
Business Impact: VPN compromise can lead to remote access, credential theft, lateral movement, and domain-wide ransomware deployment.
What You Need to Know
Qilin ransomware operators are exploiting CVE-2026-0257, a PAN-OS GlobalProtect authentication bypass, as an initial access vector. Arctic Wolf’s investigation details multiple intrusions where attackers moved from perimeter compromise to domain-wide encryption, while The Hacker News’ coverage summarizes the active ransomware use case.
Why This Matters
VPN appliances remain one of the fastest routes into the enterprise.
Authentication bypass means attackers may not need valid credentials to get started.
Qilin has shown it can move quickly from access to encryption.
Executive Actions
🩹 Patch PAN-OS and validate GlobalProtect gateway and portal versions.
🔒 Restrict VPN access using MFA, certificates, device posture, and geo/IP controls.
🔎 Hunt for anomalous VPN sessions, new source geos, and unusual post-login activity.
🔑 Rotate credentials for users who authenticated through vulnerable gateways during the exposure window.
Leadership Insight:
This week’s message is not subtle: attackers are targeting the places where trust turns into access.
SharePoint stores the business. ServiceNow maps the business. GlobalProtect opens the business. AI agents build the business. Microsoft 365 communicates the business. Ransomware disrupts the business.
The executive takeaway is simple: if a system can store context, grant access, automate work, or recover operations, it belongs in the security boundary.
Stop typing what you could say in 10 seconds.
Wispr Flow turns your voice into clean, professional text inside any app. Emails, Slack, client updates — speak once, send without editing. 4x faster than typing.
Risk Level: High
Business Impact: AI coding agent compromise can expose source code, developer secrets, CI/CD tokens, cloud credentials, and trusted host tooling.
What You Need to Know
Researchers demonstrated sandbox escapes across Cursor, Codex, Gemini CLI, and Antigravity by having AI agents write files that trusted host tools later execute, load, or process outside the intended sandbox. BleepingComputer’s sandbox escape report explains the architectural problem, while TechRadar’s coverage notes that the issue affects assumptions around how secure agentic development tools really are.
Why This Matters
AI agents can write files that later influence trusted tools outside the sandbox.
Developer environments are packed with secrets, repo access, and deployment permissions.
“The agent followed the rules” is not reassuring if host tools execute the outcome.
Executive Actions
🤖 Disable auto-approval for AI coding agents in privileged repositories and CI/CD workflows.
🔐 Limit agent access to secrets, shell execution, and write-capable repository scopes.
🧪 Treat README files, Makefiles, config files, and dependency scripts as untrusted agent input.
📦 Require human review before agents modify workflows, execute commands, or publish artifacts.
Risk Level: High
Business Impact: Microsoft 365 abuse can hide command-and-control traffic inside legitimate collaboration workflows, complicating detection and response.
What You Need to Know
A malware component called HollowGraph uses Microsoft Graph API and a compromised Microsoft 365 calendar as a two-way dead drop for command-and-control. SecurityWeek’s HollowGraph report explains that attackers plant tasking as calendar events while the implant exfiltrates data by creating its own encrypted calendar attachments.
Why This Matters
Calendar traffic looks legitimate, which makes C2 harder to distinguish from normal SaaS use.
Microsoft 365 accounts are increasingly being used as attacker infrastructure.
Collaboration platforms are now both data targets and communication channels for malware.
Executive Actions
📅 Monitor for unusual calendar events, far-future dates, abnormal attachments, and Graph API spikes.
🔐 Review app registrations, client secrets, and delegated Microsoft Graph permissions.
🔎 Hunt for suspicious files like logAzure.txt, unusual Entra ID usage, and DNS tunneling.
🧾 Treat compromised Microsoft 365 accounts as infrastructure incidents, not just mailbox incidents.
Risk Level: High
Business Impact: Ransomware against production and supply chain operations can disrupt business continuity, expose confidential data, and create extortion pressure.
What You Need to Know
The Anubis ransomware group claims it stole 1 TB of confidential data from Coca-Cola subsidiary Fairlife and is threatening to leak it unless a ransom is paid. SecurityWeek’s Fairlife ransomware report notes that Coca-Cola had already disclosed production disruption at Fairlife, and that Anubis uses double extortion with a wiper-mode capability.
Why This Matters
Ransomware is still a business interruption problem first and a data theft problem second.
Production disruption creates revenue, logistics, customer, and reputational impact.
Wiper-style features raise the stakes because paying may not restore operations.
Executive Actions
🧯 Validate offline and immutable backups for production and supply chain systems.
📊 Confirm restore timelines for manufacturing, logistics, and revenue-impacting workflows.
🔐 Review privileged access into OT-adjacent and production support systems.
📣 Pre-stage legal, comms, and executive decision paths for extortion and data leak pressure.
🩹 Patch SharePoint, ServiceNow, and PAN-OS GlobalProtect immediately based on validated exposure
🔑 Rotate SharePoint machine keys, VPN credentials, and Microsoft 365 app secrets where compromise is plausible
🔎 Hunt for SharePoint machine key theft, ServiceNow exploitation, suspicious VPN sessions, and HollowGraph-style calendar abuse
🤖 Restrict AI coding agents: no auto-approval, limited secrets, and human review for workflow changes
🧯 Validate ransomware recovery paths for production, supply chain, and business-critical systems
📊 Require proof of running fixed versions and credential rotation, not just patch ticket closure
💡 If your collaboration platform, ticketing system, VPN, AI agents, calendar, and recovery plan all need attention in the same week, that is not chaos. That is the business telling you where the real perimeter lives. 💡
J.W.
(P.S. Check out our partners! It goes a long way to support this newsletter!)
Win AI Search Without a Big Team
92% of VCs use AI to find companies. 58% of buyers start there too. If you're not showing up in AI answers, you're invisible before the conversation even starts. Join HubSpot for Startups, Anthropic, and Marketing Against the Grain on July 16 (11am ET) for a live AEO teardown. Real startup. Real recs. Register and unlock the free Startup Visibility Bundle.
