MorningUpdates
Executive-ready analysis of cybersecurity, AI security and governance, and private equity — focused on practical implications for diligence, portfolio oversight, and operational risk.
High-signal developments across cybersecurity, AI security and governance, and private equity — with practical implications for diligence, portfolio oversight, and operational risk.
Executive Summary
Top Cybersecurity Incidents and Trends
ConnectWise suspends Cloud Backup after supply-chain access abuse.
ConnectWise shut down Cloud Backup, SaaS Security, and Cloud Management services after unauthorized access tied to a compromised Cloud Backup credential; Huntress observed unauthorized sign-ins in 162 tenants and Microsoft warned that attackers likely abused previously granted customer permissions, including highly elevated Microsoft 365 permissions.
Review all third-party enterprise applications with Microsoft 365 delegated permissions, revoke stale grants, hunt for new service principals/role assignments, rotate vendor-related credentials, and do not assume vendor credential rotation alone removes persistence.
Bullhorn takes staffing platform offline after cyberattack.
Bullhorn disabled access to its staffing and recruitment platform after a cyber incident caused a system-wide outage; the company serves more than 10,000 customers and had not yet determined whether customer data was exfiltrated.
Portfolio companies using sector-specific SaaS platforms should validate outage playbooks, data export options, SSO enforcement, API secret rotation, and contractual notification/SLA language for critical business systems.
Ahold Delhaize confirms malicious activity behind grocery e-commerce outages.
The retailer said malicious actors forced it to take systems offline across most U.S. and European brands, disrupting online services while stores remained open; it said it had no evidence of sensitive data theft but the investigation remains ongoing.
Retail and consumer portfolio companies should pressure-test segmentation between store operations, e-commerce, loyalty/customer data, and fulfillment workflows so containment does not require broad revenue-impacting shutdowns.
CISA flags active BeyondTrust Remote Support exploitation.
CISA warned that CVE-2026-102764, an authentication bypass in BeyondTrust Remote Support 26.2.3 and earlier, is being actively exploited for initial access, persistence, privilege escalation, and lateral movement; self-hosted appliances require urgent patching and forensic review.
Treat remote-support infrastructure as privileged access infrastructure: patch, restrict administrative exposure, isolate suspected appliances, forward logs to SIEM, hunt for unauthorized admin sessions and session-policy changes, rebuild from known-good state, and rotate all credentials stored or injected by the appliance.
Cyber Regulatory and Enforcement Changes
CISA’s KEV process accelerates remediation pressure on exploited products.
CISA added multiple exploited vulnerabilities in October, including BeyondTrust Remote Support CVE-2026-102764, Fortinet CVE-2025-61624, and Atlassian CVE-2026-21589; the catalog now ties remediation to BOD 26-04 risk-based timelines and, for several entries, forensic triage requirements.
Use KEV status as a portfolio-wide escalation trigger: identify internet-exposed instances, verify compensating controls, require evidence of patching or isolation, and document forensic triage where CISA flags it.
HHS/OCR settles Ambry Genetics phishing investigation for $700,000.
HHS/OCR’s resolution agreement states that a January 2020 targeted phishing attack may have involved PHI of 225,370 individuals and alleged failures around risk analysis, termination of ePHI access, and unique user identification/tracking.
Healthcare and life-sciences diligence should verify phishing-resistant MFA, joiner/mover/leaver access termination, unique user IDs, auditability, and documented risk analysis—not just security tooling.
FTC sues Digital Forensics over alleged sextortion relief deception.
The FTC alleges Digital Forensics Corporation and affiliates misled sextortion victims into paying at least $60 million for promised cyber assistance services, including allegedly fake investigative work, suppressed negative reviews, and fake positive reviews.
Cyber incident-response, forensic, recovery, and crisis-services companies should review marketing claims, refund policies, testimonial practices, and consumer-facing scripts; PE owners should treat deceptive cyber-services claims as an enforcement and reputational risk.
SEC schedules roundtable on machine-understandable data and “Super Intelligence.”
The SEC will host a November 9 roundtable on machine-understandable data, SI use, related risks, and costs, signaling that AI-readability of disclosures and market data is becoming a regulatory topic.
Public-company and sponsor-backed compliance teams should monitor how AI-readable reporting, XBRL/data quality, and AI-driven investor analysis may affect disclosure controls and investor-relations workflows.
Threat Intelligence and Adversary Activity
ARMORY CASTLE exploited BeyondTrust as a zero-day.
Mandiant reports that ARMORY CASTLE, a suspected PRC-nexus espionage actor formerly tracked as UNC5221, exploited CVE-2026-102764 in late August 2026 and used SAWFLY webshell and BARECLAW tunneling malware while deleting logs and artifacts.
Beyond patching, hunt for appliance-specific persistence, missing logs, unusual proprietary CLI usage, unexpected reverse-proxy routing, and downstream credential abuse from remote-support access.
Ransomware operators target PAN-OS CVE-2026-0257.
Unit 42 updated its threat brief to report attempted exploitation of PAN-OS CVE-2026-0257 by ransomware operators including Spikey Scorpius/Qilin and Tarnished Scorpius/Inc; the flaw can allow unauthorized attackers to circumvent controls and initiate GlobalProtect VPN connections.
Check PAN-OS exposure, validate fixed versions and mitigations, hunt for gateway-connected events tied to known indicators, and treat successful VPN sessions as incident-response events even if lateral movement is not yet visible.
Microsoft details Chinese-language phishing-kit ecosystem at global scale.
Microsoft says Storm-2990 phishing-kit operators and the Storm-3119/YAM Store marketplace helped facilitate more than one billion phishing emails and SMS messages to more than 75,000 organizations worldwide since January 2026, targeting sectors including healthcare, financial services, wholesale distribution, and IT.
Defenders should tune for phishing-kit infrastructure, brand impersonation, SMS/email credential theft, and cloud-credential resale patterns; identity controls and conditional access remain the most important downstream containment layer.
China-linked actors combine automated scanning, botnets, and hands-on exploitation.
A joint CISA advisory says China government-linked actors enabled by Integrity Technology Group target sensitive data worldwide, including U.S. critical infrastructure, using scanning tools, botnets, password spraying, VPN persistence, and exploitation of older CVEs.
Portfolio companies should not leave older internet-facing software as “accepted risk”; reduce exposed services, enforce MFA, sanitize web inputs, patch legacy vulnerabilities, and monitor VPN and Exchange authentication anomalies.
AI News, Security, and Governance
Microsoft frames AI agents as “newest insiders.”
Microsoft argued that agents can be manipulated, compromised, or drift from user intent while operating at machine speed, and announced controls such as agent discovery, distinct agent identity, dynamic permissions, dynamic agent risk, and conditional access for agents.
Enterprise AI programs should assign first-class identities to agents, prevent inheritance of full user permissions, constrain tool/API/MCP access, require human approval for consequential actions, and monitor agent behavior continuously.
OpenAI disrupts AI-enabled false-front influence operations.
OpenAI banned Russia- and Iran-origin operations that used models alongside traditional methods to support false-front entities, with the Russian operation assessed as Category 5 on the IO Breakout Scale and the Iranian operation as Category 4.
AI governance should include influence-operation risk, provenance review, editorial due diligence, and employee training on fake personas and front organizations—not only model safety and data leakage.
Anthropic updates usage policy for deception, surveillance, weapons, and high-risk AI.
Anthropic’s updated policy, effective November 12, consolidates deceptive-campaign restrictions, clarifies election and surveillance prohibitions, covers weapons-related software/components, and reiterates human-in-the-loop requirements for health, legal, finance, livelihood, and essential-service decisions.
Companies building on Claude or similar models should map use cases to prohibited/dependent controls, add human review and user disclosure for high-risk decisions, and validate restrictions for surveillance, election, and physical-action workflows.
OpenCode breach exposes API keys amid AI distillation concerns.
BleepingComputer reported that the open-source AI coding project OpenCode notified a small number of users that API keys used with external providers were exposed, reinforcing that AI developer tools can concentrate secrets across multiple model providers.
Treat AI coding tools as sensitive development infrastructure: inventory stored tokens, require scoped/rotated API keys, monitor provider usage anomalies, and separate experimentation keys from production credentials.
Private Equity News
EQT agrees to acquire Cleanaway Waste Management in A$9.3 billion transaction.
Private Equity Wire, citing Reuters, reported EQT agreed to acquire Australia’s Cleanaway Waste Management for approximately A$9.3 billion including debt, with Cleanaway shares surging almost 20% after the announcement.
Infrastructure investors remain active in essential-services assets with durable demand; diligence should focus on regulated operations, labor, environmental obligations, fleet/route optimization, and technology-enabled margin expansion.
KKR to sell Infinity Transportation to Ares for $2.9 billion.
PitchBook reported that KKR agreed to sell railcar lessor Infinity Transportation to Ares in a deal valued at about $2.9 billion; Infinity owns and manages about 50,000 railcars leased to more than 350 North American customers.
Sponsor-to-sponsor exits are still possible for scaled, contracted asset platforms; underwrite residual values, maintenance capex, customer concentration, insurance-capital linkages, and macro sensitivity to diesel, rail volumes, and trucking competition.
Carlyle to acquire $7.7 billion Truist personal-loan portfolio.
Carlyle agreed to acquire approximately $7.7 billion of unsecured personal loans from Truist Financial, expanding its asset-backed finance strategy as banks continue using private capital to manage balance sheets.
Private credit teams should expect continued bank partnership flow, but diligence must stress borrower credit quality, seasoning, rate exposure, servicing controls, consumer-compliance obligations, and loss curves under recession scenarios.
Malwarewolves Watchlist / Suggested Follow-Ups
Run a portfolio flash check for BeyondTrust Remote Support, PAN-OS/GlobalProtect exposure, Fortinet products, Atlassian Data Center products, and ConnectWise Cloud Backup delegated Microsoft 365 permissions.
Publish a short operator note on “AI agents as insiders”: identity, least privilege, agent registry, MCP/tool boundaries, human approval, and runtime monitoring.
Add a diligence module for SaaS business interruption risk: status-page dependency review, data-export availability, SSO/API secret rotation, vendor app permissions, and contractual incident notification.
Monitor FTC/HHS/OCR enforcement for cyber-services claims, phishing controls, and healthcare access-management failures; these are becoming practical diligence questions, not just legal footnotes.
