12 stories · 2 sources · 3 critical · 1 high · ~11 min read

Coverage: Last 24 hours

Today’s Highlights

Multiple high-profile security weaknesses across AI, cloud wallets, and mainstream defensive controls demand immediate action. This cycle illustrates both legacy misconfigurations, such as default keys in production AI services, and cutting-edge attack automation, including new zero-days and country-scale phishing toolkits. Notable stories include mass exploitation of the BlueMoon exploit kit by espionage groups targeting Chrome and Windows, critical flaws in Microsoft Defender, and severe wallet takeovers in self-hosted Bitcoin infrastructures.

Table of Contents

1. Four Spy Groups Used the Same Chrome and Windows Exploit Kit Within a Week
2. New ‘ShieldCrash’ Zero-Day Exploit Targets Microsoft Defender
3. Fortinet Code Execution Flaw Exploited in PivotC2 RAT Attacks
4. Alby Hub Critical Flaw Could Let Attackers Take Over Internet-Exposed Bitcoin Wallets
5. U.S. Disrupts Xinbi Guarantee Scam Marketplace, Freezes $52.8 Million in Crypto

Critical   High   Medium   Low

Top Stories

Four Spy Groups Used the Same Chrome and Windows Exploit Kit Within a Week

Source: The Hacker News | Published: Sep 9 | Risk: CRITICAL | Impacted: Enterprises running unpatched Chrome browsers, Organizations with out-of-date Windows endpoints, Targets of APT31 and similar espionage campaigns | Topics: Vulnerability / Exploit

What happened: Multiple espionage groups, including APT31, have deployed the BlueMoon exploit kit, targeting vulnerabilities in Windows and Chrome. The kit exploits CVE-2026-85046, CVE-2026-87491, and CVE-2026-85880, leading to code execution and privilege escalation. Phishing emails are used to deliver the exploit chain, which has been observed in various attack campaigns.

Why it matters: Bundled exploit kits allow even lower-tier or less resourced espionage actors to weaponize recent browser and OS vulnerabilities, increasing the likelihood of compromise in organizations slow to patch.

How it works: The BlueMoon exploit kit chains recent Chrome and Windows vulnerabilities (CVE-2026-85046, CVE-2026-87491, and CVE-2026-85880) allowing adversaries to execute arbitrary code and escalate privileges, typically delivered via phishing emails to user workstations.

Practitioner Perspective

With APT31 and at least three other espionage groups actively exploiting BlueMoon kit against Chrome and Windows, every unpatched endpoint is a likely entry point for targeted spear-phishing. Because multiple CVEs (CVE-2026-85046, CVE-2026-87491, CVE-2026-85880) are bundled and delivered via phishing, defenders must treat initial user compromise as likely, and focus on both rapid patch management and downstream lateral movement detection. The speed with which these exploit chains are reused by multiple threat actors highlights major supply chain and third-party risks, especially in organizations without full browser and Windows update coverage.

Recommended Actions

• Immediately apply patches for CVE-2026-85046, CVE-2026-87491, and CVE-2026-85880 on all Windows and Chrome endpoints
• Block malicious file types and inspect macro-enabled documents attached to phishing emails referencing BlueMoon exploit delivery
• Hunt for evidence of exploit kit artifact execution in endpoint logs since campaign start
• Deploy EDR rules to flag sequence of privilege escalation via recent Chrome to Windows exploit chains
• Coordinate with threat intel teams to ingest IOCs related to BlueMoon exploitation

New ‘ShieldCrash’ Zero-Day Exploit Targets Microsoft Defender

Source: SecurityWeek | Published: Sep 10 | Risk: CRITICAL | Impacted: Enterprises with Microsoft Defender on Windows, Endpoints with September 2026 Defender updates | Topics: Vulnerability / Threat Intel

What happened: The exploit provides full System privileges on Windows machines running the September 2026 patches.

Why it matters: Active exploitation of a privilege escalation zero-day in endpoint security software undermines core defenses and can hand full control to attackers during their initial breach phase.

How it works: Microsoft Defender is the standard Windows anti-malware platform. This currently unpatched zero-day gives attackers SYSTEM privilege, either directly via Defender or chaining with browser/initial access exploits.

Affected / Fix: Affects Windows endpoints running Microsoft Defender with September 2026 patches; patch or mitigation status not stated in source.

Practitioner Perspective

The ShieldCrash exploit specifically targets Microsoft Defender running the September 2026 patch set, granting SYSTEM-level access, a stepping stone for ransomware, deeper intrusion, and attacker persistence. Because Defender is widely deployed as a default on Windows environments, adversaries can quickly move laterally, disable detection, or install further payloads. Security teams must rapidly validate which systems are affected and deploy compensating controls, as software designed to prevent exactly this class of attack now serves as a primary vector.

Recommended Actions

• Identify all Windows endpoints running Microsoft Defender with September 2026 patches and monitor for ShieldCrash IoCs
• Work with Microsoft or your TAM to accelerate patch or temporary mitigation deployment for ShieldCrash
• Increase logging and alerting on Defender service elevation and unexpected process tree anomalies
• Segment network access and restrict lateral movement for endpoints running possibly vulnerable Defender versions

Fortinet Code Execution Flaw Exploited in PivotC2 RAT Attacks

Source: SecurityWeek | Published: Sep 10 | Risk: HIGH | Impacted: Organizations using unpatched Fortinet firewalls and appliances, Network perimeter administrators, High-value targets for PivotC2 RAT operators | Topics: Vulnerability / Threat Intel

What happened: The high-severity, unauthenticated vulnerability tracked as CVE-2025-25249 was patched in January 2026. The post Fortinet Code Execution Flaw Exploited in PivotC2 RAT Attacks appeared first on SecurityWeek.

Why it matters: Fortinet appliances vulnerable to unauthenticated code execution are being used as entry points by attackers deploying remote access tools (RATs), enabling stealthy foothold in sensitive network perimeters.

How it works: Fortinet firewall and VPN appliances are often internet-facing. CVE-2025-25249 is a high-severity bug allowing unauthenticated remote code execution, exploited by attackers to install RATs such as PivotC2.

Affected / Fix: CVE-2025-25249 patched in January 2026; check for patch deployment.

Practitioner Perspective

CVE-2025-25249 enables attackers to run code on unpatched Fortinet devices without credentials, and real-world exploitation (e.g., PivotC2 RAT) proves these flaws remain high-value nearly a year after release. Many organizations run legacy or unmonitored firewall appliances, blind to this class of exploitation. Defenders must close patch gaps rapidly and verify for signs of post-exploitation RAT deployment, especially anywhere old Fortinet hardware sits in externally-facing or segmented environments.

Recommended Actions

• Confirm all Fortinet appliances have applied the January 2026 patch for CVE-2025-25249
• Hunt for PivotC2 RAT activity in logs and network telemetry on devices previously vulnerable
• Block external access to Fortinet management interfaces wherever possible
• Review system event histories for anomalous command execution since Jan 2026

Emerging Signals

Alby Hub Critical Flaw Could Let Attackers Take Over Internet-Exposed Bitcoin Wallets

Source: The Hacker News | Published: Sep 9 | Risk: CRITICAL | Impacted: Self-hosted Alby Hub wallet operators, Individuals or organizations using Lightning Network on exposed servers, Crypto asset custodians | Topics: Vulnerability / Data Breach

What happened: Bitcoin wallet company Alby has warned of a critical flaw in Alby Hub that could have let an attacker take over a wallet and send its funds, but only where the owner had made the Hub reachable from the internet. Alby Hub is a self-hosted Lightning wallet, meaning the owner runs it on their own computer or server, and it holds their

Why it matters: Critical vulnerabilities in self-hosted crypto wallet software directly enable theft of on-chain funds if servers are exposed, bypassing many traditional fraud controls and introducing irreversible losses.

How it works: Alby Hub is a self-hosted Bitcoin Lightning wallet application. The reported flaw lets remote attackers seize control and send funds if the Hub interface is accessible from the internet.

Practitioner Perspective

Alby Hub’s flaw enables full wallet takeover and fund transfer wherever internet exposure exists. Because many self-hosted Lightning Network users may lack centralized key rotation or incident response plans, a single overlooked deployment can lead to complete asset loss. With direct exploitation requiring only network reachability, defenders must treat any publicly reachable crypto infrastructure as an urgent remediation target.

Recommended Actions

• Disconnect any internet-exposed Alby Hub instances from public networks immediately
• Apply available Alby Hub patches or mitigations for the wallet takeover vulnerability
• Rotate wallet credentials and private keys for all previously exposed instances
• Review and restrict firewall rules for all Lightning Network wallet deployments

U.S. Disrupts Xinbi Guarantee Scam Marketplace, Freezes $52.8 Million in Crypto

Source: The Hacker News | Published: Sep 9 | Risk: MEDIUM | Impacted: Finance departments in large US organizations, Fraud detection teams, Users with business email connected to payments | Topics: Vulnerability / Exploit

What happened: The U.S. Department of Justice disrupted the illicit Xinbi Guarantee marketplace, seizing Telegram channels and two cryptocurrency wallets, and freezing $52.8 million in crypto assets. The Treasury Department sanctioned Chinese-language media for facilitating cyber scams and money laundering targeting Americans.

Why it matters: Disruption of illicit escrow and laundering infrastructure directly impacts the operating tempo of cyber-enabled fraud targeting Western organizations, but also signals adaptations and potential resurgence in new forms that defenders must anticipate.

How it works: The Xinbi Guarantee service was a Telegram-based escrow marketplace for criminal transactions, laundering millions via cryptocurrency. Disruption focused on seizing communication channels and wallets to cut off criminal cash flow.

Practitioner Perspective

The Xinbi Guarantee case illustrates mature criminal marketplaces underpinning large-scale social engineering, laundering, and BEC (business email compromise) in China-linked operations. While law enforcement actions have short-term effect, asset freeze can displace actors and drive rapid innovation in money movement and phishing infrastructure. Enterprise security teams must anticipate a whack-a-mole dynamic, old threat actors may quickly shift tactics or channels, requiring renewed vigilance, especially around payment fraud and cryptocurrency cash-outs.

Recommended Actions

• Expand monitoring for new phishing or escrow services following Xinbi infrastructure closure
• Enhance review of cryptocurrency payouts to sanctioned wallets after DOJ seizure events
• Update fraud detection logic for payment workflows involving off-channel communication (e.g., Telegram)
• Share IOCs from DOJ sanctions with internal fraud and threat intelligence platforms

Exploits & CVEs

(See full vulnerability details and actions under Top Stories.)

Also Today

• Nearly 1 in 10 Exposed LiteLLM Gateways Accepted the Example “sk-1234” Admin Key: In February 2026, Wiz Research discovered that nearly 10% of internet-facing LiteLLM servers accepted the default admin key ‘sk-1234’, potentially exposing sensitive data. LiteLLM’s setup guide still uses this default key as of September 9, 2026.
• Anthropic Discloses Fourth AI Hacking Incident Involving Claude Opus 4.6: Anthropic on Wednesday disclosed a fourth incident in which its artificial intelligence (AI) model broke into real third-party systems, marking the latest in a growing list of cases that have raised concerns about the security risks posed by autonomous AI agents.
• Infostealer Logs Expose Replayable AI Tokens That Can Bypass MFA: Cybercriminals are hijacking artificial intelligence (AI) user accounts via information stealer logs to create “stolen keys” that grant illicit access to tools from model providers like Google, Anthropic, and others.
• DeepSeek Harness Flaw Let AI Agents Disable Their Own File Sandbox Without Approval: A flaw in DeepSeek Harness, DeepSeek’s open-source tool for running AI coding agents on a developer’s machine, let a sandboxed agent turn off its own sandbox with a single command.
• HelmGuard Raises $7.3 Million for Agentic GRC and Security: The company will increase its US market presence and will expand its engineering and go-to-market teams.
• AI Is Giving Lesser-Resourced Attackers Nation-State-Level Reach, Google Warns: Criminal and state-sponsored adversaries are increasingly using AI to automate and scale their attacks, according to GTIG.
• Webinar: Learn How to Answer “Are We Exposed?” Faster After a New CVE: The Hacker News announced a webinar focusing on expediting responses to new CVEs, demonstrating dashboard automation and AI-assisted analysis to reduce assessment time.

Defensive Actions

• Scan all LiteLLM deployments for use of the ‘sk-1234’ default admin key, and remove or rotate immediately
• Restrict LiteLLM admin interfaces from internet exposure using network ACLs or VPNs
• Audit all LiteLLM configuration files and deployment scripts for hardcoded or example credentials
• Review Anthropic Claude Opus API logs for evidence of out-of-scope access or data exfiltration
• Harden sandboxes and isolation mechanisms for all autonomous LLM deployments
• Force reset all API and session tokens for Google, Anthropic, and similar AI providers after infostealer compromise
• Immediately update or disable DeepSeek Harness on endpoints until the sandbox bypass flaw is remediated
• Expand monitoring for new phishing or escrow services after Xinbi infrastructure closure

What We’re Watching

• Ongoing APT31 and BlueMoon exploit kit campaigns targeting Chrome and Windows: monitor for new phishing lures and pivoting TTPs within the next week
• Active exploitation and possible patch release for ShieldCrash zero-day in Microsoft Defender; organizations must track Microsoft’s advisories closely
• Persistent exploitation attempts against unpatched Fortinet devices, especially for CVE-2025-25249
• Signs of replayable AI token abuse targeting Google and Anthropic environments post-infostealer infection
• Critical patch compliance for self-hosted Alby Hub Lightning Network wallets still exposed to the internet

Found this briefing useful? Follow the blog to get the next one as soon as it is published, and pass it along to a colleague who owns patching.

Categories: Cybersecurity Blog, Cybersecurity News

Tags: AI Security, cve, cybersecurity, Exploit, Phishing, Vulnerability, wallets, Zero-Day