Since early May 2026, a large-scale phishing campaign has targeted individuals with fraudulent text messages claiming their T-Mobile rewards points are about to expire. The messages create urgency by stating that point balances, typically cited as 18,400 points, will expire imminently if not redeemed. Recipients are directed to click on phishing links using rotating domains designed to mimic legitimate T-Mobile websites. The campaign has generated over 1,000 closely related message templates with only superficial variations in salutations, dates, and point balances. The operation experienced two major spikes in activity before declining. The criminals employed at least 81 domains over four months, all following a recognizable pattern. These messages use generic greetings and formal language to appear legitimate, exploiting social engineering tactics to trick recipients into divulging login credentials, personal information, or payment details.
BragJack is a security research discovery demonstrating how a single browser extension can hijack internal browser agents across five Chromium-based browsers: Chrome, Comet, Edge, Opera Neon, and Claude in Chrome. The attack leverages the declarativeNetRequest (DNR) API to weaken security headers and redirect JavaScript resources, enabling code execution within privileged browser contexts. This allows access to sensitive capabilities such as local file access, browsing history, camera and microphone, and screenshots without user interaction. The research resulted in two CVEs and $20,000 in bounty rewards from major vendors. The vulnerabilities primarily exploit weaknesses in how these browsers handle network requests and security policies within embedded web views. No patch or remediation status is provided in the source data.
A Chinese-speaking threat actor group known as Red Heron exploited a recently disclosed remote code execution vulnerability (CVE-2026-60004) in Gitea, a self-hosted Git service, in a multinational campaign. The campaign targeted internet-facing Gitea instances across multiple countries, including Canada, Argentina, Taiwan, the United States, and Sri Lanka, focusing on sectors such as defense, elections, energy, aerospace, telecommunications, government, and research. The attackers used automated tools to steal source code, credentials, and maintain persistent access, including root-level control on some infrastructure. They deployed a novel Linux implant named JITTERLY with extensive post-exploitation capabilities and embedded a previously undocumented rootkit called SIXZUT to maintain stealth and persistence. The campaign demonstrates rapid weaponization of n-day vulnerabilities in development platforms and highlights significant risks to source code confidentiality and infrastructure integrity.
Cybercriminals are using fake websites that mimic the legitimate Bitrefill site to defraud users. These lookalike sites appear in search engine results and use domains similar to Bitrefill's, including Punycode variants. Victims are tricked into completing checkout processes that resemble Bitrefill's, paying cryptocurrency to attacker-controlled addresses with no recovery option. The campaign uses sophisticated analytics to optimize victim conversion, indicating organized criminal activity.
A campaign named Miasma exploited a vulnerability in the Model Context Protocol (MCP) tool configuration process to gain OS-level access to 73 GitHub repositories, including one owned by Microsoft Azure. The attack leverages the fact that MCP clients trust tool descriptions as executable instructions without re-prompting or sandboxing, allowing malicious edits to MCP config files to run with developer privileges. This attack class, known as MCP tool poisoning, was first documented in 2025 and escalated to real-world exploitation by mid-2026. The campaign abused the approval process tied to server names rather than content, enabling malicious commands to execute automatically after initial approval. OWASP ranks this attack third in its MCP Top 10 threats. Mitigations include rigorous code review of MCP config changes, pinning tool definitions, scoping privileges per tool, maintaining allowlists of MCP servers, and sandboxing agent execution.
A malicious browser extension named 'Twitch Enhanced Viewer | JeetBot' distributed via Chrome Web Store and Firefox Add-ons steals Twitch OAuth tokens from approximately 30,000 Chrome users and 552 Firefox users. The extension masquerades as a tool to improve user experience by blocking ads and unlocking streams but secretly captures and exfiltrates OAuth session tokens to Russian-controlled proxy servers. The stolen tokens are sent during video playlist redirects or via POST requests to collection endpoints. The operator is linked to a commercial Russian bot service called JeetBot, with infrastructure hosted in Germany and cloud providers.
Threat actors exploited trusted brands and cloud services in a sophisticated web campaign combining fraudulent DocuSign workflows, Florida healthcare screening lures, and deceptive cloud infrastructure to deploy ConnectWise ScreenConnect Access clients. The attack utilized Cloudflare Pages hosting with fake Cloudflare verification workflows to establish legitimacy. Victims were socially engineered to download a ZIP archive containing a malicious HTA file that employed Base64-encoded VBScript, fake Adobe interfaces, UAC privilege escalation, and Microsoft Defender SmartScreen registry modifications. The attack leveraged living-off-the-land techniques using native Windows tools like mshta.exe, curl.exe, and msiexec.exe for silent ScreenConnect installation, ultimately providing unauthorized remote access. The campaign was classified as Zero Hour Fraudulent and blocked at the web entry point before payload delivery could occur.
Threat actors are leveraging generative AI to enhance financial fraud campaigns targeting enterprise organizations. Between August 3-5, over one million phishing emails were distributed through third-party infrastructure, primarily targeting US-based organizations (87.7%). The attacks employed sophisticated executive impersonation, specifically CEOs and CFOs, combined with fabricated ServiceNow invoices requesting ACH transfers of approximately $50,000. The campaign demonstrated multiple AI-assisted indicators including extensive HTML comments, structured section labeling, and uniform template construction. Attackers registered lookalike domains and created elaborate forwarded email threads between spoofed executives to establish legitimacy. The fraudulent invoices contained detailed branding, personalized recipient information, and specific payment instructions to attacker-controlled bank accounts. Multiple layered social engineering techniques were deployed to reduce recipient skepticism and convince acc...
A threat actor exploited CVE-2026-15409, a critical unauthenticated server-side request forgery vulnerability in SonicWall SMA1000 appliances, to gain command execution and steal credentials. The attacker used a modified public proof-of-concept exploit to access internal Erlang services on the appliance, enabling remote code execution. This allowed extraction of LDAP configurations, Active Directory credentials, and deployment of tools to dump secrets from internal Windows systems. The campaign targeted at least 250 SonicWall SMA1000 devices across multiple countries and sectors, with confirmed credential theft in France, India, Italy, and the US. The attack leveraged compromised appliances as pivots into internal networks, exposing sensitive Active Directory data and enabling DCSync attacks against domain controllers. The targeting was opportunistic and technology-driven rather than sector-specific. The campaign was uncovered through an open directory left exposed by the attacker, providing a comprehensive view of the operation.
A widespread IT impersonation and voice-phishing campaign designated PREY-0058 is actively targeting Microsoft 365 and SaaS platforms. Threat actors impersonate IT helpdesk staff via phone or text to trick executives and senior personnel into divulging credentials through adversary-in-the-middle phishing portals. After bypassing multi-factor authentication, attackers conduct rapid automated data exfiltration across email, file storage, and cloud repositories without deploying ransomware. The operation relies heavily on NodeMaven residential proxy infrastructure to blend with legitimate traffic. Extortion demands are delivered via TOX messaging within hours of compromise, typically with 72-hour deadlines and threats of public data exposure. The campaign exhibits tradecraft overlapping with UNC6671 and involves multiple extortionware brands including BlackFile, Redact, Pink, and Helix.
A large-scale distributed denial-of-service (DDoS) campaign targeted AI Weekly following their investigative reporting on Meta's advertising partner GatherOne and associated nudify apps with child-safety concerns. The attack generated roughly four billion HTTP requests, primarily targeting the specific investigative article. The campaign caused significant server resource exhaustion, including maxing out Nginx worker connections, filling disk storage, and causing numerous server errors and job failures. The attack persisted in waves, shifting focus from the article to the homepage when mitigated, and returning weeks later. There is no evidence implicating Meta, GatherOne, or related entities in orchestrating the attacks. The campaign weaponized infrastructure to disrupt reporting rather than engage in direct rebuttal or criticism.
Multiple cloud accounts have been compromised through passkey-themed social engineering campaigns since May 2026. Attackers impersonate IT helpdesk staff via phone calls or SMS, directing victims to phishing sites that mimic Microsoft sign-in pages. After obtaining credentials through adversary-in-the-middle or device code authentication flows, attackers establish persistence by adding unauthorized MFA methods. They then conduct extensive reconnaissance using Microsoft Graph API to map users, groups, permissions, and resources. The intrusion culminates in high-volume data collection from SharePoint, OneDrive, and Exchange using automated tools, with exfiltration occurring over hours or days at controlled rates to avoid detection.
Cisco Talos is tracking a cryptocurrency theft campaign that abuses the Google Visualization API to deliver obfuscated JavaScript from Google Sheets, which victims are socially engineered to inject into their browser sessions. The attack uses a variation of ClickFix social engineering, convincing targets to paste malicious JavaScript into the Chrome address bar or install it via the Tampermonkey extension for persistence. The malicious script acts as a web skimmer, hooking the browser's fetch API to replace cryptocurrency deposit addresses and manipulate clipboard contents, deceiving users with fake bonus interfaces. The campaign targets users on cryptocurrency and hacking forums with a lure about a nonexistent API vulnerability to entice quick financial gain. The actors moved their hosting to Google Docs and Sheets after disruptions on paste sites. While this campaign primarily targets individuals, the techniques used could be adapted for broader supply-chain or web-based attacks.
BigBear 2.0 is a phishing-as-a-service campaign leveraging Evilginx2 to target Microsoft 365 credentials worldwide. It uses adversary-in-the-middle techniques combined with geo-matched residential proxies to bypass multi-factor authentication, including disabling FIDO2/WebAuthn MFA. The campaign has exfiltrated thousands of credentials and session cookies, impacting over 3,300 unique victim IPs across more than 40 countries. The platform is operated by an entity known as 'General Boss' and leased to multiple affiliates. It automates cookie replay and real-time data exfiltration via Telegram.
North Korean-aligned threat actors have deployed a sophisticated Linux espionage toolkit targeting automotive and media organizations in South Korea. The toolkit embeds a custom backdoor within HAProxy version 2.8.12, enabling stealthy long-term surveillance through HTTP traffic interception and injection. It also includes trojanized Linux tools and a curl-based RAT for remote command execution, credential harvesting, and persistence. Initial access was gained via exploitation of a Groupware login portal vulnerability. The campaign uses watering-hole techniques and mimics legitimate web traffic to evade detection. The activity is attributed to North Korean threat actors, likely linked to Lazarus or APT37 groups. No patch or remediation guidance is provided in the source.
The Kimsuky threat group continues Operation GitPower campaigns utilizing malicious LNK files disguised as financial and business documents. Thirteen LNK variants collected between August 11-19, 2026, employ GitHub Personal Access Tokens (PAT) for command-and-control communications, delivering obfuscated PowerShell loaders through custom decoders. Notable evolution includes anti-analysis routines detecting virtualization tools, Pastebin as alternative C2 infrastructure, and diversified decoy formats (PDF, XLSX, PNG). Metadata analysis reveals AI-generated content using the 'opencode' AI coding agent and HeadlessChrome PDF conversion, with placeholder text remaining unreviewed. The group maintains persistence through hidden scheduled tasks masquerading as legitimate software (BitLocker, MATLAB), while hardcoded GitHub PATs enable raw content retrieval. Despite increased sophistication in evasion techniques and decoy production automation, endpoint behaviors remain detectable through behavioral correlation a...
A large-scale phishing campaign used invisible Unicode characters from the deprecated Unicode Tags block to obfuscate keywords related to loans and financing, bypassing email filters that rely on contiguous ASCII string matching. The campaign sent between 1 million and 2.37 million messages daily at its peak in early 2026, primarily targeting recipients with fraudulent loan offers. The messages were routed through the legitimate marketing platform ActiveCampaign, leveraging its reputation to evade detection. There is no software vulnerability or patch associated with this technique; it exploits differences in how email clients and filters process Unicode text. Mitigation requires enhanced email filtering that inspects raw Unicode content and flags or removes invisible characters to detect obfuscated phishing keywords.
Attack campaigns targeting Korean users deploy remote control tools Radmin and UltraVNC to compromise systems. Attackers download compressed files containing batch scripts and remote administration software. After installing Radmin, they deploy UltraVNC along with proxy tools such as Netch-gateway and CCProxy to use compromised systems as proxy nodes. Recent variants also deploy SoftEther VPN to establish VPN servers on infected infrastructure. PowerShell scripts with Chinese language comments and tools familiar to Chinese-speaking actors suggest attribution to Chinese threat operators. These campaigns enable remote system control and abuse of compromised infrastructure for proxy services.
Microsoft researchers identified a high-volume phishing campaign utilizing invisible Unicode tag characters (U+E0000 to U+E007F), a technique originally associated with AI prompt injection research known as ASCII Smuggling. The attackers inserted these invisible characters into financial keywords like 'funding' to evade email filters rather than hiding instructions from users. The campaign began February 9, 2026, generating millions of daily messages for approximately three months with a distinctive weekday-only pattern. Finance-themed disposable domains were used to send business loan and credit-line phishing through a legitimate email marketing platform. The technique, while designed for AI security contexts, proved effective at bypassing traditional keyword-based detection by splitting words with invisible characters that appear normal to recipients but break signature matches and alter ML tokenization.
Two multi-stage network intrusion campaigns are actively targeting organizations in Latin America, specifically in Mexico, Ecuador, and Brazil. The Mexican campaign targets transportation and government entities using living-off-the-land techniques and self-hosted NextChat instances. The Brazilian campaign targets financial organizations with custom remote access trojans (RATs) and tunneling tools, including a Go-based SOCKS5 proxy. Both campaigns leverage AI tools such as Claude and GPT-4.1 to generate scripts and troubleshoot operations. Operational security failures by attackers, including exposed staging directories and unsecured interfaces, have allowed defenders visibility into their activities. Indicators include multiple duckdns.org domains, an IP address, and various file hashes. There are no known exploits in the wild or confirmed active threat actors at this time.
