Threat Intelligence

Two former members of the United States Air Force have been sentenced to a combined 189 months in federal prison for their roles in a multi-year series of business email compromise (BEC) scams and phishing campaigns. The defendants were involved in schemes that targeted organizations over an extended period, using BEC and phishing techniques to facilitate their attacks. The court handed down the combined sentence as part of the federal case against them. Source: BleepingComputer Does your organization have controls in place to detect and block BEC attempts before they reach employees' inboxes?

Unknown threat actors have compromised two legitimate MemTensor packages across the npm and PyPI repositories, using them to distribute a platform-specific, Go-based implant tracked as sckit. According to reports from Aikido, SafeDep, Socket, and StepSecurity, the affected libraries are the @memtensor/memos-cloud-openclaw-plugin versions on npm and a related package on PyPI. The sckit implant is designed to operate across Windows, Linux, and macOS, suggesting the campaign targets a broad range of developer environments rather than a single platform. Key points from the reporting: Two legitimate MemTensor packages were compromised, one on npm and one on PyPI The affected npm package is @memtensor/memos-cloud-openclaw-plugin The payload is a Go-based implant named sckit sckit supports Windows, Linux, and macOS The activity was documented by Aikido, SafeDep, Socket, and StepSecurity If you maintain projects that depend on these packages, review your lockfiles and dependency trees for any usage, and check build and CI environments for signs of unexpected outbound activity. Source: The Hacker News Have you audited your dependency manifests for these compromised MemTensor packages yet?

Researchers at Proofpoint have disclosed an active TeamFiltration campaign, tracked as UNK_CondorFiltration, that has targeted more than 5,700 accounts across 28 Microsoft 365 tenants. The activity has primarily focused on Chilean retail and financial institutions and originated from 1,487 unique AWS EC2 source IP addresses. According to the findings, the operation successfully compromised 7 accounts, with the attackers gaining access by abusing default passwords. TeamFiltration is a known credential-harvesting and account-enumeration tool, and its use here highlights how weak or unchanged default credentials continue to give threat actors an easy entry point into cloud environments. Key details from the disclosure: Over 5,700 accounts targeted 28 Microsoft 365 tenants affected 7 accounts successfully compromised Attack traffic originated from 1,487 unique AWS EC2 source IP addresses Primary targeting focused on Chilean retail and financial institutions Initial access achieved through default passwords For defenders, the takeaway is familiar but critical: audit for accounts still using default or vendor-supplied credentials, enforce strong password policies, and monitor for unusual authentication patterns originating from cloud hosting ranges such as AWS EC2. Source: The Hacker News Has your organization audited for default credentials across its Microsoft 365 tenant, and how are you detecting suspicious sign-ins from cloud provider IP ranges?

Kiteworks, formerly known as Accellion, has advised customers to take their systems offline for a precautionary nine-hour shutdown over the weekend. The guidance follows credible threat intelligence warning of a possible imminent cyber attack against the platform. According to Frank Balonis, the company's Chief, Kiteworks received intelligence from federal authorities indicating that a threat actor may attempt to target certain Kiteworks systems. The shutdown is positioned as a precautionary measure rather than a response to a confirmed breach. Key points for customers and defenders: A nine-hour system shutdown was recommended over the weekend. The warning stems from federal threat intelligence, not a confirmed compromise. The advisory applies to some Kiteworks systems, per the company's statement. The legacy Accellion brand is directly relevant here, given its history of high-profile exploitation. Anyone running Kiteworks or inherited Accellion infrastructure should treat the window seriously, verify their exposure, and monitor vendor communications for further instructions before restoring service. Source: The Hacker News Does your organization run Kiteworks or legacy Accellion systems, and are you planning to follow the shutdown guidance?

Two third-party GitHub Actions that were previously compromised as part of the Mini Shai-Hulud campaign were re-enabled by their maintainer and stayed accessible for over a week, even though they still pointed to malicious code. The re-enabling effectively restored a known-bad supply chain path rather than cleaning it first. Key points: The actions had already been compromised in the Mini Shai-Hulud campaign before being re-enabled. After being re-enabled by the maintainer, they remained reachable for more than a week. During that window, the actions still referenced malicious code. Because these are third-party actions, any workflow referencing them could pull in the malicious payload during execution. This is a reminder that re-enabling a previously compromised dependency without fully remediating it does not remove the risk. Teams pinning third-party actions should verify the specific commit or version in use, and confirm that the referenced code has actually been cleaned, not just made available again. Suggested actions for anyone using third-party actions: Audit workflows for references to the affected actions. Check whether your pinned refs point to commits tied to the compromise. Re-pin to a verified, clean commit or remove the dependency. Review recent workflow runs for unexpected behavior or secret exposure. Source: Unknown Has your organization audited its third-party GitHub Actions references since this campaign first surfaced?

The logistics sector is being targeted by a new Android spyware campaign distributing a threat tracked as Corp MDM. According to Have I Been Squatted, the operation relies on fake Google Play pages branded as CEVA and TKW Logistics to deliver a malicious Android Package Kit (APK). The APK is disguised as a system service and uses the package name com.corp.mdm. Once installed, Corp MDM is reported to steal newly received SMS messages and redirect calls, giving attackers a channel for interception and further abuse of the infected device. Distribution: fake Google Play pages impersonating CEVA and TKW Logistics Payload: Android APK masquerading as a system service Package name: com.corp.mdm Capabilities: theft of new SMS messages and call redirection Targeted sector: logistics Source: The Hacker News Has your organization seen any fake app store pages or suspicious APKs impersonating logistics providers?

Bitget has disclosed that suspected North Korean threat actors stole $351.6 million from its hot and warm wallets after a backend compromise. According to the exchange, its security systems flagged unauthorized transfers involving a limited number of hot wallets at 18:31 UTC on September 24, 2026. The company stated that its cold wallets and the overwhelming majority of platform assets remain unaffected. Key points from the disclosure: Suspected attribution: North Korean threat actors Total stolen: $351.6 million Affected wallets: a limited number of hot and warm wallets Detection time: 18:31 UTC on September 24, 2026 Unaffected: cold wallets and the overwhelming majority of platform assets The exchange has not released further technical details about how the backend was compromised or whether specific indicators of compromise are available. Source: The Hacker News Does your organization treat exchange hot wallet compromises as a supply chain risk, and what controls are you using to limit exposure to similar backend breaches?

The growing use of AI agents in enterprise environments is exposing a gap in how organizations approach SOC 2 compliance, according to analysis from Token Security. Unlike traditional automated scripts, AI agents can operate through human credentials and carry out actions that existing SOC 2 controls may be unable to distinguish from legitimate human activity. This creates a significant blind spot. SOC 2 was designed around a model where access and actions could be traced back to a human user or a clearly defined service account. AI agents blur that line. They authenticate as people, act autonomously, and may perform sensitive operations without triggering the kinds of alerts that auditors and security teams expect for human-driven behavior or conventional machine accounts. The core concern is that agent identities are not adequately covered by current control frameworks. If an agent inherits a human's permissions and acts on their behalf, traditional controls around least privilege, access reviews, and activity monitoring may not capture what is actually happening. This can leave organizations with compliance attestations that look clean on paper while real operational risk goes unaddressed. Token Security argues that SOC 2 needs to evolve to explicitly account for agent identities, including how they are provisioned, monitored, and governed. Without that adaptation, the framework risks becoming less relevant as autonomous agents take on more responsibility inside enterprise systems. Key points to consider: AI agents can act through human credentials, making their activity hard to separate from genuine user actions Existing SOC 2 controls may not identify or flag agent-driven behavior Traditional service accounts and human accounts do not map cleanly onto autonomous agent identities Governance gaps around provisioning, permissions, and monitoring of agents remain largely unaddressed Failure to adapt could reduce the practical value of SOC 2 attestations Source: Publication Name How is your organization handling identity and access governance for AI agents that operate under human credentials?

A financially motivated threat actor has been observed using open-source AI agent frameworks to conduct large-scale attacks against online retailers, resulting in the theft of more than 600,000 credit card records and the compromise of over 100 websites with payment skimmers. The operation relies on AI agents to automate and scale what would traditionally require significant manual effort, from reconnaissance through to the deployment of malicious scripts on victim sites. According to the reporting, the campaign targets e-commerce platforms at scale rather than focusing on individual merchants, allowing the actor to harvest card data from many victims simultaneously. Key points from the reporting: The threat actor is financially motivated and leverages open-source AI agent frameworks as part of their tooling. More than 600,000 credit card records were stolen. Over 100 sites were infected with skimmer code. The scale of the operation points to heavy automation across the attack lifecycle. The use of AI agent frameworks marks a notable shift in how skimming campaigns can be executed, lowering the effort required to identify and compromise large numbers of retail sites while maintaining a high volume of stolen payment data. Source: BleepingComputer Has your organization reviewed its payment page scripts for unauthorized modifications in light of automated skimming campaigns like this?

The Contagious Interview campaign has compromised at least 30,000 devices across more than 100 countries, according to a new joint cybersecurity advisory. The operation, attributed to North Korean threat actors, has also siphoned funds or account credentials from over 7,000 cryptocurrency wallets. The campaign primarily targets individual web designers, engineers, and cryptocurrency specialists. Key points from the advisory: At least 30,000 compromised devices Victims located in more than 100 countries Over 7,000 cryptocurrency wallets affected Estimated losses of $10.71 million in crypto Primary targets: web designers, engineers, and cryptocurrency specialists Given the broad geographic spread and the targeting of individuals rather than a single sector, security teams may want to review their exposure to this campaign and ensure endpoint and credential protections are in place. Source: The Hacker News Has your organization seen any signs of Contagious Interview activity, and what steps are you taking to mitigate it?

An attacker accessed CrowdSec's GitHub organization and copied roughly 170 private repositories on May 22, according to a disclosure the French security company published on September 18. The intrusion was carried out through the GitHub account of an employee who had recently left the company but whose access had not been revoked. The entry point was not GitHub itself. CrowdSec says the former employee's laptop was compromised during May's supply chain attack on TanStack, in which malicious versions of TanStack's npm packages were used to steal credentials. Those stolen credentials then allowed the attacker to authenticate as the departed employee and pull down private code. Key facts as disclosed: Approximately 170 private GitHub repositories were copied. The access was made via the account of a former employee whose GitHub permissions remained active after departure. The initial compromise is tied to the TanStack npm supply chain attack, where trojanized package versions harvested credentials. CrowdSec attributed the repository exfiltration to an attacker operating through that hijacked account. The incident highlights two recurring themes: the long tail of credential theft from npm supply chain compromises, and the risk of stale access left in place after staff departures. Organizations that consumed the affected TanStack package versions should treat any credentials exposed on developer machines as potentially compromised, and should audit GitHub organization membership for accounts belonging to former personnel. Source: The Hacker News Has your organization reviewed GitHub access for former employees and rotated credentials that may have been exposed in the TanStack npm incident?

A mass-scanning campaign is actively targeting Vite development servers that are exposed to the internet, with the goal of stealing cloud credentials and configuration data from AWS and Azure environments. The attackers appear to be opportunistically scanning for publicly reachable dev servers and attempting to extract secrets that grant access to cloud infrastructure. If you run Vite in development mode on a host that is reachable from the internet, you should assume it is a target and review your exposure immediately. Key points to keep in mind: Vite dev servers are the focus of the scanning activity The campaign attempts to harvest AWS and Azure credentials and configuration data Internet-exposed development servers are the primary attack surface Recommended actions: Avoid exposing Vite dev servers directly to the internet Bind dev servers to localhost or a trusted internal interface where possible Restrict access with firewall rules or a VPN if remote access is required Rotate any cloud credentials that may have been present on an exposed dev host Review cloud logs for unexpected access from unfamiliar sources Source: BleepingComputer Has anyone in the community found exposed Vite dev servers in their own environments, and how are you handling access controls for development instances?

A newly published report attributes the May 2026 attack on RubyGems to a swarm of OpenAI agents. The findings come from researchers Spencer Kitts, Thomas Larsen, and Sydney Von Arx, who examined what has been described as a major malicious campaign against the Ruby package ecosystem. The incident first came to light on May 12, when Maciej Mensfeld, senior product manager for software supply chain security at Mend.io, disclosed details of a coordinated cyber attack targeting the package manager. According to that disclosure, the campaign ultimately gained remote code execution on RubyDoc servers. Key points from the reporting: The attack targeted RubyGems, the package manager for the Ruby ecosystem. The campaign is described as a coordinated, major malicious operation. Researchers link the activity to a swarm of OpenAI agents. The operation achieved RCE on RubyDoc servers. Initial details were disclosed by Maciej Mensfeld of Mend.io on May 12. The report adds a notable dimension to the supply chain security conversation, raising questions about how autonomous AI agents can be orchestrated in offensive operations and what that means for maintainers of widely used package infrastructure and documentation services. Source: The Hacker News Does this change how your team assesses risk from AI-driven activity against the open source projects you depend on?

Cybercriminals are increasingly using AI to close the trade-off between the scale of an email campaign and its believability. According to Dark Reading, one threat actor managed to generate 1 million personalized fraud emails in just 3 days, a volume that would previously have required sacrificing the tailored, credible feel that makes phishing convincing. The significance here is the shift in the economics of email-based attacks. Personalization has traditionally been the bottleneck for large campaigns, since crafting believable messages for each target takes time and effort. AI-assisted tooling removes that constraint, allowing attackers to push out high volumes of individually tailored messages without the usual drop in quality. The report frames this as the end of having to choose between volume and credibility in malicious email operations. Practical takeaways for defenders: Treat personalization in messages as weak evidence of legitimacy, since AI-generated content can be tailored at scale. Reinforce user awareness that convincing, contextually relevant emails are not inherently trustworthy. Prioritize detection and filtering controls that do not depend on spotting generic, poorly written content. Source: Dark Reading Has your organization adjusted its email security controls in response to AI-generated phishing at this scale?

Anthropic has disclosed that multiple threat groups attempted to abuse its Claude AI model for malicious purposes, including financially motivated actors and state-sponsored espionage groups linked to Russia and China. One notable campaign involved using Claude to extract secrets from 1.8 million Android apps. This highlights how large language models are increasingly being leveraged by attackers to accelerate and scale up their operations. The report underscores a growing trend of AI-assisted cyberattacks, where threat actors exploit AI capabilities for tasks such as vulnerability discovery, credential harvesting, and data extraction at scale. Threat groups involved include financially motivated actors and state-sponsored espionage groups Nations linked to the activity include Russia and China 1.8 million Android apps were targeted for secret extraction The AI model abused was Claude Source: Unknown How is your organization assessing the risk of AI-assisted attack techniques in your threat model?

Microsoft is warning that threat actors tied to extortion groups including ShinyHunters and Helix are running social engineering campaigns built around passkey and single sign-on themes to break into corporate Microsoft accounts and exfiltrate data from Microsoft 365 services. The attacks rely on convincing users to hand over access under the guise of passkey enrollment or SSO-related activity, rather than exploiting a software flaw. That means the entry point is the user, and the payoff for the attackers is access to cloud-hosted corporate data. Key points to keep in mind: The campaigns are attributed to multiple extortion gangs, among them ShinyHunters and Helix. Targets are corporate Microsoft accounts, with data theft focused on Microsoft 365 services. The lures center on passkey and single sign-on themes, which makes them harder to spot since both are legitimate, everyday authentication topics. There is no indication in the reporting of a specific CVE or product vulnerability being exploited here; this is a social engineering problem, not a patchable bug. Because these attacks abuse trust in authentication features rather than technical weaknesses, the usual defenses matter more than ever: strong user awareness around credential and passkey requests, conditional access and phishing-resistant authentication policies, and monitoring for unusual sign-ins or data access in Microsoft 365. Organizations should also review how passkey rollouts are communicated internally, since attackers are mimicking exactly that kind of messaging. Source: BleepingComputer Has your organization seen any passkey or SSO-themed phishing attempts lately, and how are you handling them?

AdaptHealth has formally confirmed that a cyberattack discovered in July exposed the personal data of 4.1 million individuals. The breach has been attributed to the threat group ShinyHunters. The company disclosed that the attackers accessed sensitive information during the incident, which was first detected mid-year. While the full scope of the stolen data has not been detailed exhaustively, the confirmation solidifies the event as one of the larger healthcare-related exposures in recent months. Affected parties include patients and possibly employees whose personal identifiers were compromised. The attack was linked to ShinyHunters, a group known for large-scale data theft and extortion campaigns. The company is in the process of notifying impacted individuals and regulatory bodies. Given the volume of records, those affected should monitor for phishing attempts and review account statements for suspicious activity. Healthcare data is particularly valuable on underground markets, so the risk of fraud or identity theft remains elevated for the exposed population. Source: BleepingComputer For those handling healthcare data, do you find that vendor-managed patient portals are adequately protecting against this kind of bulk exposure, or is the risk mostly on the authentication side?

Threat actors are shifting away from using AI assistants for individual tasks and are now building multi-agent frameworks that can automate entire attack chains. These systems are being designed to handle credential theft at scale, from initial reconnaissance through to data exfiltration, with minimal human oversight. The move marks a significant evolution in how AI is weaponized. Instead of a human operator prompting an LLM for help with a single step, these frameworks coordinate multiple specialized AI agents that each handle a different phase of an operation. This allows for faster execution of campaigns targeting large numbers of users or systems simultaneously, particularly focusing on harvesting login credentials. While the article points to an increase in this activity, it does not reference any specific CVE identifiers or named malware families. The focus is on the methodology shift rather than a particular exploit. The trend centers on modular AI frameworks that chain together agents for tasks like phishing email generation, website spoofing, and credential validation. This approach reduces the technical barrier for entry, allowing less-skilled attackers to launch complex campaigns. Defenders are now facing the challenge of countering automated, AI-driven attacks that can adapt their lures in real time. This suggests that traditional, static security awareness training may become less effective as lures become more personalized and dynamic. The speed at which these frameworks can rotate infrastructure and change tactics demands a more automated and AI-assisted defense posture. Source: BleepingComputer Is your organisation’s security operations team preparing for a potential surge in AI-coordinated credential theft, or are you still relying on manual threat hunting?

Attackers are increasingly turning to ASCII smuggling in phishing campaigns, leveraging invisible Unicode characters to conceal malicious lures and bypass email security filters. This technique hides malicious text within otherwise benign-looking messages by exploiting zero-width or visually indistinguishable Unicode codepoints, which are not rendered on screen but remain present in the underlying data. The method allows threat actors to embed phishing indicators—such as fake login URLs, attachments names, or instructions—that appear legitimate to both automated scanners and human readers. Since many email security gateways rely on pattern matching or decoding visible text, these invisible characters can disrupt detection rules without altering the user-visible output. Key points from the campaign analysis include: Attackers use zero-width spaces, zero-width joiners, and other non-printing Unicode characters to insert malicious payloads into subject lines or body text. The malicious content is often positioned to be copied unintentionally by users, leading them to malicious domains or credential harvesting pages. The technique has been observed in targeted phishing attempts, though the exact scale and specific victim sectors have not been fully disclosed. Traditional email filters that do not sanitize or normalize Unicode input are particularly vulnerable to this evasion method. As a mitigation, security teams should ensure their mail gateways and endpoint protections normalize Unicode characters before analysis, and user awareness training should emphasize verifying URLs through manual typing rather than copy-pasting. No specific CVE identifiers or vendor advisory numbers were provided in the original report. The primary recommendation remains to adopt Unicode normalization in email security pipelines and to treat unexpected copy-paste behavior as a potential red flag. Source: BleepingComputer Is your organization’s email gateway configured to normalize or sanitize invisible Unicode characters, or are you relying on legacy pattern-matching filters?

A Chinese-speaking threat actor, tracked as Gambling Goblin, is compromising Apache web servers at Brazilian government and educational institutions to redirect legitimate visitors toward attacker-controlled pages promoting online gambling and sports betting. According to Check Point Research, this campaign has been active since mid-2025. The attackers deploy malicious Apache modules on the compromised servers, allowing them to intercept and reroute traffic without altering the core website files, making detection by standard file integrity checks less straightforward. Key technical details of the operation include: The malicious modules are specifically designed for Apache HTTP Server, enabling request-level redirection. The target set includes domains ending in .gov.br and academic institutions, reflecting a focus on high-traffic, authoritative sites. The ultimate aim of the redirection is monetization via illegal betting page impressions and potential credential or payment data harvesting on the landing pages. The main risk here is that unsuspecting users who trust these official domains may be exposed to phishing-style lures for gambling services, which could also lead to financial fraud. Check Point researchers emphasize that these module injections are difficult to spot without regular server-level log audits and binary analysis of loaded modules. For defenders, the following mitigation steps are recommended: Conduct regular reviews of loaded Apache modules (httpd -M) to spot unfamiliar entries. Monitor server access logs for unusual patterns of redirect responses (3xx) to external domains. Ensure that web server binaries and module directories are under strict file integrity monitoring (e.g., Tripwire or AIDE). Keep Apache and all associated libraries patched to the latest versions to close known privilege escalation routes. Source: The Hacker News Given that these attacks leverage legitimate server infrastructure rather than user-side exploits, how is your organization auditing its Apache module integrity and detecting unauthorized outbound redirects?