data exfiltration | Breaking Cybersecurity News | The Hacker News

A novel attack technique named EchoLeak has been characterized as a "zero-click" artificial intelligence (AI) vulnerability that allows bad actors to exfiltrate sensitive data from Microsoft 365 Copilot's context sans any user interaction. The critical-rated vulnerability has been assigned the CVE identifier CVE-2025-32711 (CVSS score: 9.3). It requires no customer action and has been already addressed by Microsoft. There is no evidence that the shortcoming was exploited maliciously in the wild. "AI command injection in M365 Copilot allows an unauthorized attacker to disclose information over a network," the company said in an advisory released Wednesday. It has since been added to Microsoft's Patch Tuesday list for June 2025, taking the total number of fixed flaws to 68. Aim Security, which discovered and reported the issue, said it's instance of a large language model (LLM) Scope Violation that paves the way for indirect prompt injection , leadi...

Google on Wednesday disclosed that the Chinese state-sponsored threat actor known as APT41 leveraged a malware called TOUGHPROGRESS that uses Google Calendar for command-and-control (C2). The tech giant, which discovered the activity in late October 2024, said the malware was hosted on a compromised government website and was used to target multiple other government entities. "Misuse of cloud services for C2 is a technique that many threat actors leverage in order to blend in with legitimate activity," Google Threat Intelligence Group (GTIG) researcher Patrick Whitsell said . APT41, also tracked as Axiom, Blackfly, Brass Typhoon (formerly Barium), Bronze Atlas, Earth Baku, HOODOO, RedGolf, Red Kelpie, TA415, Wicked Panda, and Winnti, is the name assigned to a prolific nation-state group known for its targeting of governments and organizations within the global shipping and logistics, media and entertainment, technology, and automotive sectors. In July 2024, Google reve...

Cybersecurity researchers have uncovered malicious packages uploaded to the Python Package Index (PyPI) repository that act as checker tools to validate stolen email addresses against TikTok and Instagram APIs. All three packages are no longer available on PyPI. The names of the Python packages are below - checker-SaGaF (2,605 downloads) steinlurks (1,049 downloads) sinnercore (3,300 downloads) "True to its name, checker-SaGaF checks if an email is associated with a TikTok account and an Instagram account," Socket researcher Olivia Brown said in an analysis published last week. Specifically, the package is designed to send HTTP POST requests to TikTok's password recovery API and Instagram's account login endpoints to determine if an email address passed as input is valid, meaning there exists an account holder corresponding to that email address. "Once threat actors have this information, just from an email address, they can threaten to dox or spam, c...

Several ransomware actors are using a malware called Skitnet as part of their post-exploitation efforts to steal sensitive data and establish remote control over compromised hosts. "Skitnet has been sold on underground forums like RAMP since April 2024," Swiss cybersecurity company PRODAFT told The Hacker News. "However, since early 2025, we have observed multiple ransomware operators using it in real-world attacks."  "For example, in April 2025, Black Basta leveraged Skitnet in Teams-themed phishing campaigns targeting enterprise environments. With its stealth features and flexible architecture, Skitnet appears to be gaining traction rapidly within the ransomware ecosystem." Skitnet , also called Bossnet , is a multi-stage malware developed by a threat actor tracked by the company under the name LARVA-306. A notable aspect of the malicious tool is that it uses programming languages like Rust and Nim to launch a reverse shell over DNS and evade detecti...

Cybersecurity researchers have discovered three malicious Go modules that include obfuscated code to fetch next-stage payloads that can irrevocably overwrite a Linux system's primary disk and render it unbootable. The names of the packages are listed below - github[.]com/truthfulpharm/prototransform github[.]com/blankloggia/go-mcp github[.]com/steelpoor/tlsproxy "Despite appearing legitimate, these modules contained highly obfuscated code designed to fetch and execute remote payloads," Socket researcher Kush Pandya said . The packages are designed to check if the operating system on which they are being run is Linux, and if so retrieve a next-stage payload from a remote server using wget. The payload is a destructive shell script that overwrites the entire primary disk (" /dev/sda ") with zeroes, effectively preventing the machine from booting up. "This destructive method ensures no data recovery tool or forensic process can restore the data, as...

The malware loader known as MintsLoader has been used to deliver a PowerShell-based remote access trojan called GhostWeaver. "MintsLoader operates through a multi-stage infection chain involving obfuscated JavaScript and PowerShell scripts," Recorded Future's Insikt Group said in a report shared with The Hacker News. "The malware employs sandbox and virtual machine evasion techniques, a domain generation algorithm (DGA), and HTTP-based command-and-control (C2) communications." Phishing and drive-by download campaigns distributing MintsLoader have been detected in the wild since early 2023, per Orange Cyberdefense . The loader has been observed delivering various follow-on payloads like StealC and a modified version of the Berkeley Open Infrastructure for Network Computing (BOINC) client. The malware has also been put to use by threat actors operating e-crime services like SocGholish (aka FakeUpdates) and LandUpdate808 (aka TAG-124), distributing via p...

As the field of artificial intelligence (AI) continues to evolve at a rapid pace, fresh research has found how techniques that render the Model Context Protocol ( MCP ) susceptible to prompt injection attacks could be used to develop security tooling or identify malicious tools, according to a new report from Tenable. MCP, launched by Anthropic in November 2024, is a framework designed to connect Large Language Models (LLMs) with external data sources and services, and make use of model-controlled tools to interact with those systems to enhance the accuracy, relevance, and utility of AI applications. It follows a client-server architecture, allowing hosts with MCP clients such as Claude Desktop or Cursor to communicate with different MCP servers, each of which exposes specific tools and capabilities. While the open standard offers a unified interface to access various data sources and even switch between LLM providers, they also come with a new set of risks, ranging from exc...

Government and telecommunications sectors in Southeast Asia have become the target of a "sophisticated" campaign undertaken by a new advanced persistent threat (APT) group called Earth Kurma since June 2024. The attacks, per Trend Micro, have leveraged custom malware, rootkits, and cloud storage services for data exfiltration. The Philippines, Vietnam, Thailand, and Malaysia are among the prominent targets. "This campaign poses a high business risk due to targeted espionage, credential theft, persistent foothold established through kernel-level rootkits, and data exfiltration via trusted cloud platforms," security researchers Nick Dai and Sunny Lu said in an analysis published last week. The threat actor's activities date back to November 2020, with the intrusions primarily relying on services like Dropbox and Microsoft OneDrive to siphon sensitive data using tools like TESDAT and SIMPOBOXSPY. Two other noteworthy malware families in its arsenal include r...

Microsoft is calling attention to an ongoing malvertising campaign that makes use of Node.js to deliver malicious payloads capable of information theft and data exfiltration. The activity, first detected in October 2024, uses lures related to cryptocurrency trading to trick users into installing a rogue installer from fraudulent websites that masquerade as legitimate software like Binance or TradingView. The downloaded installer comes embedded with a dynamic-link library ("CustomActions.dll") that's responsible for harvesting basic system information using Windows Management Instrumentation (WMI) and setting up persistence on the host via a scheduled task. In an attempt to keep up the ruse, the DLL launches a browser window via " msedge_proxy.exe " that displays the legitimate cryptocurrency trading website. It's worth noting that "msedge_proxy.exe" can be used to display any website as a web application. The scheduled task, in the meanwhile...

Cybersecurity researchers have discovered a new, sophisticated remote access trojan called ResolverRAT that has been observed in attacks targeting healthcare and pharmaceutical sectors. "The threat actor leverages fear-based lures delivered via phishing emails, designed to pressure recipients into clicking a malicious link," Morphisec Labs researcher Nadav Lorber said in a report shared with The Hacker News. "Once accessed, the link directs the user to download and open a file that triggers the ResolverRAT execution chain." The activity, observed as recently as March 10, 2025, shares infrastructure and delivery mechanism overlap with phishing campaigns that have distributed information stealer malware such as Lumma and Rhadamanthys, as documented by Cisco Talos and Check Point last year. A notable aspect of the campaign is the use of localized phishing lures, with the emails crafted in the languages predominantly spoken in the targeted countries. This includ...

Cybersecurity researchers have disclosed details of a now-patched privilege escalation vulnerability in Google Cloud Platform (GCP) Cloud Run that could have allowed a malicious actor to access container images and even inject malicious code. "The vulnerability could have allowed such an identity to abuse its Google Cloud Run revision edit permissions in order to pull private Google Artifact Registry and Google Container Registry images in the same account," Tenable security researcher Liv Matan said in a report shared with The Hacker News. The security shortcoming has been codenamed ImageRunner by the cybersecurity company. Following responsible disclosure, Google addressed the problem as of January 28, 2025. Google Cloud Run is a fully managed service for executing containerized applications in a scalable, serverless environment. When the technology is used to run a service, container images are retrieved from the Artifact Registry (or Docker Hub) for subsequent depl...

The financially motivated threat actor known as FIN7 has been linked to a Python-based backdoor called Anubis (not to be confused with an Android banking trojan of the same name) that can grant them remote access to compromised Windows systems. "This malware allows attackers to execute remote shell commands and other system operations, giving them full control over an infected machine," Swiss cybersecurity company PRODAFT said in a technical report of the malware. FIN7, also called Carbon Spider, ELBRUS, Gold Niagara, Sangria Tempest, and Savage Ladybug, is a Russian cybercrime group known for its ever-evolving and expanding set of malware families for obtaining initial access and data exfiltration. In recent years, the threat actor is said to have transitioned to a ransomware affiliate. In July 2024, the group was observed using various online aliases to advertise a tool called AuKill (aka AvNeutralizer) that's capable of terminating security tools in a likely ...

Cybersecurity researchers have shed light on a new China-linked threat actor called Earth Alux that has targeted various key sectors such as government, technology, logistics, manufacturing, telecommunications, IT services, and retail in the Asia-Pacific (APAC) and Latin American (LATAM) regions. "The first sighting of its activity was in the second quarter of 2023; back then, it was predominantly observed in the APAC region," Trend Micro researchers Lenart Bermejo, Ted Lee, and Theo Chen said in a technical report published Monday. "Around the middle of 2024, it was also spotted in Latin America." The primary targets of the adversarial collective span countries such as Thailand, the Philippines, Malaysia, Taiwan, and Brazil. The infection chains begin with the exploitation of vulnerable services in internet-exposed web applications, using them to drop the Godzilla web shell for facilitating the deployment of additional payloads, including backdoors dubbed VA...

Cybersecurity researchers have discovered several cryptocurrency packages on the npm registry that have been hijacked to siphon sensitive information such as environment variables from compromised systems. "Some of these packages have lived on npmjs.com for over 9 years, and provide legitimate functionality to blockchain developers," Sonatype researcher Ax Sharma said . "However, [...] the latest versions of each of these packages were laden with obfuscated scripts." The affected packages and their hijacked versions are listed below - country-currency-map (2.1.8) bnb-javascript-sdk-nobroadcast (2.16.16) @bithighlander/bitcoin-cash-js-lib (5.2.2) eslint-config-travix (6.3.1) @crosswise-finance1/sdk-v2 (0.1.21) @keepkey/device-protocol (7.13.3) @veniceswap/uikit (0.65.34) @veniceswap/eslint-config-pancake (1.6.2) babel-preset-travix (1.2.1) @travix/ui-themes (1.1.5) @coinmasters/types (4.8.16) Analysis of these packages by the software supply chain ...

An advanced persistent threat (APT) group with ties to Pakistan has been attributed to the creation of a fake website masquerading as India's public sector postal system as part of a campaign designed to infect both Windows and Android users in the country. Cybersecurity company CYFIRMA has attributed the campaign with medium confidence to a threat actor called APT36 , which is also known as Transparent Tribe.  The fraudulent website mimicking India Post is named "postindia[.]site." Users who land on the site from Windows systems are prompted to download a PDF document, whereas those visiting from an Android device are served a malicious application package ("indiapost.apk") file. "When accessed from a desktop, the site delivers a malicious PDF file containing ' ClickFix ' tactics," CYFIRMA said . "The document instructs users to press the Win + R keys, paste a provided PowerShell command into the Run dialog, and execute it – potentia...