The solana-utils-terminalmoney1 package on PyPI is a malicious package that executes a stealer script during installation. It exfiltrates sensitive environment variables, credentials, and cryptocurrency-related data from various local files and browser extensions. The package impersonates the legitimate solana-py project by using similar metadata and README content. The stolen data is sent to a hardcoded external endpoint tagged with the installer's hostname.
Two malicious versions (1.26.1 and 1.26.2) of the Microsoft Agent Framework Python package on PyPI were released in a supply chain attack. These versions steal Azure, Azure DevOps, and GitHub Actions credentials by executing code automatically when Python starts, even if the package is not explicitly imported. The malicious packages retained legitimate functionality while adding credential-stealing code. The compromised versions have been removed from PyPI, and the latest clean release is 1.26.0. Organizations are advised to check for these versions, replace them with verified packages, rotate exposed credentials, and investigate suspicious activity.
On November 17, 2026, two malicious versions (1.26.1 and 1.26.2) of Microsoft's agent-framework Python package were briefly published on PyPI before removal. These compromised packages deployed credential-stealing malware via Python .pth files that executed automatically on interpreter startup. The malware targeted Azure and CI/CD environments, extracting secrets from environment variables, credential files, and tokens including Azure managed-identity and OIDC tokens. Version 1.26.2 added the ability to steal GitHub Actions OIDC tokens from workflows with federated credential access. Stolen data was exfiltrated through a hardcoded webhook.site URL. The root cause appears to be unauthorized direct upload to PyPI outside the normal release process, though this remains unconfirmed.
Microsoft's agent-framework package on PyPI was compromised on November 17th, 2026. Versions 1.26.1 and 1.26.2 contained malicious code that installs Python startup hooks via .pth files, which execute silently when the interpreter initializes. The malicious code collects credentials related to Azure, GitHub Actions, and various CI/CD environments, including environment variables, configuration files, and cloud identity tokens. It actively requests fresh Azure managed-identity and GitHub Actions OIDC tokens. The stolen data is compressed, base64-encoded, and exfiltrated to a webhook.site endpoint. The attack targets credentials from multiple tools such as pip, npm, Docker, and Git.
The npm package @testsecuiitytestme/cache-helper is a malicious package that installs a credential stealer and self-propagating worm. It executes a native addon at install time that loads an obfuscated Bun program to collect a wide range of credentials and secrets from developer machines. The malware targets credentials from npm, GitHub, GitLab, PyPI, RubyGems, cloud providers (AWS, GCP, Azure), HashiCorp Vault, Kubernetes, SSH keys, browser logins, and password managers. It also propagates by publishing malicious packages and injecting CI/CD workflow changes and AI assistant configurations. The package was published and updated multiple times on October 27, 2026, then unpublished shortly after. Users who installed any version should consider all accessible credentials compromised and rotate them immediately.
The LiteLLM project's release pipeline was compromised on October 21, 2026, resulting in the distribution of malicious versions of the LiteLLM PyPI package and Docker images. The compromised version 1.107.0 contained malware that steals credentials, secrets, environment variables, and other sensitive data from affected systems. The malware executes automatically on Python startup by placing malicious files in the Python library directory. The attack impacted all PyPI wheels for LiteLLM 1.107.0 and Docker images tagged with v1.107.0 and related tags. The malicious packages have been removed, and users are advised to stop affected deployments, isolate infected systems, and rotate exposed secrets. The recommended clean version is LiteLLM 1.106.2.
On October 21, 2026, the LiteLLM Python proxy library version 1.107.0 was compromised in a supply chain attack. The malicious release was available on PyPI and official Docker images for approximately two hours. Attackers used stolen GitHub CLI OAuth tokens from three engineers to deploy a malicious fork via a private release workflow. The payload included a credential stealer targeting cloud, developer, AI-tool, browser, and LiteLLM proxy credentials, as well as a second-stage remote access trojan backdoor. The malware executes automatically whenever Python starts by placing files in the site-packages directory. This is the second known supply chain compromise of LiteLLM in 2026, unrelated to the previous incident linked to TeamPCP activity.
The npm packages pixel-map-testmetest versions 0.0.1 and 0.0.2 contain malicious code that executes at install time via a node-gyp binding.gyp script. This script runs an obfuscated loader that decrypts and decompresses an encrypted payload, then executes it using the Bun runtime. The package masquerades as a benign placeholder but actually delivers a credential stealer payload. The malicious behavior is hidden through multi-layer obfuscation and use of an alternate runtime, making static analysis difficult.
The npm package test-brotli2-testmetest version 1.0.0 contains malicious code that executes during installation. Its install script triggers node-gyp to run an obfuscated JavaScript loader (index.js) that decrypts and executes a Brotli-compressed payload. This payload acts as a credential stealer. The package was published by the npm account testnewtestme2 on 2026-10-21.
log-sink-testmetest version 1.0.1 is a malicious npm package that executes a concealed credential stealer during installation. It abuses the node-gyp build process by embedding a binding.gyp file that triggers execution of an encrypted payload without any native source code. The payload is heavily obfuscated and decrypts an AES-128-GCM encrypted blob, which is then executed either directly in Node.js or via the Bun runtime to evade detection. The package was published on 2026-10-21 by the npm account 'testnewtestme2'. Version 1.0.0 is clean and unaffected.
The npm package event-emitter-testmetest contains malicious code in specific versions that use multi-layer obfuscation to execute attacker-controlled payloads. Versions 1.0.6 and 1.0.7 run an obfuscated loader that decrypts and executes code on load or during npm install via a binding.gyp hook. Versions 4.1.0 and 5.0.0 include a stealer payload executed similarly through binding.gyp. Other versions do not contain executable payloads.
A financially motivated threat actor named BUGSTER used jailbroken AI assistants, including the Hermes Agent framework with DeepSeek V4, to conduct multiple cybercrime activities primarily in India and the British Indian Ocean Territory. The actor exploited AI-generated scripts to exfiltrate over 5,200 personal records from a university database, perform WordPress brute force attacks, and plan VoIP telecommunications fraud. Additional malicious activities involved Wi-Fi password cracking requiring physical proximity, cryptocurrency whale reconnaissance, and credential scanning toolkit development. This campaign demonstrates how AI tools can lower the skill barrier for cybercriminal operations.
Two critical PHP remote code execution vulnerabilities will be patched in an emergency release on October 21. Fixes are only expected for supported PHP 8.x branches. Key takeaways CVE-2026-109618 and CVE-2026-109619 are critical vulnerabilities in PHP, both rated CVSSv4.0 9.2 by the PHP security team. Public reporting describes both as remote code execution vulnerabilities. watchTowr says it believes it has reproduced CVE-2026-109618, and Cloudflare CTO Dane Knecht says the company is seeing early attacks emerging. The patch release is scheduled for October 21, 2026 at 14:00 UTC. Version numbers reported before the release point to PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35. End-of-life branches are not expected to receive fixes. Background Tenable's Research Special Operations (RSO) team has compiled this blog to answer Frequently Asked Questions (FAQ) regarding two critical vulnerabilities in PHP that were announced on October 20, 2026. FAQ When were these PHP vulnerabilities first disclosed? On October 20, 2026, the CVE identifiers circulated on Hacker News and social media. Later that day, the PHP security team published an advance notice of a security release scheduled for October 21 at 14:00 UTC (10:00 AM ET) to address two critical vulnerabilities. The notice said the team could release earlier if details became public or if exploitation appeared in the wild. CERT-EU also published an advisory (2026-018) naming both CVEs. According to watchTowr , the vulnerabilities were originally set to be disclosed on Thursday, October 22, under embargo. A video shared by researcher Kirill Firsov demonstrates the exploit chain, dubbed php2shell. Firsov credited the video and exploit to researcher LCFR and Calif.io. The PHP security team is giving advance notice: Upcoming security release Wed, Oct 21st 2026, 14:00UTC, two critical vulnerabilities, both rated CVSSv4 9.2. PHP users are encouraged to prepare to update. Earlier release if details become public or active exploit evidence. — The PHP Foundation (@ThePHPF) October 20, 2026 What are CVE-2026-109618 and CVE-2026-109619? CVE Description CVSSv4.0 CVE-2026-109618 Unspecified PHP Vulnerability 9.2 CVE-2026-109619 Unspecified PHP Vulnerability 9.2 CVE-2026-109618 and CVE-2026-109619 are critical vulnerabilities in PHP. Public reporting describes both as remote code execution vulnerabilities. The PHP security team has not published technical details before the patch release. CVSS scores differ across sources. The PHP security team rates both vulnerabilities as CVSSv4.0 9.2. Some early reports, including watchTowr and a Hacker News post, describe them as CVSS 10.0. We will update this blog once PHP publishes its advisory. Which PHP versions are affected? CERT-EU says these vulnerabilities could affect all PHP versions from PHP 5 onwards, and the demonstration video references PHP 5.4.0 and later. PHP will publish the official list of affected and fixed versions with the release. Before the release, security researcher Kevin Beaumont reported the following fixed versions: Affected Versions Fixed Version (expected) PHP 8.5.x 8.5.12 PHP 8.4.x 8.4.27 PHP 8.3.x 8.3.36 PHP 8.2.x 8.2.35 PHP 5.x through 8.1.x End-of-life, no fix expected Linux distributions that ship older PHP packages may backport fixes, so organizations using distribution packages should check their vendor's advisories. How widespread is PHP, and how many hosts are potentially affected? According to W3Techs , PHP is used by 69.3% of websites whose server-side language is known. Platforms such as WordPress, Drupal, Joomla, Laravel and Magento all run on PHP. Security researcher Silas Cutler's rough estimate is that over 1.3 million internet-exposed hosts are potentially vulnerable. A Shodan search by Henk Poley counts over 1.9 million internet-exposed hosts running PHP. Has exploitation been observed in the wild? On October 21, Cloudflare CTO Dane Knecht said the company is "seeing early attacks emerging" and is updating its miti…
QuartzLine, also known as CLOUDHAZE, is malware disguised as a legitimate Windows PDF viewer that contains a backdoor. It has been distributed via malicious advertising campaigns, leveraging AWS S3 buckets for hosting. The malware establishes persistence through scheduled tasks and communicates with command-and-control servers. It operates with user-level privileges, allowing attackers to execute arbitrary Windows commands and deploy additional malware such as the Caira and IronBark remote access trojans. Early activity includes theft of Chrome browser data within one week of infection. The malware is signed with a now-revoked legitimate EV code-signing certificate, appearing as paperline.exe and fielddesk.exe.
The Psychedelic Stealer is an active malware campaign targeting Ukrainian small-business websites as of September 2026. It tricks victims with a fake Cloudflare CAPTCHA page to manually run msiexec.exe commands that install malware. This malware steals browser credentials, session tokens, and cryptocurrency wallet data from wallets such as MetaMask and Trust Wallet. It maintains persistence through browser extensions using native messaging and scheduled tasks, and communicates with a REST API for further commands. The campaign uses signed Microsoft binaries and avoids encoded PowerShell scripts to evade detection. It is financially motivated and specifically targets Ukrainian victims.
2CLoader is a newly identified malware loader active since August 2026 that delivers information stealers such as Vidar and Remus, as well as the XWorm RAT. It features advanced anti-analysis and evasion techniques, including anti-VM, anti-debug, indirect system calls via Hell's Gate, and inline trampoline hooks to bypass endpoint security. The loader protects its payloads with rolling XOR and AES-GCM encryption stored as PE resources and supports multiple execution methods. It communicates with command-and-control servers using encrypted JSON over HTTP POST and maintains persistence through registry keys, scheduled tasks, and startup folders.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (8f03abfe6149423144b6d3143a66cbca7925bb6eb736b1ece350d5a344684eb9)
@hrmony/pladddform-frontend@48.99.0 ships an index.js whose top-level code, executed on require/import, POSTs to the hardcoded attacker host https://hrnmn.dd.h4x.tv/save_instance_info_from_pkge_hrmony_pladddform-frontend and then invokes two harvesters. aws_get() queries the AWS EC2 Instance Metadata Service at 169.254.169.254 using IMDSv2 (PUT /latest/api/token) and sweeps /latest/meta-data/iam/security-credentials/ to collect instance identity, IAM role names, and temporary IAM security credentials. local_get() enumerates every entry of process.env and recursively reads /var/run/secrets/, capturing Kubernetes-mounted service-account tokens and secret files. Both payloads are POSTed to the same hardcoded hrnmn.dd.h4x.tv endpoint. The package has an empty description, no functional code beyond the stealer, an intentionally misspelled scope and name mimicking an internal 'harmony/platform-frontend', and declares a dependency on the similarly look-alike '@hrmony/kit-5', consistent with a dependency-confusion typosquat targeting an internal package name.
## Source: ghsa-malware (37ed7f1bf9d1ce61730644e3ac7eaf7f1525da3d26b9f9ba1f171d2ff2b72ae6)
Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it.
Threat actors exploited ChatGPT's Custom GPT feature to impersonate legitimate models and lure victims via sponsored Google ads to malicious Custom GPTs named 'Plus 5.6'. Victims were shown fake service notices and redirected to Google Sites hosting ClickFix lures disguised as CloudFlare CAPTCHA checks. They were tricked into running PowerShell commands that downloaded malicious MSI installers. The attack used DLL sideloading with legitimate Canon- and Stardock-signed executables to establish persistence via registry Run keys and scheduled tasks. The multi-stage infection involved obfuscated scripts, encrypted payloads hidden in WAV files and NuGet packages, and deployed a remote access trojan with capabilities including remote desktop access, browser hijacking, credential theft, and follow-on payload delivery. Approximately 40 incidents were investigated by Huntress.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (37cd33c0908189c0085fba153be1c42776411968aa5ad652dd7ffdef88b318bc)
hardhat-zet@2.0.1 is published under a name that impersonates the Hardhat Ethereum tooling ecosystem, but its README, type definitions and docs/ tree are copied verbatim from the unrelated `pino` logger project (badges and links point at npmjs.com/package/pino and github.com/pinojs/pino). The exported `middleware` in index.js is a no-op Express middleware and matches neither the `hardhat-zet` name nor the pino README. The package's declared main (index.js) top-level-requires./lib/config, which is not a configuration file but a 4,498,609-byte obfuscator.io-style bundle: a 26,113-entry rotated string array behind two runtime decoders (functions v and F), hex-encoded identifiers, and control-flow flattening. All strings and destination URLs are hidden behind the decoders, so this code runs automatically on any `require('hardhat-zet')` and its behavior is not observable without deobfuscation. Publisher metadata (author 'Robert King <hello@jsonspack.com>', bugs URL https://jsonspack.com/issues) ties to neither the Hardhat nor the pino identity. The combination — ecosystem-typosquat name targeting crypto/Hardhat developers, cover-story README lifted from an unrelated popular package, no-op public API, and a multi-megabyte obfuscated payload auto-executed on import — is the shape of a supply-chain stealer aimed at Ethereum developers.
## Source: ghsa-malware (781273d2c802d3781fecccd4434ba418c93ce88edf920b2a1163dbacf7576254)
Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it.
A Python-based Malware-as-a-Service builder enables operators to generate customized Windows infostealer executables. The system comprises a builder component and an embedded payload, using Nuitka or PyInstaller compilation to evade detection. The builder features automatic dependency installation, webhook configuration with XOR and Base64 encoding, and multiple compilation backends. The payload targets Chromium and Firefox browsers, extracting credentials, cookies, and credit card data. It harvests Wi-Fi passwords, Discord tokens, and Roblox session cookies while employing anti-analysis techniques including debugger detection, VM process blacklisting, disk size checks, and timing evasion. Persistence is established through registry Run keys and scheduled tasks. All stolen data is packaged into in-memory ZIP archives and exfiltrated via attacker-controlled webhooks, following a scalable affiliate model.