The npm package 'typecript-cli' version 1.0.0 is a typosquatting malicious package impersonating 'typescript'. Its postinstall script performs an install-time dropper attack by downloading and executing an unsigned executable from a personal GitHub repository. It uses obfuscation techniques such as XOR-decoding payload strings and assembling IP addresses from split arrays to evade detection. On Windows, it downloads and runs 'main.exe' without verification. Under Windows Subsystem for Linux (WSL), it escapes the Linux sandbox to execute the same payload via PowerShell. The script also sends a beacon with platform information to a hardcoded IP address, likely for command and control or host selection purposes.
The npm package 'typescipt-core' version 1.0.0 contains malicious code that executes during installation. It uses obfuscated scripts to download and run a Windows executable and a PowerShell bridge, sending host profile data to a hardcoded IP address. This behavior enables attacker-controlled code execution on the installer's system.
The npm package 'tyepescript-cli' is a typosquatting malicious package impersonating the legitimate 'typescript' package. Its postinstall script uses XOR obfuscation to decode attacker-controlled URLs and commands. On Windows, it downloads and executes a malicious binary from a suspicious GitHub account. Under Windows Subsystem for Linux (WSL), it uses a PowerShell bridge to execute the same payload on the Windows host. The script also sends an install-time beacon to a hardcoded IP address. This behavior indicates an active attempt to compromise systems during package installation.
Attackers are exploiting Steam forums by posting malicious PowerShell commands disguised as troubleshooting advice. The command downloads and executes a script that installs a crypto miner (XMRig) on victims' Windows machines. The script mimics a system optimization tool to deceive users and disables antivirus scanning on the miner's folder. The miner hijacks system resources to mine Monero cryptocurrency without user consent, degrading performance and hardware lifespan. The attack relies on social engineering to trick gamers into running the harmful command with administrator privileges.
CVE-2026-50523 is a command injection vulnerability in Microsoft PowerShell 7.4 that allows an authorized local attacker to execute arbitrary code due to improper neutralization of special elements in commands. The flaw affects multiple versions of PowerShell 7.4, 7.5, and 7.6 prior to their respective fixed releases. A high-severity vulnerability with a CVSS score of 7.8, it requires local privileges but no user interaction. Official patches have been released to address this issue.
Improper control of generation of code ('Code Injection') in the settings feature in Devolutions PowerShell Universal 2026.2.3 and earlier allows an authenticated user with settings management permission to execute arbitrary PowerShell code via a crafted setting value that is not properly escaped when written to the settings configuration file.
OpenClaw CLI's system.run allowlist mode did not properly detect PowerShell encoded-command wrappers such as -EncodedCommand, -enc, or -e. This allowed callers with access to system.run to execute PowerShell inline payloads without the usual approval step required for equivalent -Command invocations. The issue was fixed in version 2026.3.7 by updating the shell-wrapper parsing to recognize these encoded-command aliases, ensuring allowlist mode enforces approval consistently.
OpenClaw CLI versions from 2026.1.29 up to but not including 2026.5.12 have a vulnerability where PowerShell encoded-command aliases may bypass execution allowlist checks. This allows encoded PowerShell commands to run without proper allowlist validation if abbreviated alias forms are used. The vulnerability affects configurations where the feature is enabled and reachable, potentially allowing execution of unauthorized encoded PowerShell content. The issue does not alter the trusted-operator model of OpenClaw. A patch is available starting with version 2026.5.12. Until patched, it is recommended to avoid allowlisting PowerShell wrapper forms and require approval for encoded commands. Additional mitigations include narrowing allowlists, avoiding shared Gateways between untrusted users, and disabling the affected feature if not needed.
APT-C-06 (Darkhotel) is an APT organization that has been active since at least 2007, targeting corporate executives, defense industries, and electronics sectors. In April 2026, the group launched phishing attacks using a decoy document titled 'North Korean Central Television Real-time Broadcasting Program Instructions.' The document instructs users to download an application for watching North Korean Central Television. By late May, attacks evolved to deliver malicious MSI files through phishing emails. These MSI files execute VBS code that creates scheduled tasks to download and execute PowerShell scripts, which then retrieve subsequent payloads. The malware employs ChaCha20 encryption and ultimately deploys shellcode. PowerShell has become a high-frequency component in APT-C-06's attack chain since 2025, handling payload downloads and persistence mechanisms.
A threat hunting investigation identified suspicious PowerShell content served from an IP address (203.188.171.166) and domain (dorenzaa.com), both retrieving ZIP archives from Vercel-hosted infrastructure. The PowerShell loaders extract and execute payloads locally, including Grape.exe, UltraToolliteSetup.exe, and draw.io.exe. Analysis revealed heavily obfuscated PowerShell stages utilizing Base64 encoding, XOR-based obfuscation with the key 'Write', dynamically constructed IEX commands, and hidden PowerShell execution. A decoy 'Verification complete!' message disguised as Google.com was presented to victims during execution. Multiple Vercel instances hosted additional artifacts including loader scripts and executables. The initial infection vector remains unidentified, suggesting these PowerShell-hosting URLs represent second-stage delivery points in a multi-stage attack chain.
The Kimsuky threat group has integrated artificial intelligence capabilities into its attack operations, establishing local large language model environments using Ollama, GPT4All, and Msty. Evidence indicates the group is accumulating technologies to incorporate AI across attack operations, including AI-generated decoy documents and retrieval-augmented generation for document analysis. The campaign, dubbed Operation GitPower, continues targeting foreign diplomatic missions and sectors including military, security, and virtual assets. Attacks utilize malicious LNK files contained in ZIP archives, executing obfuscated PowerShell scripts that abuse Git-based repositories as command-and-control infrastructure. The group distributes encrypted AsyncRAT payloads disguised as image files through GitHub. Linguistic indicators including North Korean vocabulary patterns such as "싸이트", "가입리력", and "로출되였는지" support attribution to North Korean state-sponsored operations under the Reconnaissance General Bureau.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (9476bbe7c2ce12348142e93deffe53c9d4d6dfef3400030af81cff902bac5935)
This package impersonates the @rbxts/services Roblox typings library and re-exports its public API in index.js as cover. Its postinstall script (scripts/postinstall.js) is gated to win32 and uses \x hex escapes to hide the strings 'child_process', 'spawn', 'powershell', and related identifiers. It spawns a hidden PowerShell process with -ExecutionPolicy Bypass, passing a base64-encoded body that decodes to an AMSI bypass (patched via reflection with strings split across variables to evade signature scanning), downloads a binary from https://files.catbox.moe/a3loxy.bin, allocates RWX memory via VirtualAlloc with protection 0x40, copies the downloaded bytes into it, and executes them via CreateThread. Installing this package on Windows results in arbitrary attacker shellcode executing in the installer's process on `npm install`.
The npm package @rbxst/services version 1.0.756 contains malicious code that executes during installation on Windows systems. The postinstall.js script downloads a ZIP archive from an anonymous file-hosting service, extracts it using PowerShell, and runs a hidden Python payload with the installer's privileges. This behavior is suspicious and matches known malicious patterns of remote payload execution. The package falsely advertises Roblox-TS service utilities but has no legitimate reason to fetch and run external Python code. This threat does not have a CVSS score but is assessed as high severity due to the execution of opaque remote code with user privileges.
The npm package trimprompt versions 1.0.35, 1.0.42, 1.0.46, 1.0.47, 1.0.48, and 1.0.49 contains heavily obfuscated JavaScript modules that execute malicious behavior. These modules include scripts that spawn PowerShell processes during installation and runtime on Windows hosts, and perform HTTP POST requests sending host identifiers to remote endpoints. This behavior is inconsistent with the package's intended functionality and indicates a supply-chain compromise with install-time execution and host beaconing capabilities.
The elephant-tusk-runner npm package versions 1.0.0 and 1.0.1 contain malicious code that starts an Express and WebSocket server bound to all network interfaces with open CORS and no authentication. This server exposes endpoints that allow remote attackers to spawn interactive shells and execute arbitrary commands on the host machine. The exposed functionality includes spawning bash or PowerShell shells, executing commands like git clone, npm install, PHP artisan commands, and killing arbitrary processes. This creates a severe remote code execution and remote shell access risk on any machine running the tool, especially on shared or reachable networks.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (534ea1156519b64249657c1495faacdb3cc5d8e250a00e0dc2f8fd4e991af61c)
The package's bin entries (`wormgpt-cli`, `wormgpt`, `deepholegpt`, `wgpt`) map to `bin/victim.js`, which silently spawns a detached implant on any invocation (including `--help`/`--version`). The implant connects to a hardcoded C2 at http://13.60.13.215:7771 over an AES-256-GCM/HMAC-signed protocol with pool failover and a DNS TXT dead-drop channel (`protocol.js` `pollDnsDeaddrop`) that rotates the C2 URL and shared secret at runtime. On beacon, `implant.js` `gatherInfo()` collects hostname, user, OS, architecture, PID, cwd, home, LAN IP, admin flag, and a persistent victim_id, and accepts opcodes for shell/PowerShell execution, interactive PTY reverse shell, file up/download, keylogging (`GetAsyncKeyState` PowerShell loop), clipboard capture, screenshots, LAN lateral-movement scanning, privilege-escalation checks, and anti-forensics wipe. `stealers/browser.js` reads Chrome/Edge/Brave/Opera/Vivaldi Login Data SQLite databases, decrypts DPAPI-protected passwords via PowerShell with an AMSI bypass stub (`AMSI_BYPASS_B64` base64 blob run through `powershell -EncodedCommand`), parses Firefox `logins.json`, and regex-scans Discord/Chrome LevelDB for auth tokens (`/[\w-]{24}\.[\w-]{6}\.[\w-]{27}|mfa\.[\w-]{84}/g`), exfiltrating results via C2 opcode OP_BROWSER_DATA. `persist.js` installs persistence across six autostart layers per OS: Windows HKCU/HKLM Run keys, Startup.lnk, WMI __EventFilter/CommandLineEventConsumer permanent subscription, scheduled tasks, and hidden `Windows Helper` directories via `attrib +H +S`; Linux XDG autostart, systemd --user units, `@reboot` crontab, shell RC injection (`.bashrc`/`.zshrc`/`.profile`/`.xinitrc`), and `/etc/rc.local`; macOS LaunchAgent plist with KeepAlive and Login Items via osascript. `loader.js` runs a watchdog sibling process (poll every 20s, `MAX_REVIVES = 50`) plus `uncaughtException` respawn to keep the implant alive.
CVE-2026-71312 is a high-severity OS command injection vulnerability in rclone versions prior to 1.75.0. The flaw arises from improper escaping of certain Unicode single-quote characters in PowerShell commands used for server-side hashing of remote SFTP paths. This allows an attacker to inject and execute arbitrary PowerShell commands with the privileges of the victim's SSH account. The issue is fixed in rclone version 1.75.0.