Last updated: November 2, 2026 at 1:20 PM UTC
All 957 Vulnerability 399 Breach 144 Threat 407 Defense 7

GodDamn ransomware uses a Microsoft-signed malicious driver to disable defenses

Symantec detailed GodDamn, a ransomware operation that disables endpoint defenses using PoisonX, a malicious kernel driver its developers managed to get signed by Microsoft, an unusual escalation over the more common tactic of abusing a legitimate vulnerable driver. In an early-June attack, the operators used AnyDesk for remote access and a credential-harvesting toolkit that pulls passwords from browsers, Windows Credential Manager, cached domain credentials, email clients, and network traffic, before deploying the ransomware. Alongside the signed driver, they ran a user-mode tool disguised as a Symantec product to blind security software. Symantec links GodDamn to a developer it tracks as Hyadina and says the group is actively improving its defense-evasion capabilities.

Check
Watch for bring-your-own-driver activity and processes masquerading as security products, audit remote-access tools like AnyDesk in your environment, and monitor for credential-harvesting across browsers and Windows credential stores.
Affected
Windows organizations where attackers gain a foothold; GodDamn uses a Microsoft-signed malicious driver to switch off endpoint defenses, harvests credentials broadly, then encrypts systems, making detection before deployment much harder.
Fix
Enable driver block lists and tamper protection, restrict who can load kernel drivers, tightly control remote-access software, enforce phishing-resistant MFA, and keep monitored offline backups so encryption stays recoverable.

GigaWiper bundles disk wiping and fake ransomware into one Windows backdoor

Microsoft detailed GigaWiper, a destructive Windows backdoor that stitches several older malware families into one implant, letting operators choose how to destroy a machine. Its commands include a raw disk wiper that overwrites the drive and partition table, a multi-pass wiper for the Windows drive, and a fake-ransomware module that encrypts files with keys it never saves, so there is no recovery and no ransom note, only destruction disguised as extortion. It also offers remote access and log clearing. To hide, it poses as OneDrive through a scheduled task and routes command traffic over legitimate services like RabbitMQ and Redis. Because it runs after intrusion, offline backups are the real defense.

Check
Hunt for GigaWiper's signs, including a scheduled task posing as OneDrive Update, command traffic to RabbitMQ or Redis services you do not use, and unexpected changes to disk partitions or recovery settings.
Affected
Windows organizations an attacker has already breached; GigaWiper can destroy disks and files irreversibly while disguising the destruction as ransomware, and can also spy and maintain remote access before it strikes.
Fix
Keep tested, offline backups since wiped data cannot be decrypted, enable tamper protection and endpoint detection in block mode, monitor for suspicious scheduled tasks and disk operations, and restrict outbound traffic.

HalluSquatting registers the fake package names AI coding tools invent

Researchers have shown an attack, called HalluSquatting, that weaponizes the tendency of AI coding assistants to invent plausible-sounding names for software packages that do not exist. The attackers work out which fake names an AI reliably hallucinates, register those names first on a package registry or GitHub, and upload a trap that instructs the coding agent to install a reverse shell or runs code directly. When the assistant fetches the made-up dependency on a developer's machine, it runs the attacker's payload. Tested against nine AI coding assistants, the technique could let attackers pool compromised developer machines into a botnet for crypto mining, denial-of-service attacks, or ransomware.

Check
Review whether developers or AI coding assistants automatically install and run packages the AI suggests, and check that dependencies are verified against known-good sources before installation.
Affected
Developers using AI coding assistants that fetch and run packages; if the assistant hallucinates a package name an attacker has pre-registered, the developer's machine runs attacker code and can join a botnet.
Fix
Do not let AI coding agents install or run fetched packages without human verification, confirm each suggested dependency actually exists and is legitimate, pin known-good dependencies, and isolate agent execution from credentials.

Attackers phone Microsoft 365 users to walk them through fake passkey setup

Okta warns of a campaign that phones Microsoft 365 users and talks them through what looks like setting up a passkey, but is actually a phishing kit that hands their account to the attacker. Active since April, the operators register passkey-themed domains and call targets, exploiting unfamiliarity with how passkeys really work. The kit mimics Microsoft's passkey enrollment without registering a real passkey, and pushes the victim to "save a recovery key" that the attacker controls, capturing the access needed to take over the account. The campaign, aimed at extortion, notably targets the passkey adoption process itself, turning a security upgrade into a social-engineering opening.

Check
Tell staff that Microsoft passkey setup happens through a device system prompt, not a phone call or web form, and that anyone calling to walk them through passkey registration is suspicious.
Affected
Microsoft 365 users unfamiliar with passkey enrollment; a convincing phone call plus a look-alike registration page can trick them into handing over the access an attacker needs to take over the account.
Fix
Train users on the genuine passkey enrollment flow, restrict who can register new authentication methods and recovery keys, monitor Entra for unexpected authentication-method changes, and verify unsolicited passkey calls internally.

Fake payment SDKs on npm and PyPI steal developer and cloud credentials

Socket found a coordinated campaign of 17 malicious packages on npm and PyPI that impersonate the payment SDKs for Paysafe, Skrill, and Neteller to steal developer and cloud credentials. The fake libraries mimic real SDKs, returning success responses without contacting the real services, while quietly harvesting environment secrets such as Paysafe API keys, AWS keys, and GitHub and npm tokens and sending them to a server on AWS. Because these packages get pulled into build pipelines, a single one reaching a continuous integration runner can hand over the broad credentials those runners hold. The npm versions trigger when a payment key is present; the PyPI versions run on install regardless.

Check
Audit your dependency trees and continuous integration logs for the named fake Paysafe, Skrill, and Neteller packages, and search for the PAYSAFE_API_KEY variable appearing alongside unfamiliar package names.
Affected
Developers and CI/CD pipelines integrating Paysafe, Skrill, or Neteller that pulled the malicious packages; the fake SDKs steal payment API keys, AWS credentials, and GitHub and npm tokens from build environments.
Fix
Remove any affected packages, rotate exposed API keys, cloud credentials, and registry tokens, pin and verify dependencies against official SDK names, and limit the credentials and permissions available inside CI/CD runners.

A single GitHub issue can trick AI workflows into leaking private repository data

Researchers at Noma Labs showed that GitHub's new Agentic Workflows, which let an AI agent read issues and act on repositories automatically, can be tricked into leaking private code through nothing more than a public issue. The technique, GitLost, is indirect prompt injection: an attacker opens an ordinary-looking issue in an organization's public repository, buries plain-English instructions in it, and the agent, which often holds a token with read access across the org's repositories, follows them, fetches files from a private repo, and posts the contents in a public comment. No credentials, coding, or write access are needed. GitHub was notified, but researchers frame it as an architectural weakness.

Check
Review any GitHub Agentic Workflows or similar AI automations, and check whether their agents hold organization-wide or cross-repository read tokens while also processing untrusted content like public issues and pull requests.
Affected
Organizations using GitHub Agentic Workflows where the agent has broad cross-repository read access and reacts to public issues; an attacker can steer it into leaking private repository contents with one crafted issue.
Fix
Scope agent tokens to only the repositories a workflow needs, keep agents reading untrusted input away from private data, restrict which events trigger automation, and require human review before agents post publicly.

RedWing rents out ready-made Android banking malware through a Telegram bot

Zimperium found RedWing, an Android bank-fraud operation rented out on Telegram as a finished product, complete with subscription tiers, guides, and a bot that builds each buyer a custom malicious app on demand, so no coding skill is needed. It spreads through phishing links leading to fake app-store pages that convincingly imitate Google Play and other stores. Once installed and granted permissions, it overlays fake login screens on real banking and crypto apps, reads incoming texts and screen content to capture one-time codes, and can silently forward the victim's calls to defeat phone-based verification. It also offers live screen control, keylogging, and camera access.

Check
Remind users to install apps only from official stores, distrust app updates arriving by link or text, and never grant Accessibility or default-texting access to an app without a clear reason.
Affected
Android users who sideload apps and approve broad permissions; RedWing then overlays fake login screens, steals one-time codes from texts and the screen, and forwards calls to defeat phone-based verification.
Fix
Keep installs restricted to official app stores, avoid enabling unknown sources, review and limit Accessibility and default-messaging permissions, use app-based rather than SMS authentication, and deploy mobile threat defense on banking devices.

Fake job interviews from big brands phish marketing staff for Google accounts

A phishing campaign is impersonating more than 30 well-known brands, including Adobe, Netflix, Coca-Cola, and OpenAI, in fake job-interview lures aimed at marketing professionals, with the goal of stealing their Google account credentials. Posing as recruiters from desirable companies, the attackers draw targets into a process that leads to a convincing Google sign-in page under their control. Because marketers often manage valuable brand, advertising, and analytics accounts tied to Google, a stolen login can open the door to ad fraud, data access, and further impersonation. The lure works by exploiting excitement about a career opportunity to lower the target's guard.

Check
Warn staff, especially in marketing, that recruiters asking them to sign in with Google to view interview or job details may be phishing, and verify any unexpected job outreach through official channels.
Affected
Marketing and other professionals targeted by fake job offers from impersonated big brands; stolen Google credentials can expose advertising, analytics, and brand accounts and enable further fraud and impersonation.
Fix
Use phishing-resistant multi-factor authentication like passkeys on Google accounts, verify recruiter outreach independently, never enter credentials on pages reached through unsolicited links, and check the domain before signing in anywhere.

SkillCloak hides malicious AI agent skills from the scanners meant to catch them

Researchers at the Hong Kong University of Science and Technology showed that the static scanners meant to vet add-on "skills" for AI coding agents like Claude Code can be fooled while the malware keeps working. Their technique, SkillCloak, either rewrites the tell-tale bytes a scanner looks for, using look-alike characters and line breaks, or hides the whole payload in a directory scanners skip, such as .git, behind a decoder that rebuilds it only when the agent runs the skill. Across eight scanners and 1,613 real malicious skills, the packing trick evaded detection more than 90 percent of the time. The researchers argue static scanning is not enough and released a runtime checker.

Check
If you scan AI agent skills, test whether your scanner inspects ignored directories like .git and normalizes look-alike characters, and add runtime behavior monitoring rather than relying on static analysis alone.
Affected
Anyone installing third-party skills for AI coding agents; these run with the agent's permissions, and static scanners, even strong ones, miss most cloaked malicious skills while the payload still executes.
Fix
Do not rely on static scanning alone for AI agent skills, add runtime monitoring of what skills do, vet skill sources strictly, restrict skill permissions, and treat marketplace skills as untrusted code.

Hidden website prompts trick AI agents into sending cryptocurrency to attackers

Zscaler found attackers using search-engine poisoning and hidden instructions on malicious websites to manipulate AI agents into making cryptocurrency payments. In one case, a hidden element on the page tells an AI agent that it must "resolve an error" by completing a payment, alongside code that starts a crypto transfer to a hardcoded wallet; the same page also shows human visitors ordinary payment options. Another campaign typosquats a decentralized-finance portfolio tracker and uses hidden prompts to convince agents the fake site is the real one. The attacker is seeding the scheme through several GitHub repositories, showing how autonomous agents that browse and act can be steered by content they read.

Check
Review any AI agents that can browse the web and initiate payments, and check what guardrails, approvals, or spending limits sit between an agent reading a web page and moving money.
Affected
Users and organizations running autonomous AI agents that can browse and make payments; hidden instructions on poisoned or typosquatted websites can redirect the agent into sending cryptocurrency to an attacker's wallet.
Fix
Require human approval for agent-initiated payments, isolate untrusted web content from action-taking, apply spending limits and allowlisted destinations, and treat everything an agent reads online as untrusted rather than as instructions.