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

DeepSeek AI agent tool flaw lets an agent disable its own sandbox

Researchers at VulnCheck found a flaw in the DeepSeek Harness, a tool that runs an AI agent's commands inside an operating-system sandbox so an agent handling untrusted files cannot write outside its workspace. Through an authentication bypass using a spoofed host header, an attacker needing no credentials or API key can call the tool's own web interface to invoke privileged commands with full-access permissions, raise the session's approval policy to unrestricted execution, and read every stored conversation. In effect, the sandbox meant to contain the agent can be switched off from outside. It is a reminder that an AI agent's isolation is only as strong as the authentication protecting its control interface.

Check
If you run the DeepSeek Harness or similar agent-sandboxing tools, restrict and authenticate access to their control interfaces, keep them off untrusted networks, and apply vendor fixes for the host-header authentication bypass.
Affected
Deployments using the DeepSeek Harness to sandbox AI agents; an unauthenticated attacker who reaches its control interface can spoof the host header to escalate to full-access command execution and dump conversations.
Fix
Authenticate and lock down agent-sandbox control planes, never expose them to untrusted networks, validate host headers, patch the flaw, and design agent isolation assuming the control interface itself is a target.

Attackers plant a stealthy Linux rootkit on F5 BIG-IP access gateways

Researchers at Sophos and ESET found attackers breaching F5 BIG-IP APM access gateways and installing a stealthy Linux rootkit that ESET calls PoisonedRefresh. Rather than dropping a file on disk, it hides a web shell in memory, hooks into the Apache and PHP components, tampers with SELinux settings, and uses disguised requests to run commands while blending into normal traffic. Crucially, it persists across device upgrades, so patching the appliance alone does not remove it. The intrusions likely began by exploiting a critical remote code execution flaw that F5 had earlier downgraded to a mere denial-of-service issue, which may have led some organizations to deprioritize patching it.

Check
Treat internet-facing F5 BIG-IP APM devices as potentially compromised, patch the underlying remote code execution flaw, and hunt for in-memory web shells, Apache and PHP hooks, and altered SELinux settings.
Affected
Organizations running F5 BIG-IP APM access gateways, especially internet-facing ones on the vulnerable version; attackers install a memory-resident rootkit that survives upgrades and turns the gateway into a persistent, covert foothold.
Fix
Patch the F5 flaw, but because the rootkit survives upgrades, inspect and rebuild affected devices from known-good images, restrict management exposure, rotate credentials the gateway handled, and re-evaluate vendor DoS-only ratings.

Researchers build a zero-click WeChat worm that spreads through voice calls

Security researchers built a zero-click worm that hijacks WeChat accounts through an incoming voice call on both iPhone and Android, without the target ever answering. The exploit fires during the ringing phase, before the user declines or picks up, abusing a memory-corruption flaw in WeChat's call-handling code to run commands on the device and take over the account. Because it can spread from a compromised contact to their contacts, it behaves like a worm. Notably, the researchers used AI to find the bug and write the exploit in about two days. Tencent patched it in late August and added a server-side mitigation, and saw no in-the-wild abuse before the fix.

Check
Make sure WeChat is updated to the patched version on all devices, since the fix landed in late August, and treat messaging apps with call features as a real remote attack surface.
Affected
WeChat users on iPhone and Android not updated before the late-August patch; a malicious incoming call could take over the account with no interaction, and the worm could spread to their contacts.
Fix
Keep messaging and calling apps updated promptly, prioritize patches for zero-click and call-handling flaws, and recognize that AI is shortening the time between a bug and a working exploit.

Attackers use autonomous AI agents to steal thousands of credentials in hours

Google's threat-intelligence team reported that a financially motivated attacker used an autonomous, multi-agent AI framework to compromise thousands of third-party credentials in under six hours, a pace that would take a human far longer. It is part of a broader shift in which criminals fold autonomous and coding-focused AI agents into operations, using them to get past defenses, reverse-engineer software, and sift stolen data. Google also found infostealer malware now specifically targeting AI developer configurations and credentials, and actors running open-weight models on compromised machines to sidestep restrictions on commercial services. The takeaway is that AI is collapsing attack timelines from days to hours, and AI access itself is now worth stealing.

Check
Assume attackers can now move at machine speed, and shorten detection and response for credential abuse: enforce phishing-resistant authentication, monitor authentication closely, and protect AI developer keys and configurations as sensitive credentials.
Affected
Organizations exposed to large-scale automated credential attacks; autonomous AI agents can test and abuse stolen credentials across many services in hours, and AI keys and developer configurations are now specific theft targets.
Fix
Adopt phishing-resistant, device-bound authentication, monitor for rapid credential-testing and anomalous automation, secure and rotate AI provider keys and developer configs, and assume the window between a leak and its abuse is shrinking.

PEEP toolkit hijacks Chrome and Edge into backdoors that run host commands

Researchers disclosed a post-exploitation toolkit called PEEP that turns Chrome and Edge into backdoors for stealing credentials and running commands on the host. After an attacker already has administrative or code-execution access, PEEP's installer injects a malicious extension disguised as a bookmarks tool directly into browser profiles, forging Chromium's own integrity settings to bypass the Web Store and skip user approval prompts. The planted extension then serves as a covert channel to harvest credentials and execute operating-system commands. Because it abuses a trusted, ever-present browser and hides its extension from the usual checks, it can persist quietly on a compromised machine, a reminder that browsers themselves are a rich post-compromise attack surface.

Check
Hunt for tampering with Chromium's secure preferences file and for extensions loaded outside the Web Store, and treat an unexpected browser extension on a server or admin workstation as a possible backdoor.
Affected
Windows systems an attacker already compromised with admin or code-execution access; PEEP silently installs a browser extension by forging integrity settings, then uses it to steal credentials and run host commands.
Fix
Enforce enterprise extension allowlisting through browser policy, monitor for secure-preferences tampering and unapproved extensions, restrict local administrator rights that enable the install, and include browser artifacts in endpoint detection and incident response.

BigBear phishing service bypassed MFA at 258 organizations and disabled passkeys

Researchers at CloudSEK gained access to the control panel of BigBear, a phishing-as-a-service platform that defeated multi-factor authentication at 258 organizations and stole thousands of Microsoft 365 credentials. It uses an adversary-in-the-middle proxy based on Evilginx to sit between victims and Microsoft's real login, capturing passwords, multi-factor codes, and the session cookie, then replaying the cookie to hijack the already-authenticated session. The panel logged over 5,000 stolen records across 40-plus countries and is rented to multiple affiliates who receive stolen data through Telegram bots. Notably, it runs JavaScript that disables the browser's passkey support, forcing victims off phishing-resistant login onto weaker methods it can intercept.

Check
Enforce phishing-resistant passkeys with conditional access that requires managed devices, so attacker-in-the-middle kits cannot simply capture and replay session cookies or quietly downgrade users to weaker authentication.
Affected
Microsoft 365 organizations relying on passwords plus standard multi-factor authentication; BigBear steals the post-login session cookie to hijack accounts and can disable passkey support in the browser to force weaker login methods.
Fix
Adopt device-bound phishing-resistant authentication, require managed devices through conditional access, revoke sessions and refresh tokens on suspicion, monitor for impossible-travel and residential-proxy sign-ins, and train users on help-desk and login-page lures.

JSCeal malware steals session cookies to log into Google without the password

Researchers at Check Point analyzed JSCeal, malware compiled into a hard-to-analyze bytecode format that steals browser cookies and authentication tokens to hijack accounts. By replaying stolen session cookies, an attacker can access a victim's Google account without the password or a second factor, and the malware also grabs saved passwords, autofill data, and OAuth tokens to automate further account access. It additionally targets cryptocurrency wallets and platforms and includes keylogging, screenshots, and messaging-session theft. The compiled format and layered obfuscation push it outside analysts' usual tooling, though Check Point released a deobfuscator. It is a reminder that stolen session cookies quietly defeat passwords and multi-factor authentication alike.

Check
Treat session cookies as sensitive credentials: monitor endpoints for access to browser cookie databases and suspicious script-to-runtime execution chains, and shorten session lifetimes so stolen cookies expire sooner.
Affected
Users whose browsers are infected by this stealer; theft of session cookies and OAuth tokens lets attackers replay authenticated Google sessions without the password or second factor, and reach crypto accounts.
Fix
Bind sessions to devices where supported, expire and revoke sessions on anomalies, deploy endpoint detection for cookie theft and in-memory browser attacks, and monitor for session replay from unfamiliar locations.

REVSTEALER modules disable Windows Update and Defender to hide a crypto miner

Researchers at Elastic documented four persistent programs tied to the REVSTEALER infostealer that stay on a machine even after the stealer deletes itself. One disables Windows Update services and Microsoft Defender, adds Defender exclusions, and kills update and malware-removal tasks before hiding a cryptocurrency miner inside legitimate Windows processes. The malware also bypasses Chrome's app-bound encryption by launching the browser in a debugger to read the decryption key from memory, and steals session cookies to take over accounts without passwords. It spreads through game-cheat lures on hijacked video channels and pirated or fake application installers. Because these modules outlive the stealer, a confirmed infection warrants reimaging rather than cleanup.

Check
Treat any REVSTEALER or infostealer detection as an incident and reimage the machine, since companion modules persist after the stealer removes itself, and watch for disabled Defender and high CPU usage.
Affected
Windows users who run game cheats or pirated and fake software; the modules disable protection, mine cryptocurrency, bypass Chrome's encryption to steal cookies, and persist after the stealer deletes itself.
Fix
Restrict local administrator rights so malware cannot disable Update and Defender, block game-cheat and pirated-software sources, monitor for security-tool tampering and mining activity, and reimage confirmed infections rather than deleting individual files.

Shai-Hulud npm worm now hunts credentials across 469 different locations

Researchers at GitGuardian found that a recent variant of the self-spreading Shai-Hulud npm worm has expanded its credential theft to scan 469 distinct locations on infected developer machines. The targets now span developer environments, continuous integration and deployment tooling, cloud configuration files, and even the configuration of AI tools. That breadth turns a single compromised package into a wide net for secrets, from cloud and registry credentials to keys held by developer and AI tooling. It reflects how supply-chain worms are industrializing secret collection, treating any credential a developer's machine can reach as fair game once malicious code runs during installation or use.

Check
Scan your dependencies and lockfiles for known-compromised packages, rotate any credentials that a developer machine or pipeline can reach, and reduce the number of long-lived secrets stored in reachable configuration files.
Affected
Developers and CI/CD systems that install compromised npm packages; the worm harvests credentials from 469 locations across developer, pipeline, cloud, and AI-tool configurations, then uses them to spread and steal further secrets.
Fix
Pin and vet dependencies, use scoped short-lived tokens instead of long-lived secrets, isolate build environments, monitor for credential access during installs, and keep secrets out of files developer and AI tools read.

Attackers abuse the trusted Node.js runtime to run malware past defenses

Symantec reported that threat actors are abusing the legitimate, digitally signed Node.js runtime to run malicious JavaScript while slipping past security tools, in attacks on government, technology, and hospitality targets since February. Because the Node.js executable is a trusted developer tool, defenses rarely flag it, so instead of dropping a malicious program the attackers stage the genuine runtime and keep their harmful logic in interpreted scripts. They gain persistence through a Windows registry startup key and, in one case, pulled command-and-control instructions from the blockchain using a technique called EtherHiding. The activity has been tied to a ClickFix social-engineering entry point and an initial-access broker.

Check
Hunt for unexpected Node.js installations on machines that should not run developer tools, suspicious registry startup entries invoking the runtime, and outbound traffic to blockchain endpoints used for command and control.
Affected
Windows environments where a signed Node.js runtime can be introduced and run scripts unnoticed; attackers use it to execute malicious JavaScript, persist through registry keys, and evade tools that trust the binary.
Fix
Apply application control to restrict where the Node.js runtime may run, alert on its use outside development, monitor script execution and registry run-key changes, and block known blockchain command-and-control and ClickFix infrastructure.