🚨 ACTIVE EXPLOITATION
UPDATE: Attackers chain three Ivanti Connect Secure zero-days to gain root on fully patched VPN appliances
Threat actor UTA0576 chained three Ivanti Connect Secure zero-days to gain unauthenticated root access on fully patched VPN appliances.- Affects Ivanti Connect Secure VPN appliances; a fully patched appliance running 22.8R2.4 (Build 19097) was compromised, with attacks confirmed against a limited number of customers.
- The chain exploits CVE-2026-104935 (CVSS 8.6, authentication bypass in welcome.cgi), CVE-2026-111357 (CVSS 8.8, privilege escalation via dsinvoked trusting DSCache data), and CVE-2026-118527 (CVSS 4.3, predictable filename in wu.cgi).
- The attacker bypasses authentication via an X-DSCGI-PATH header mismatch, uploads a Perl script through wu.cgi and races its deletion for code execution, then patches library code in memory to run commands as root.
- UTA0576 stole stored domain credentials, Windows LAPS admin passwords, BitLocker recovery keys, and LDAP traffic, and signaled preparation for lateral movement; activity was first confirmed October 1, 2026.
- The attack leaves no disk artifacts — no backdoor or web shell — evading Ivanti's Integrity Checker, and fixes shipped in releases 25.1.1.3 and 22.8R2.6 on October 29, 2026.
Drupal warns of actively exploited Commerce Core access bypass flaw CVE-2026-118539
CVE-2026-118539
Drupal has warned that attackers are actively exploiting a critical access bypass vulnerability in its Commerce Core module.- Affects sites using the Drupal Commerce Core e-commerce module, used by more than 42,000 websites including Fortune 500 companies and research universities.
- CVE-2026-118539 is a missing authorization flaw allowing unauthenticated attackers to bypass access protections on certain routes and entity fields, potentially causing financial impact to vendors.
- Affected versions: Commerce Core 0.0.0 through 2.41 and 3.0.0 through 3.3.13; fixed releases are 3.3.13 (3.x branch) and 8.x-2.41 (earlier versions).
- Drupal tagged the advisory as highly critical and confirmed the vulnerability is being actively exploited in the wild, urging admins to audit orders for inconsistencies.
Sources 📄 Source: drupal.org · 📎 Coverage: bleepingcomputer.com · 👁 via BleepingComputer, The Hacker News
💥 BREACHES & INCIDENTS
Trojanized Claude Code Extension on Open VSX Plants Persistent SSH Backdoor on Developer Systems
A trojanized anthropic.claude-code extension (v2.1.313) on Open VSX installs an SSH backdoor giving attackers persistent remote access to developer machines.- Affects developers who installed the tampered anthropic.claude-code package version 2.1.313 from Open VSX; the analyzed sample targeted macOS on Apple Silicon with injected Linux and Windows code branches.
- A 66-line JavaScript loader prepended to extension.js adds an attacker-controlled Ed25519 SSH key (comment svc-telemetry, SHA256:0Z6R+DF64Do2nwQic8a7kkfDLNNk4IiK+2U2Rci4htk) to ~/.ssh/authorized_keys and enables Remote Login/SSH services.
- The implant opens a reverse SSH tunnel to tunnel@134.209.82.225 (forwarding localhost:22) with auto-reconnect every 60 seconds, using autossh or ssh.
- Persistence uses a macOS LaunchAgent (com.apple.sysmond.plist), ~/.config/.smt and .smtk files, an @reboot cron entry on Linux, and a SysMon scheduled task or sysmon.cmd startup script on Windows.
- Collected hostname, username, OS, architecture, Git email, and up to 40 environment variable names are sent via HTTP beacon to http://134.209.82.225:8099/beacon and http://127.0.0.1:8099/beacon; malicious extension.js SHA-256 is ecb58cb29078d9646ad1dcdbe51526f2d9cfea06be5fb2ab183e8ecc3ab6bf15.
Sources 📄 Source: opensourcemalware.com · 📎 Coverage: cybersecuritynews.com · 👁 via Cyber Security NewsPentagon treats August freezer failures at US military bases as suspected cyberattack
The Pentagon is investigating freezer malfunctions that spoiled food at US military base commissaries as a suspected cyberattack.- At least six on-base commissaries were hit in late August, including Fort Huachuca, Arizona, where a freezer went into defrost mode and heated up all the food
- A Cyber Command unit and the Army Criminal Investigation Division's Cyber Field Office are investigating with law enforcement and French counterparts
- The case mirrors an October French cyberattack that overrode temperature controls on commercial cooling units at more than 350 supermarkets, causing food losses and fires
- The French attack exploited Danfoss controllers that were obsolete or running outdated firmware; whether Danfoss units were used at US bases is unconfirmed
🕵️ RESEARCH & DEEP DIVES
- Unauthorized AI relay service disables OpenAI Codex and Claude Code sandboxes, reroutes enterprise data
Symantec has identified a grey market relay service that installs genuine OpenAI Codex and Claude Code CLIs and reconfigures them to route all prompts, files, and credentials through its own servers with sandboxes disabled.- A dozen organizations affected since early 2026, mostly in mainland China plus subsidiaries in South Korea, Taiwan, and Malaysia, across sectors including industrial automation, electronics, semiconductors, pharmaceuticals, shipping, petrochemicals, and data center services.
- The service is not malware or a vulnerability in Codex/Claude Code — it installs the genuine CLI tools via npm, then rewrites config files (~/.codex/config.toml, ~/.claude/settings.json) to point at relay-controlled domains.
- Delivery uses a fileless PowerShell or bash setup script fetched by a one-line command from lookalike domains such as vibe.aiok[.]me, which users appear to run themselves minutes after installing the CLI.
- The script sets sandbox mode to 'danger-full-access', enables network access for the write sandbox, disables telemetry and error reporting, and replaces user API credentials with relay-issued tokens that persist across reboots.
- The relay backend has rotated across at least eight domains impersonating OpenAI- or Anthropic-branded chat services; on one host it logged the user out of a legitimate Anthropic session and deleted their auth variables before reissuing relay credentials.
📋 ADVISORIES
Trezor warns of suspected hardware implant in Safe 5 wallet sold by CryptoBilis
Trezor has warned of a suspected hardware implant in a Trezor Safe 5 bought via reseller CryptoBilis that could steal recovery phrases and transmit them over a mobile network.- Affects Trezor devices bought from CryptoBilis (including stores it supplied) from March 2026 onward; one suspected implanted Safe 5 reported by a customer in Indonesia
- An extra circuit board sits between the main board and screen, containing a cellular module and antenna
- The implant allegedly reads information shown on the display, including the wallet backup/recovery phrase, and sends it over the mobile network while the device is plugged in
- Separately, hardware hacker Joe Grand found implants in Trezor One and Model T devices from unauthorized sellers, where replaced microcontrollers generated pre-determined seed phrases attackers could guess
- Trezor removed CryptoBilis from its authorized reseller network on October 9 and is working with Ledger, which confirmed physical tampering in some of its devices from the same reseller
Trezor warns of suspected hardware implant in Safe 5 that could steal recovery phrases
Trezor has warned that a suspected hardware implant in a Trezor Safe 5 sold via reseller CryptoBilis could steal recovery phrases.- Affects Trezor devices bought from reseller CryptoBilis (on Tokopedia) from March 2026 onward; one suspected Trezor Safe 5 was reported by a customer in Indonesia who bought it in August 2026.
- The implant reportedly is an extra circuit board between the main board and screen, containing a cellular module and antenna.
- It allegedly reads information shown on the display, including the wallet backup, and transmits it over a mobile network while the device is plugged in.
- A stolen recovery phrase lets an attacker restore the wallet elsewhere and control its funds; Trezor has not confirmed the implant through lab testing and reports no stolen funds from CryptoBilis buyers.
- The warning follows Ledger's confirmation of unauthorized hardware implants in CryptoBilis-supplied devices; Trezor removed CryptoBilis from its authorized reseller list.
📄 Source for ESET APT report: Russia-aligned hackers hit Ukraine; Gamaredon adopts Turla's Rezzuar tool — web-assets.esetstatic.com
📄 Source for Top 10 UEBA Tools in 2026 ranked: Exabeam leads list of nine platforms — cryptika.com
📄 Source for Supply chain attack via DebugBear script drains over $2 million from Polymarket traders — x.com
📄 Source for Belarus-aligned FrostyNeighbor APT escalates attacks on Ukrainian government with new tooling — eset.com
📄 Source for Researcher uses Claude AI chat to extract internal Anthropic binary and rebuild valid Google Cloud token — bishopfox.com