Anthropic appears to be preparing the public rollout of Claude Mythos - the restricted security-focused frontier model that uncovered 10,000 high or critical vulnerabilities in its first month under Project Glasswing. References to 'claude-mythos-1-preview' have briefly appeared in the public Claude Code and Claude Security products, with at least one user reportedly seeing a toggle to enable Mythos before it was pulled. Anthropic originally announced Mythos in early preview on April 7 and held back the public release pending guardrails, warning the model 'can automatically develop functional cyberattacks at a highly professional level.' Pricing and tier availability are not yet disclosed.
NCC Group's Fox-IT has documented RemotePE, a previously private cross-platform RAT used by the North Korea-linked Lazarus Group against DeFi, financial, and cryptocurrency organizations. The chain starts with social engineering on Telegram (impersonating a trading-firm employee with fake Calendly and Picktime meeting links), then drops DPAPILoader (Iassvc.dll) which uses Windows DPAPI to decrypt RemotePELoader. That loader fetches RemotePE entirely in memory from aes-secure[.]net, evading EDR via Hell's Gate and ETW patching. RemotePE itself is a C++ RAT supporting six command categories. Fox-IT believes the toolset is reserved for high-value, long-dwell access leading to large-scale financial theft. Activity dates from mid-2023.
Socket has detailed TrapDoor, a coordinated cross-ecosystem supply-chain campaign that has published 34+ malicious packages across 384+ versions on npm, PyPI, and Crates.io since May 22. Targets are crypto, DeFi, Solana, and AI developers. The npm packages deploy trap-core.js, which scans for credentials, validates AWS and GitHub tokens via API, and persists via cron, systemd, Git hooks, shell rcfiles, and SSH; Rust crates use build.rs to trigger; Python packages auto-execute on import to fetch JavaScript from ddjidd564.github[.]io. Notable twist: the campaign also plants .cursorrules and CLAUDE.md in PRs to popular AI repos to trick AI coding assistants into running 'security scans' that exfiltrate secrets.
FBI Director Kash Patel's merchandise website basedapparel[.]com was taken offline on Friday after researchers documented a multi-stage WooCommerce compromise that stole payment data and targeted Mac users with a ClickFix attack. The site displayed a fake Cloudflare CAPTCHA prompting visitors to paste a command into their terminal; the macOS-specific shell command then downloaded a script-based infostealer that targets browsers, password vaults, and cryptocurrency wallets before compressing the data, exfiltrating to monterushy[.]com, and deleting itself. Researchers WifiRumHam and 'debbie' analyzed the live campaign on May 21-22; the site went offline on May 22. Similar infections seen across many compromised WooCommerce sites.
Qianxin XLab has documented a large-scale ClickFix campaign exploiting CVE-2026-26980, an SQL injection in Ghost CMS that was disclosed and patched on February 19. The vulnerability lets unauthenticated attackers read arbitrary database content including admin API keys, which are then used to inject malicious JavaScript into articles. More than 700 domains are confirmed compromised, including Harvard, Oxford, and Auburn universities and DuckDuckGo. Victim browsers receive a fingerprinted iframe overlay impersonating a Cloudflare prompt that instructs users to paste a command into the Windows command prompt, dropping DLL loaders, JS droppers, or the UtilifySetup.exe Electron-based payload. Two distinct activity clusters compete for compromised sites.
Anthropic has named the program behind its Claude Mythos Preview model 'Project Glasswing' and disclosed the first-month results. Working with AWS, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks, the program flagged 6,202 high or critical vulnerability candidates across 1,000+ open-source projects; 1,726 were validated by human reviewers and 1,094 confirmed as genuine high or critical severity. A WolfSSL certificate-forgery flaw (CVE-2026-5194, CVSS 9.1) is the named-and-shamed example. 97 upstream patches and 88 security advisories have landed. Anthropic itself warns that finding flaws is far easier than fixing them.
Socket has detailed a coordinated supply-chain campaign that planted malicious code in eight Composer packages on Packagist, including moritz-sauer-13/silverstripe-cms-theme, crosiersource/crosierlib-base, devdojo/wave, devdojo/genesis, katanaui/katana, elitedevsquad/sidecar-laravel, r2luna/brain, and baskarcm/tzi-chat-ui. The attackers placed the payload not in composer.json but in package.json - meaning teams scanning only PHP dependencies would miss the Node.js lifecycle hook bundled inside. The postinstall script downloads a Linux binary from a GitHub Releases URL (github[.]com/parikhpreyash4/systemd-network-helper-aa5c751f), saves it as /tmp/.sshd, and runs it backgrounded with execute permissions. Socket found the same payload referenced in 777 GitHub files, including two GitHub Actions workflows - hinting at a broader campaign.
GitHub has shipped npm CLI 11.15.0 introducing a 'staging' workflow that lets maintainers run 'npm stage publish' to push a candidate to a staging area before going live - with the constraint that the package must already exist on the registry and have 2FA enabled on the account. Three new install flags (--allow-file, --allow-remote, --allow-directory) extend the existing --allow-git to give developers an explicit allowlist for every non-registry install source. GitHub is also encouraging maintainers to pair staging with trusted publishing via OIDC. The changes respond to the TeamPCP supply-chain wave that compromised hundreds of packages over the past several weeks.
The FBI has issued a warning about Kali365, a phishing-as-a-service platform that fueled large Microsoft 365 attacks in April. Instead of stealing passwords, Kali365 customers trigger Microsoft device-login requests and trick victims into completing the authorization, capturing OAuth access and refresh tokens that grant immediate mailbox access. Arctic Wolf, which infiltrated the system, says Kali365 sells in three tiers from $250 for 30 days to $2,000 for the year and generates branded phishing lures impersonating Adobe, DocuSign, and SharePoint in dozens of languages. Threat actors set malicious inbox rules to suppress security notifications and extend dwell time.
Italian Guardia di Finanza has dismantled CINEMAGOAL, an unusual piracy operation whose customers installed an app on their devices that authenticated directly to legitimate Netflix, Disney+, Spotify, Sky, and DAZN. A network of virtual machines in Italy captured fresh authentication and decryption codes from real subscriptions (opened under false identities) every three minutes and redistributed them to subscribers, who streamed at full quality with their real IPs masked. Operation 'Tutto Chiaro' executed 100 searches across Italy, seized servers in France and Germany, and identified about 70 resellers. The first 1,000 subscribers have been fined between €154 and €5,000.
TrustStrike Labs