Splunk has patched a critical flaw in Splunk Enterprise that lets an unauthenticated attacker run code on the server, a serious risk given Splunk often sits at the heart of a company's security monitoring. The bug (CVE-2026-20253, rated 9.8) is in the PostgreSQL sidecar service added in Splunk 10, whose internal API has no authentication yet is reachable through the main web app's proxy. An attacker can write or overwrite files on the host and chain that into remote code execution. The sidecar is off by default on on-premises Windows but enabled out of the box on Splunk Enterprise running in AWS. Splunk Cloud is not affected.
Researchers at Tenet Security have disclosed Agentjacking, a new attack that turns AI coding assistants like Claude Code, Cursor, and Codex into tools for running an attacker's code on a developer's machine. The trick abuses Sentry, a widely used error-tracking service: anyone can submit a fake error event using a project's DSN, a public write-only key embedded in website code, and the AI agent, fetching that event through Sentry's MCP integration, cannot tell the malicious instructions from real diagnostics and runs them with the developer's privileges. No phishing, malware, or server breach is needed, and it bypasses traditional controls because every step is technically authorized. Tenet found 2,388 exposed organizations.
Attackers hijacked more than 400 packages in the Arch User Repository (AUR), the community add-on store for Arch Linux, in a supply-chain attack dubbed Atomic Arch. Rather than exploiting a flaw, they adopted abandoned packages and quietly edited the build recipe (PKGBUILD) to pull in a malicious npm package, atomic-lockfile, at install time. The payload is a Rust credential stealer that grabs browser logins, SSH keys, crypto wallets, and developer tokens; when run as root it also loads an eBPF rootkit that hides its processes, files, and network connections. Only the AUR is affected, not Arch's official repositories. The package names and histories looked completely normal.
Sygnia has detailed Operation Highland, a campaign in which the China-linked group Velvet Ant hid inside the Linux authentication stack itself for close to a decade, with traces back to 2016. Instead of dropping detectable malware, the attackers replaced the trusted PAM login module (pam_unix.so) and OpenSSH binaries with backdoored versions, found in nine distinct variants. Some accepted a hardcoded secret password; others silently logged real usernames, passwords, and every command typed, with a hidden switch to turn logging off. Because login programs are trusted and rarely inspected, the activity looked like normal administration and evaded scanners on a network with no direct internet access.
A critical flaw in phpBB, the open-source forum software running on thousands of sites, lets an unauthenticated attacker obtain a valid login session as any user, including an administrator, with a single HTTP request. The bug (CVE-2026-48611, rated 9.4) works in the default configuration and traces back to code from 2014. An admin session gives full read, write, and delete access to the forum and, on the latest branch, opens a path to remote code execution and full server takeover. A second, lower-severity flaw affecting only OAuth-configured installs was also fixed. phpBB released version 3.3.17 to patch both.
Check Point has disclosed three now-patched flaws in LangGraph, the popular LangChain framework for building AI agents, that can be chained for remote code execution on self-hosted servers. The chain combines an SQL injection (CVE-2025-67644) with an unsafe msgpack deserialization bug (CVE-2026-28277): an attacker who can reach the agent's stored-state endpoint plants a malicious checkpoint that runs code when loaded. A compromised LangGraph server exposes everything the agent can touch, including model API keys, customer data, and internal network access. It is only exploitable in self-hosted deployments using the SQLite or Redis checkpointer; LangChain's managed LangSmith platform is not affected.
The Iran-linked group Handala claims it breached California Water Service (Cal Water), one of the largest US investor-owned water utilities, and published a 5GB sample to prove it. Analysts say the attackers reached a customer billing database holding personal data (names, addresses, account and payment details) and an internal GPS-correction server, leaking administrative credentials in the process. Handala framed the attack as retaliation for US actions against Iran and boasted it could disrupt water supply, but researchers stress the evidence does not support that claim, neither system controls water treatment, and the group is known to exaggerate. Cal Water has not yet publicly confirmed the incident.
Novo Nordisk, the pharmaceutical giant behind Wegovy and Ozempic, has disclosed that attackers copied data from its internal IT systems, including information on patients in some of its clinical trials. The company stressed the patient data was de-identified, containing fields like patient ID, year of birth, sex, biomarkers, and lifestyle factors rather than names or direct identifiers. Novo has not said how many people are affected or named the attacker, and is not offering credit monitoring, instead advising patients and healthcare professionals to stay alert for unexpected messages or calls. Pharma firms are increasingly targeted for their valuable research and patient data.
France's government messaging platform Tchap, the in-house, Matrix-based app that civil servants are required to use instead of WhatsApp or Signal, was breached after a threat actor hijacked a single user account, no software exploit needed. The cyber agency ANSSI detected it on June 7. Officials say data tied to about 73,000 accounts, roughly 9 percent of users, was exposed: the attacker scraped everything shared in public chat rooms, which are not encrypted, while private end-to-end conversations stayed protected. The haul includes over 13.5GB of documents and media plus hardcoded LDAP credentials leaked in a PowerShell script. Entry was via the education ministry's server.
Google has filed suit against a Chinese cybercrime network it says abused its Gemini AI to mass-produce phishing text messages and fake websites targeting Americans. The group runs a phishing-as-a-service kit called Outsider and used Gemini to generate fraudulent pages and large smishing campaigns. The texts impersonate trusted brands, warning of "brokerage account issues" or dangling carrier "rewards," and link to lookalike sites that harvest personal and financial details. Google says the lawsuit aims to dismantle the network's infrastructure. The case underscores how criminals are folding mainstream AI tools into industrialized phishing operations.
TrustStrike Labs