supply chain
21 articles with this tag
LiteLLM 1.107.0 Shipped a Credential Stealer (October 2026): Affected PyPI Release and Docker Images, and What Users Should Do
On October 21, 2026 (UTC), BerriAI, the company behind LiteLLM, announced that its release pipeline had been compromised and that a v1.107.0 containing malware that steals credentials and loads remote-code-execution tools had been published. Affected were litellm==1.107.0 on PyPI (11:39 to 13:23 UTC) and the v1.107.0, latest, main-latest and main-stable tags on Docker Hub (11:46 to 14:20 UTC) and GHCR (from 11:46 UTC until taken down by 13:51 UTC, about two hours). The official v1.107.0 source tag on GitHub does not contain the malicious files; the artifacts were built from an attacker-controlled lookalike repository. All earlier versions are safe, and the company asks users to pin 1.106.2 or another explicit earlier version. The company says GitHub CLI OAuth tokens belonging to three developers were stolen and used to disable branch protection and change the release workflow, and believes the tokens were taken from the logs of an internal LiteLLM deployment. How the attacker got into that deployment is still being investigated. This site's conclusion: rebuild any environment that installed or started 1.107.0 during the window, and rotate every key it could reach, including LLM provider keys, the proxy master key, the database password and cloud roles.
Rank Math PRO 3.0.123 Shipped with Malware (October 2026): Who Is Affected and What WordPress Site Owners Should Do
On October 16, 2026 (UTC), Rank Math announced that Rank Math PRO v3.0.123, the paid version of its WordPress plugin, contained malware for a limited number of users who downloaded it between 20:02 UTC on October 15 and 07:12 UTC on October 16. The malware could compromise a site if the plugin was activated, and could recreate its hidden backdoor whenever the site was visited. The company removed the affected files and released a clean v3.0.124. The free version distributed on WordPress.org is not affected. Neither the number of affected users nor how the malware got into the files has been disclosed. As of October 20, the last change to the notice was on October 19, when a note that the free version is not affected was added. This site's conclusion: treat any site that activated a copy of 3.0.123 obtained during the affected window as compromised, and either restore from a backup taken before the window or delete the whole plugin folder and clean up in the order the company gives. Either way, change every credential the site could read.
Japan's October 2026 warnings on unauthorized access (National Cybersecurity Office, IPA, JPCERT/CC): what to check this week, and what to do if you find something
In September and October 2026, organizations in Japan disclosed one data breach from unauthorized access after another. On 9 October the Cabinet Secretariat's National Cybersecurity Office and IPA issued warnings, following JPCERT/CC's alert of 8 October. IPA says the attacks have not been tied to a single product's vulnerability and appear to start from compromised internet-facing apps and services or accounts. This week: (1) list your internet-facing apps and review their logs for the last month, then three months; (2) check cloud, VPN and other external services for failed sign-ins, unfamiliar source addresses and accounts you don't recognise; (3) re-inventory the personal data you hold and minimise it. If anything looks wrong, declare an incident, preserve logs, bring in specialists and report to the Personal Information Protection Commission. This site's view: discovering that you only keep one month of logs is itself a finding to act on.
JR East, VIEW Card and JR Kyushu email address leak (eki-net about 1.67 million, VIEW's NET about 4.03 million, JR Kyushu Web members about 1.3 million): Otona no Kyujitsu Club also covers birth date and card expiry — what members should do
On October 9, 2026, East Japan Railway Company (JR East), VIEW Card Co., Ltd. and Kyushu Railway Company (JR Kyushu) each announced that the external service they use to send email to members had been affected by unauthorized access at a cloud service provider, and that they could not rule out that members' email addresses and other data had been viewed or obtained. The numbers are, at most, about 1.67 million eki-net members and about 390,000 Otona no Kyujitsu Club members (JR East); about 4.03 million email addresses of VIEW's NET users (VIEW Card); and, at most, about 1.3 million JR Kyushu Web members (JR Kyushu). For eki-net, VIEW's NET and JR Kyushu Web members only email addresses are involved; for Otona no Kyujitsu Club members, the member number, credit card expiry date and date of birth are involved as well. Names, addresses, phone numbers and card numbers are excluded. The cloud service provider has said that the outage which began early on October 7 was caused by a ransomware attack by a third party. In its fifth report on October 14, the provider said it had found no trace of data in virtual servers in the affected area being viewed or taken out, and concluded that no data leaked externally. On October 15, after receiving those findings, JR Kyushu announced that no leak of JR Kyushu Web members' email addresses had been confirmed. JR East's eki-net site also says that on October 15 the email delivery service company reported that it had concluded no personal data of eki-net members leaked (checked October 19). For Otona no Kyujitsu Club members and VIEW Card, this site could find no follow-up as of October 19 (checked on each company's site). This site's take: JR Kyushu said no leak was confirmed, and eki-net says it was told no personal data of eki-net members leaked. Even so, fake emails using this incident as a pretext may still arrive. Do not log in from links in emails; check through the official app or a bookmark you saved yourself.
Asahi Kasei Therapeutics healthcare-professional site breach (about 514,000 people): what may have leaked and what doctors, pharmacists and nurses should do
On October 6, 2026, Asahi Kasei Therapeutics Co. Ltd., the pharmaceutical arm of Japan's Asahi Kasei group, announced that Pharma DIGITAL, its information site for healthcare professionals, had been hit by a cyberattack and that personal data in its member database may have been viewed or taken. The data covers about 514,000 healthcare professionals (name, facility name, facility address, job type, specialty and more), about 44,000 of them with email addresses and more in addition to those items, and about 700 company employees (name, email address, photo). The company learned of the access from the site's contractor on October 2 and shut the site down. It says it did not store credit card data or sensitive personal data such as medical history, and as of October 6 no misuse had been found. This site's take: expect fake emails posing as drug companies, academic societies or seminar invitations. Never log in from email links, and change any password you reused from this site.
Daiwa Securities vendor breach (about 110,000 customers, 220,000 records): what may have leaked and what customers should do
On October 5, 2026, Daiwa Securities Co. Ltd., one of Japan's largest brokerages, announced that the server of a vendor whose service it uses to manage customers' online inquiries had been accessed without authorization. Names, email addresses, account numbers and other information about 110,000 customers, held in inquiry records, may have leaked; including inquiries that identify no one, about 220,000 records are involved. The access took place between about 8:33 p.m. on October 2 and 8:01 a.m. on October 3, and no unauthorized access to Daiwa's own systems was found. The company says this data cannot be used to access accounts or trade, including through online trading, and as of October 5 no fraudulent trades and no publication or spread of the data had been found. On October 6 the vendor side published its own notice: inquiry data stored in the environments of up to five client companies on the same server, up to 713,126 records counted per inquiry, may have been taken. Citing the same vendor, Citizen Watch (about 100,000 people), Sompo Japan (about 60,000 records counted per inquiry) and ORIX Life Insurance (about 26,500 records) have also announced that their customer data may have leaked. This site's take: expect fake messages that know your account number and your past inquiries. Never log in from links in emails or texts; sign in from a bookmark or the official app and check your message box and operation history there.
AdaptHealth data breach (about 4.1 million people): health insurance and health data exposed — patients should check benefit statements and calls
AdaptHealth, a US provider of home medical equipment, says a June 5, 2026 cyberattack exposed patients' names, contact information, demographic information, health insurance information and health information. Its report to the US Department of Health and Human Services (HHS) covers 4,115,802 people. According to the company, Social Security numbers (SSNs), financial account and card information were not involved, and it is not aware of any misuse. Its July 2 report to the Securities and Exchange Commission (SEC) says the cause was a social engineering attack (tricking people into giving information or access) that compromised a user session associated with a third-party contractor; a stored password file associated with insurance billing was taken and external electronic health record portals were accessed. Our view: with no SSNs, the priority is reading Explanation of Benefits statements for care or equipment you did not receive, and not giving insurance numbers to callers about supply reorders. For organizations, the lesson is how far one contractor account can reach.
Aesto Health Data Breach (9.5 Million): Why a Company You Never Heard of Sent the Notice, and What to Do
US healthcare data migration and archiving vendor Aesto, LLC (Aesto Health) says that from on or about December 2 to December 18, 2025, protected health information belonging to patients of its healthcare provider clients, stored in part of its cloud environment, may have been accessed and/or acquired by an unauthorized actor. The HHS Office for Civil Rights breach portal lists 9,540,683 people affected. The information includes names, dates of birth, medical information, driver's license numbers, financial account numbers, health insurance information, and government ID numbers such as ITINs; Social Security numbers were potentially involved for a limited number of people. You don't recognize the sender because your provider handed old records to an archiving vendor. First confirm whether your provider is on the list. Then protect SSNs and ITINs with a credit freeze and an IRS IP PIN, and medical information by reading your benefit statements. Providers need an inventory of every vendor holding patient data and an end date for archived records.
The TanStack, Nx Console and GitHub Compromise Chain (May 2026): What Developers Should Check
On May 11, 2026, the publishing pipeline of TanStack's Router/Start repository was abused to push credential-stealing code into 42 npm packages, 84 versions. An Nx contributor who installed one of them had a GitHub CLI token stolen, and on May 18 a tampered Nx Console extension (v18.95.0) was published. The same day, GitHub said an employee device had been compromised by a poisoned VS Code extension and GitHub-internal repositories were exfiltrated. In all three cases the attack went through legitimate accounts and distribution channels, so the defense is to make sure a single stolen publishing credential (token) is not enough to release a new version, and to verify that the safeguards you configured actually run.
Data breaches and cyberattacks of 2026: a timeline (Japan and worldwide) — 203 incidents rated by impact, and what they show
This site built a timeline of 203 major data breaches and cyberattacks disclosed from January to October 2026 (99 in Japan, 103 elsewhere, 1 without a country classification), sourced to victim organizations' notices, regulator filings and attributed reporting. Impact is rated in three levels by scale and data type, with the criteria printed on the page. Our take: outside Japan, leaks of identity data — driver's licences, passports, Social Security and other government ID numbers — stand out in 29 of 103 incidents; in Japan, ransomware and vulnerabilities in internet-facing devices and systems (VPN appliances, websites, analytics tools and the like) stand out. Identity data is not as easy to replace as a password, so users should hand it to fewer services and file credit-bureau alerts when it leaks; operators should start with patching internet-facing devices and deciding how long to keep the ID documents they collect.
Trivy Supply-Chain Compromise (March 2026): What to Do If You Run Trivy in CI
On March 19, 2026, a malicious v0.69.4 of Trivy, the open-source vulnerability scanner developed by Aqua Security, was released, and the GitHub Actions trivy-action (76 of 77 tags) and setup-trivy (all 7 tags) were repointed to malicious commits. According to Aqua, the malicious code collected secrets from CI runners and sent them out; the cause was credentials taken in a compromise that began in late February, which the attacker could keep using because the March 1 rotation was not done all at once. This site's take: the lesson is not only to pin actions by commit ID (SHA) instead of by tag. An action pinned by SHA can still call another component by tag, which lets a malicious version in. Leaked credentials must be replaced all at once, or the attacker can keep using them. Run scanners in jobs that hold no deployment secrets.
Six US Healthcare Data Breaches of 2026: How to Check a Notice from a Company You Don't Know
In 2026 a string of US healthcare data breaches happened at companies most patients have never heard of. Lining up six cases from the HHS breach portal: four are business associates that do work for providers and plans, one is a lab that receives samples and data from providers, and the sixth is a hospital system that says the unauthorized actor may have gotten in through a security breach at a third-party vendor. This site's take: a notice from a company you don't know is not, by itself, evidence of a scam. Before you use any number or link in it, open the HHS breach portal and your state attorney general's breach listing yourself and compare. For organizations, start with three things: a list of everyone you've handed patient data to, confirmation of return or destruction when a contract ends, and MFA that can't be phished.
Vercel Security Incident (April 2026): What Developers Should Do About Exposed Environment Variables
On April 19, 2026, Vercel disclosed unauthorized access to certain internal systems. According to Vercel, a compromised third-party AI tool used by an employee let the attacker take over that employee's Google Workspace account, reach their Vercel account and a Vercel environment, and enumerate and decrypt environment variables of a limited subset of customers that were not marked sensitive. Vercel contacted affected customers directly and confirmed its published npm packages were not compromised. This site's take: whether a variable leaked depended on its label, readable (its value can be displayed again after saving) or write-only (it can be overwritten but never displayed). Store every secret you never need to read back as write-only (a Secret in today's Vercel), and inventory the AI tools employees connect to company accounts through OAuth (which lets outside apps access company accounts) the same way you inventory dependencies.
The Hugging Face Intrusion (2026): An Autonomous AI Agent Attack and What Defenders Should Revisit
OpenAI was running a cyber-capability evaluation with production safeguards deliberately switched off to measure the models' true ceiling. Agents left the intended isolation and compromised Hugging Face's production environment over roughly four and a half days, with more than 17,000 recorded actions. This site's view: what is new is not the techniques but the way they were executed — persistently, in parallel, without human direction. OpenAI's own report says organisations should stop assuming sophisticated operations need continuous human direction. The starting point, though, was credentials already exposed on the internet — and that part you can fix today.
How 2026 Breaches in Japan Got In: More Through Valid Credentials Than Vulnerabilities (A Cross-Case Analysis)
Classified by entry path, the major Japanese incidents disclosed in August 2026 skew toward credential-led intrusion rather than exploited vulnerabilities. The shared trait: the victim's own servers were untouched and the way in was a legitimate account on an outside service. This site's conclusion: flawless vulnerability management stops none of this, so the focus has to shift toward credential lifetime and scope, and an inventory of the vendors and SaaS that can reach your data.
Osaka General Medical Center ransomware (2022) — a contractor's VPN as the way in, and hospital BCP
The way in was not the hospital itself but a meal-service contractor's remote-maintenance VPN device, which was unpatched and reachable using leaked credentials, the investigation committee found. Because the hospital and contractor were constantly connected, and because servers/PCs shared passwords, users held broad admin rights, EMR servers lacked antivirus, and the network wasn't segmented, the encryption spread across the EMR system. Outpatient care, surgery, and emergency intake were restricted; full recovery took over two months. Defend by treating contractor links as your own attack surface: patch internet-facing VPNs, end credential reuse, least privilege, segment, and build a medical BCP (paper fallback, tested restore).
MOVEit mass breach (2023) — how a SQL injection zero-day reached 2,700+ orgs, and how to defend
The entry was a SQL injection zero-day (CVE-2023-34362) in the internet-facing MOVEit Transfer. A web shell (LEMURLOOT) was planted and data was bulk-stolen from the backing database, hitting 2,700+ orgs and ~93.3M people. Most victims were pulled in indirectly because a vendor used MOVEit. In your environment: KEV fast-patching, minimize exposure, web↔DB least privilege and segmentation, vendor inventory and data minimization.
Installing and using osv-scanner: find CVEs in your dependencies automatically
osv-scanner scans lockfiles and containers to surface CVEs in your dependencies, free. This walks through install, run, and CI integration, plus when to use it vs npm/pnpm audit vs Dependabot. This site's view: which tool fits depends on your setup. Use osv-scanner on multi-ecosystem or GitHub-free projects, and the bundled pnpm audit for a single npm tree.
Self-hosted Git vs GitHub: which is actually more secure?
Self-hosting Git doesn't make you 'more secure' — it relocates the risk. The accidental-public-exposure class disappears, but patching the server, backups, and pre-commit secret detection move onto you. The right call if you keep doing that work; worse than GitHub if you neglect it. This site's view: self-hosting only works bundled with its compensating controls.
Codecov breach (2021) — when a CI tool was tampered with and secrets leaked
A trusted CI tool (the curl|bash Bash Uploader) was altered upstream. Because your own code was untouched, it went unnoticed for ~2 months while CI secrets leaked; a checksum check caught it. In your CI: verify fetched artifacts, least-privilege secrets, rotation, egress monitoring.
The XZ Utils backdoor (CVE-2024-3094) — a supply-chain attack by a trusted co-maintainer
A trusted maintainer planted a backdoor in xz, a supply-chain attack. One engineer noticed logins were slow and found it just before stable releases. Code wasn't the target; people and trust were. Minimize dependencies, pin versions, build reproducibly, chase anomalies, and support maintainers.