// Threat intel

Morning Threat Brief: A cloud host tells every customer to rotate all keys, your AI coder may have posted your passwords on GitHub, a retailer breach hits 7.7 million, and a data broker's SSNs are claimed exposed

Good morning. A cloud platform that hosts apps and AI workloads told every customer to assume their stored secrets were decrypted and rotate everything, after stolen AI keys were already being abused. Researchers found more than a thousand public GitHub files where an AI coding assistant saved passwords and API keys in plain text. A Japanese retailer's third-party cloud service was breached, exposing data on about 7.7 million customers. And a ransomware group claims it exposed 1.7 billion consumer rows from a US data broker, including Social Security numbers in plaintext, a claim nobody has verified. Rotate your keys, check your repos, and assume your data broker already has your number.

1. A cloud host tells every customer to assume their secrets were decrypted and rotate all keys. Developer platform Northflank opened a security incident on October 30 after a threat alert from Wiz, and within twenty minutes told customers to assume every environment variable had been decrypted and to disable or rotate all keys. By that evening customers were reporting unauthorized use of AI provider API keys stored on the platform. A November 6 update says the first known decryption of secrets happened between 02:39 and 03:08 UTC on October 30, eight to nine hours before the first alert, and that the flaw, CVE-2026-98432, was patched across the fleet within an hour of the alert. Northflank has not published a root cause or a count of affected customers, and says running workloads were unaffected. Wiz counts Northflank as at least the ninth AI developer platform to disclose an incident in October, part of what it calls the Black October wave of LLM credential theft. Trigger.dev, another platform for building AI agents, warned of its own environment variable breach and also told customers to rotate. If you use Northflank, rotate AI and payment keys first, then database and cloud credentials, then the rest, and check usage and billing with each provider for charges you did not make.

2. Your AI coder may have posted your passwords on GitHub. Clicking "Yes, and don't ask again" in Claude Code can save the exact command you approved, password included, into the project's .claude/settings.local.json file. Security firm Apiiro searched public GitHub and found 104,185 copies of that file in 103,617 repositories; 1,122 of them hold at least one secret, 1,379 distinct secrets in total. More than a third open a database directly, including 379 database passwords, plus 122 Supabase service-role keys, 43 SSH logins for public servers, 42 GitHub, GitLab and npm tokens, and 23 AWS keys. Apiiro did not test whether the credentials still work, but one company it notified confirmed its key had been live. Two common secret scanners missed some of them: gitleaks flagged 564 of the 1,122 files and TruffleHog flagged 643, and neither flagged 215. Deleting the file does not undo the leak, because older versions stay in git history and in forks; Apiiro found 734 more secrets in older versions across 509 repositories. The bug was first reported in August 2025, and Anthropic has not published a product response. If you use Claude Code, search your repos for the file, add it to your project .gitignore, and rotate any credential it ever held, including ones from deleted versions.

3. A Japanese retailer breach hits 7.7 million customers through a third-party cloud service. Manufacturer Iris Ohyama, which runs the Iris Plaza shopping platform, disclosed on November 8 that unauthorized access to a third-party cloud service it used may have exposed personal information belonging to about 7.7 million people. The data includes names, addresses, phone numbers, email addresses, birth dates, and account passwords, covering about 5.8 million members and 1.9 million non-members. Credit card data was not stored on that service, and the company has not identified a perpetrator. The breach is a reminder that your data sits in systems you never chose: a vendor your retailer hired. If you shop at Iris Plaza, change that account password, and if you reused it anywhere else, change those too. For everyone else, the check is the same one as the bank heist last week: ask which third-party services hold your customer data and how access to them is controlled and logged.

4. A ransomware group claims it exposed 1.7 billion consumer rows from a data broker, including Social Security numbers in plaintext. UNVERIFIED. The group N0n posted a claim on November 7 saying it exposed data from US data broker Infutor Data Solutions, including a warehouse table of over 1.7 billion consumer rows with names and dates of birth, 590 million survey responses, 1.1 billion address history entries, and Social Security numbers stored in plaintext, plus client delivery credentials also in plaintext. The listing is labeled "NOT PAID - EXPOSED". N0n posted 13 other victims the same week, including Staples/SBD with 381,964 margin records, Lampenwelt with 81 million search queries, and Higginbotham insurance with 217,000 policy documents. Two independent leak-site crawlers saw the posting, which confirms only that the group published the claim. No mainstream outlet has confirmed a breach, and Infutor has said nothing. Treat the row counts as the group's claim, not a count of people affected. The practical point stands either way: data brokers hold your Social Security number, and some of them store it in plaintext. Freeze your credit if you have not already, and check your free credit reports.

Warning: the VMware ESX flaw CVE-2026-41709 is due Wednesday, November 11, and the second Ivanti flaw CVE-2026-118527 is due Friday, November 13. The Apple KEV entry CVE-2024-27834 is due November 18, and the SonicWall SSL-VPN flaw CVE-2026-19868 is due Friday, November 20.

Full writeups: kevinthecyberguy.com/vuln-watch
Indicators to hunt: kevinthecyberguy.com/ioc-library

No new CISA KEV additions today. The catalog stands at 1,769 entries, last updated November 6.

Kevin Swaney, Principal Cybersecurity Engineer

Kevin Swaney

Principal Cybersecurity Engineer with 17+ years across banking, payments, and consulting. Writing about AI security, identity, and the fundamentals that stop most incidents. kevintswaney.com

← kevintswaney.com