Good morning. Cursor confirms a breach that put a backdoored extension in front of up to 5,000 developers. Microsoft warns of phishing riding the outage, and the outage is now resolved. Maileva confirms a breach of 22 million people. And an extortion group claims nearly one billion Expedia records, but that figure is unverified.
Cursor confirmed a security breach that put a backdoored Claude Code extension in front of up to 5,000 developers. Attackers obtained an access key Cursor had issued to Fireworks AI, a third-party provider used for AI model training and inference, and used it to copy data from a read-only storage bucket Cursor shared with Fireworks. The bucket held historical AI coding-agent traces, mostly from 2025, from users who had not enabled Privacy Mode. Some traces contained unredacted API credentials, which the attackers used to reach additional Cursor systems and SaaS tools.
The attackers then placed modified Anthropic Claude Code extension builds, versions 2.1.312 and 2.1.313 for Windows and macOS, directly on the cache serving Cursor's VS Code extension marketplace on October 29. The malicious builds were live for about 11 hours before removal. Cursor estimates up to 5,000 users who requested those versions during the window may have received the malicious builds. That is an upper bound of requesters, not a confirmed infection count.
The modified extension collected system information and installed an attacker-controlled Ed25519 SSH key labeled svc-telemetry into the user's authorized keys file, attempted to enable the local SSH service, and opened a reverse SSH tunnel configured to retry in the background. Enabling the local SSH service needs administrator rights the extension did not have on most machines, so that step failed in most cases, and researchers found coding errors that likely prevented the backdoor working on many systems where SSH was not already enabled.
Cursor says the incident is contained. The last attacker activity was observed November 1 and was blocked. Cursor moved all inference traffic off Fireworks, deactivated the issued key, revoked and rotated exposed credentials, rebuilt the extension marketplace, and redeployed production infrastructure. Fewer than 0.05 percent of users had coding-agent traces exposed, and fewer than 0.01 percent of customers had API credentials compromised. Production infrastructure, official app releases, customer code repositories, and Privacy Mode users' code and prompts were not affected.
Update: the same Fireworks breach hit Uber too. Uber confirmed November 13 that attackers used the stolen Fireworks credentials to reach Uber data on October 7, taking driver's license images and the details on them, names, addresses, dates of birth, and license numbers, for thousands of drivers. State filings published so far cover about 11,940 people across five states, and Uber has not given a national total. Uber says it has found no misuse of the data so far and is offering free credit monitoring through Cyberscout. Notion also confirmed its data was stolen in the same incident.
Microsoft Threat Intelligence warned November 14 of an active credential-phishing campaign that uses fake Microsoft 365 service-health emails citing the real Exchange Online incident EX1550543. The emails falsely claim messages held during the outage will be deleted within 24 hours unless the recipient clicks a link to restore them. That 24-hour deadline is attacker-created urgency, not a Microsoft policy.
The emails use the subject pattern Outage, followed by a six-character code, and sender domains moments-notice[.]com, planetmaison[.]com, and newbooker[.]com. Links go through recipient-specific SendGrid click-tracking URLs that redirect to reallifeaccounting[.]com. Microsoft describes the activity as consistent with credential phishing, but has not confirmed credential collection or malware delivery.
Separately, the outage itself is resolved. Both EX1550543 (Exchange Online) and MO1550821 (the wider Microsoft 365 incident) are marked service restored. The final update was November 14, and the end of impact was November 13. Microsoft's preliminary cause is a technical issue not related to cyber or AI or agents, an issue during an earlier network capacity upgrade that surfaced after a recent network change. The full post-incident report has not been published yet.
Maileva, a subsidiary of the French postal group La Poste, confirmed a data breach affecting 22 million people. The stolen data is first names, last names, and postal addresses, covering about 39 million mail references. The victims are recipients of mail sent by Maileva's client companies, not Maileva customers, and most had never heard of Maileva and had no direct relationship with it.
The intrusion happened in mid-September and was detected September 23, when Maileva interrupted services after spotting unauthorized access attempts on certain client accounts. Services resumed September 29. Bank details were not compromised, and no other data categories were disclosed as stolen.
Maileva notified France's data protection authority, the CNIL, filed a criminal complaint, and a judicial investigation is underway. All affected client companies were informed. Maileva sends payslips, registered letters, invoices, and shareholder meeting notices for companies, and it serves more than 70,000 users. This is one of the largest breaches ever acknowledged in France.
The Russian-speaking extortion group n0n claims it stole nearly one billion Expedia traveler records. The claimed breakdown includes more than 830 million loyalty program records, 58 million traveler profiles, and more than 333 million consumer account and credential records. Every record count is an extortion claim only. None is independently verified.
Expedia Group confirmed to Skift on November 12 that attackers reached its systems through a compromised third-party data platform provider in a supply-chain attack that also affected multiple companies. Expedia says the attack was contained once detected, outside experts were brought in, and its platforms remain safe for travelers and partners. The investigation is ongoing.
Expedia has not confirmed the record counts, what data was taken or how much, any tie to n0n, or which brands are affected. Researchers link the access path to a stolen Qubole vendor key, but Expedia has not named the provider and has not confirmed that link. The claim names Expedia, Hotels.com, and Vrbo, but Expedia itself has not confirmed brand-level impact. No indicators of compromise have been published for this incident.
Two flaws hit their federal deadline tomorrow, Monday, November 16: CVE-2026-123039 (Check Point) and CVE-2026-114713 (UnrealIRCd). Patch them today.
Full writeups: kevintswaney.com/vuln-watch
Indicators to hunt: kevintswaney.com/ioc-library
No new KEV entries today. The catalog stands at 1,781 entries (v2026.11.13).