Understand this story
On November 14, 2026, Cursor released a security incident update explaining the course of the incident and the corresponding remediation measures. The investigation results it published show that attackers used access keys leaked by Fireworks to read coding agent trajectory data in a read-only storage bucket, exposing fewer than 0.01% of customer API credentials, and used them to gain access to some SaaS tools. The attackers also used the stolen credentials to plant a malicious copy of the Anthropic Claude Code extension in the Cursor VS Code extension marketplace cache; on October 29, up to 5,000 users who requested the specific cached version may have been affected, and the copy existed for about 11 hours. Cursor said it has migrated all inference traffic away from Fireworks, rotated all exposed credentials, and rebuilt the extension marketplace, stating that the incident is under control, and that the last attack activity occurred on November 1 and has been blocked.
AI-generated from public reports
Latest update · Discussion
Baoyu joined this discussionDevelopments
1 confirmed developments
Cursor releases security incident update and remediation measures
Source:Cursor Blog
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 16
BaoyuRelated discussion
Baoyu joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Nov 14
English text is being prepared.
Cursor Discloses Investigation Results of Security Incident: Malicious Claude Code Extension Affected Up to 5,000 Users
Cursor disclosed on local time November 13 the investigation results of the malicious Claude Code extension incident, stating that up to 5,000 users were affected and the vast majority of customers were not impacted. Attackers used access keys leaked by Fireworks to steal shared bucket data, and used stolen internal credentials to inject a tampered version of Anthropic.claude-code 2.1.312 or 2.1.313 into the Cursor extension marketplace; that version remained available for about 11 hours. The prompts and code of users using privacy mode were not affected; among users not using privacy mode, the proportion of coding agent trace data leaked was less than 0.05%, and Cursor has notified affected users one by one.
IT Home · AIRelated discussion
IT Home · AI joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Nathan BenaichRelated discussion
Nathan Benaich joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Unite.AI
Cursor Discloses Data Breach: Exposed Access Key from Fireworks AI Exploited
In its November 13 incident update, Cursor stated that attackers exploited a Cursor access key exposed by Fireworks AI to copy historical coding agent trajectories and credentials, and used this to penetrate some of the software tools Cursor uses and distribute malicious Claude Code extension builds. Fewer than 0.01% of customer API credentials were exposed, and among users who did not use Privacy Mode, the proportion of coding agent trajectories exposed was below 0.05%; Cursor has notified affected users individually. Attackers also used stolen credentials to place unauthorized builds in Cursor's VS Code extension marketplace cache, and Cursor estimates that within about 11 hours, up to 5,000 users who requested a specific cached version may have received the malicious build; Cursor has moved all inference traffic away from Fireworks and rotated all exposed credentials.
CursorFirst-party
Cursor releases security incident update and remediation measures
Cursor stated that it has released an update regarding the recent security incident, which details what happened and the corresponding remediation measures.
Nov 13
Cursor BlogFirst-partySelected
Cursor Publishes Security Incident Investigation Results: Fireworks Key Leak Led to Data Exfiltration and Malicious Claude Code Extension Listing
Cursor published the investigation results of the October security incident: attackers used access keys leaked from Fireworks to read coding agent trajectory data in a read-only storage bucket, fewer than 0.01% of customer API credentials were exposed, and used this to gain access to some SaaS tools. The attackers also used the stolen credentials to place a malicious copy of the Anthropic Claude Code extension in the Cursor VS Code extension marketplace cache; on October 29, up to 5,000 users who requested a specific cached version may have been affected, and the copy existed for about 11 hours. Cursor has migrated all inference traffic away from Fireworks, rotated all exposed credentials, and rebuilt the extension marketplace, stating that the incident is under control, and that the last attack activity occurred on November 1 and was blocked.
Heat trend for this story
- Current
- 22
- Period peak
- 34 Nov 14 22:00
- 24-hour change
- -11%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.