← All news
BleepingComputer

Cursor breach: stolen Fireworks AI credentials used to push rogue Claude Code extension

Cursor breach: stolen Fireworks AI credentials used to push rogue Claude Code extension

Cursor, the AI-powered code editor, has confirmed a security incident in which threat actors gained access to credentials belonging to Fireworks AI, a third-party AI infrastructure provider, and used them to distribute a malicious version of the Claude Code extension through Cursor's own marketplace. The company disclosed the breach after details emerged publicly, acknowledging that the compromised credentials enabled the attackers to publish the tampered extension to users.

This incident illustrates a growing risk in modern software supply chains: the compromise of a trusted downstream provider or integration partner can be leveraged to push malicious content through a seemingly legitimate channel, such as an official extension marketplace. Developer tool marketplaces and plugin ecosystems are increasingly attractive targets because a single poisoned package can reach large numbers of technically sophisticated users who may inherently trust content distributed through official channels. When AI coding assistants and code editors are involved, the stakes are heightened further, since malicious extensions can potentially exfiltrate source code, credentials, API keys, or inject backdoors into developer workflows with minimal user interaction.

Supply-chain compromises of this type typically follow a pattern: attackers first breach a smaller or less-scrutinized vendor, harvest authentication tokens or API keys, and then use those credentials to impersonate legitimate publishers on a platform with a much larger user base. This allows them to bypass direct detection at the primary platform while exploiting existing trust relationships between vendors and their customers.

For organizations using AI coding tools, extensions, or plugins, this event is a reminder to apply the same scrutiny to developer tooling supply chains as to other software dependencies. Security teams should verify the integrity and provenance of installed extensions, monitor for anomalous behavior from AI coding assistants, and review access logs for any unusual activity tied to third-party AI infrastructure credentials. Rotating API keys and credentials shared with external AI service providers, enforcing least-privilege access, and maintaining an inventory of installed extensions can reduce exposure to this category of attack. Organizations should also stay alert for official guidance from Cursor and Fireworks AI regarding remediation steps and indicators of compromise.

#supply chain attack#data breach#ai security#credential theft#software supply chain

Source: BleepingComputer

Read the original report ↗

This is an original summary written by CyberSkii for reference. All rights to the underlying reporting remain with the publisher linked above.

Need help? Contact us
Cursor breach: stolen Fireworks AI credentials used to push rogue Claude Code extension — CyberSkii News