Cursor Confirms Breach: Malicious Claude Code Extension Distributed via Marketplace

What Happened

Cursor has confirmed a security breach that allowed attackers to distribute a malicious Claude Code extension through its marketplace, according to BleepingComputer. The attackers gained access by stealing credentials from Fireworks AI, an AI infrastructure provider, and used that access to push the malicious extension into Cursor's marketplace. Full technical details of the breach and the extension's capabilities have not been disclosed.

Why It Matters

The incident involves AI-assisted development tooling and a third-party extension marketplace. Developers who treat a vendor's marketplace as a vetted distribution channel may install a malicious package that has access to local code, credentials, or developer environments. The breach follows a supply-chain pattern: attackers compromised one vendor's credentials (Fireworks AI) to reach another vendor's distribution pipeline (Cursor's marketplace).

What Defenders Should Watch For

• Inventory which Cursor and Claude Code extensions are installed across developer workstations and flag any installed or updated around the breach window.
• Review credential and secret handling for any integrations with AI infrastructure providers, including API keys and tokens used by development tooling.
• Monitor for anomalous outbound network activity, unexpected process spawning, or credential access from IDE and extension processes.
• Apply least-privilege scoping to any tokens that AI coding assistants can access.
• Watch advisories from Cursor and Fireworks AI for indicators of compromise, affected extension identifiers, or remediation guidance.

Developing Story

This is a developing report with limited technical detail disclosed so far. No CVE has been assigned. Read the original report at BleepingComputer.