AI Stack Alert High TD-2026-0114 For CISO, AppSec, AI Eng, SOC / IR Unconfirmed AI-assisted
Cursor Traces Malicious Claude Code Extension to Third-Party Breach
A compromised key issued to AI infrastructure provider Fireworks was used by threat actors to access Cursor data and credentials. The source ties the access to a malicious Claude Code extension. Details beyond this are not available in the source document.
Why it matters
Organizations using Cursor or AI infrastructure providers may have exposed data and credentials. Attackers who hold a valid third-party key can reach vendor data without triggering direct attacks on the vendor's own systems.
Do this
- Inventory use of Cursor and Claude Code extensions across developer endpoints and remove unverified extensions.
- Rotate any credentials or API keys that may have been shared with Cursor or Fireworks, and review them for unexpected use.
- Review third-party AI vendor keys for scope, expiry, and anomalous usage, and revoke unused keys.
- Hunt for connections from developer workstations to unfamiliar endpoints tied to extension activity.
Signals
- NoCISA KEV
- 31/100Priority
- 50%Confidence
Evidence
“A compromised key issued to AI infrastructure provider Fireworks was used by threat actors to access Cursor data and credentials.”
Sources
Affected / related
Cursor, Claude Code, Fireworks
Published 2026-11-17 18:38 UTC · updated 2026-11-18 01:50 UTC · AI-assisted summary, checked against quoted sources · Unconfirmed: single non-official source · How we score