Black October: at least nine AI developer platforms (Railway, Northflank, Warp, OpenCode, Fireworks AI and others) breached to steal stored AI and cloud keys, which are resold and used for LLMjacking

Wiz Research (2026-10-30) ties together a month-long wave of breaches at AI developer platforms: Fireworks AI (Oct 6), Andromeda (Oct 9), Zeabur (Oct 14, shutting down), Lightning AI (Oct 15), OpenCode (Oct 16-17), Railway (Oct 20), Warp (Oct 23), CommandCode (Oct 28) and Northflank (Oct 30). Several vendors report 'agent swarm' (AI-agent-driven) attacks; OpenCode and CommandCode call it a distillation attempt. Entry points differ (stolen credentials, privilege escalation in Railway's build infrastructure, a third-party vendor at Warp, a legacy system at OpenCode); no single vulnerability links them. Every report says the attackers hunt stored environment variables, especially AI provider keys. Wiz saw enterprise AWS Bedrock keys stolen in one of these breaches tested within 30 minutes, kept alive for days, then resold and flooded with ~800k and ~1.7M Bedrock calls (mostly Claude Opus, over $100k) through LLM reseller sites. Anyone whose secrets sat on these platforms is at risk; about $371k in crypto was reportedly stolen using keys stored in Railway. Attacker identity unknown; one reseller IP also hosted an ARTEX autonomous pentest console.

Summary written from the sources below. Check the originals before acting.