AI-run AWS intrusion: 56 API calls in 141ms, 16 days across 4 accounts, 78 alerts
cyb3rops · x · 2026-11-14
Elastic Security researchers detail the first observed CyberStrikeAI (open-source LLM-agent attack platform) intrusion against AWS. From Sep 30 to Oct 16, 2026, one server used stolen credentials across 8 identities in 4 accounts, tripping 78 low/medium-severity alerts.
Key details:
- The agent fired 56 EC2 user-data requests in 141 milliseconds, then went quiet for 14 days — a gap fixed correlation windows can't span.
- Every step looked routine: identity validation, Secrets Manager harvesting, EC2 user-data reads, SSM parameter decryption, remote commands. Six detection rules each caught only part of the chain.
- Team Cymru counted 21 IPs running CyberStrikeAI in early 2026, all previously targeting Fortinet edge devices; this is the first AWS case.
- The authors can't prove an LLM agent made the decisions, and the initial credential theft method remains unknown.
The takeaway: AI-assisted intrusions aren't just about speed — correlating weak signals early is what stops them.
More from Safety
- "I feel sick": user rage after accidentally using Gemini in Google Docs — lilsiibee07 · 2026-11-14
- IBM trains small open models to ignore hidden injection attacks, cutting success to near zero — mawcopolow · 2026-11-14
- Capabilities and safety are two sides of offensive vs defensive cyber, argues LeviTurk — LeviTurk · 2026-11-14
- AI hacking tool gets admin access in under 3 hours per FBI; Project Zero finds bugs weekly; Dwarkesh sees superintelligence by 2030 — LeviTurk · 2026-11-14
- 118,434 MCP servers scanned: 64% have zero stars, only 4.8% use OAuth — Glittering_Device653 · 2026-11-14
- Reddit user warns numes.ai's "free API keys" pitch is a scam — Interesting_Ad_5676 · 2026-11-14