Weekly briefing·Week 35 of 2026·24 Aug 2026 to 30 Aug 2026

Threat actors deployed AI agents to target government systems and exploit tools

The week had 2 incidents, down 2 from 4, including 2 real-world incidents, down 1 from 3.

incidents

−2 vs the week before (−50%)

real-world

−1 vs the week before (−33%)

attacks on AI

AI was the target

using AI

AI was the weapon

Busiest family: AI-powered offense (2) · Top targets: Framework / tooling 2, Open-weight model 1 · 32 disclosed CVEs (+26), not counted as incidents

Threat actors deployed a multi-agent framework built on Hermes and OpenClaw in attacks against Taiwanese government systems, where investigators recovered operational workspaces documenting multiple attack waves. In another incident, a threat actor configured Hermes Agent with DeepSeek as its reasoning engine to facilitate terminal access and exploit Langflow and n8n using integrated tools and custom red-team skills. The week also had 32 disclosed vulnerabilities, up 26 from 6.

What mattered

The five most notable incidents of the week: highest severity first, then most sources. Disclosed CVEs are left out.

1
Multi-Agent Framework Compromises Taiwanese Government Systems
3 Exploitedby AIAI intrusion30 Aug 2026 · 4 sources
2
Threat Actor Uses DeepSeek-Powered Hermes Agent for Exploitation
3 Exploitedby AIAI intrusion30 Aug 2026 · 2 sources