Securing AI on the Endpoint, Anthropic Dreams of AI Replicants, and the News: Michael Jenkins, ESW #481
SC Media, Monday, October 19th, 2026
Enterprise Security Weekly: ThreatLocker's CTO on controlling local AI coding agents, plus Anthropic replicant research and the week's news.
In Enterprise Security Weekly #481, ThreatLocker CTO Michael Jenkins discusses securing local endpoint AI agents such as Claude Code, OpenAI Codex and Google Antigravity, which readily download libraries and untrusted software to reach their goals; he argues execution control and ringfencing can contain the risk.
The hosts then debate Anthropic research on AI worms distributing inference across smart TVs.
The news segment covers ShinyHunters' history, backdoored Ledger wallets, exposed Danfoss refrigeration controls and a Patch Tuesday of more than 1,400 fixes.
The interview segment is sponsored by ThreatLocker.