Blog
Blog
Notes and observations on AI security, MCP, and the standards and protocols shaping how AI systems actually get deployed. Mostly short pieces, written as things come up.
ai-agents
The Agent Kill Switch Problem: A Tiered Model for Stopping Tool-Using AI Agents
Stopping an AI agent is not one button. It is a control chain across runtime, tools, MCP servers, credentials, queues, downstream systems, and evidence. This post proposes eight kill-switch layers and a three-tier model that separates what's practical now from what's still research.
ai-agents agent-security mcp governance
mcp
What Changes in MCP 2026-07-28: A Before-and-After Walkthrough
The next Model Context Protocol revision is the largest since launch. Here's what actually changes between the 2025-11-25 stable spec and the 2026-07-28 release candidate, flow by flow.
mcp protocol oauth architecture