<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>AgentPilots Blog</title><link>https://agentpilots.ai/blog</link><description>Running and monitoring AI agents</description><item><title>What is an Agent Management System (AMS)?</title><link>https://agentpilots.ai/blog/what-is-an-agent-management-system</link><guid>https://agentpilots.ai/blog/what-is-an-agent-management-system</guid><pubDate>Tue, 13 Oct 2026 09:00:00 -0700</pubDate><description>An Agent Management System is mission control for people who run many AI agents: live status, approvals, activity and stall detection in one place. Here is how it differs from LLM observability.</description></item><item><title>MCP for monitoring: report agent status with 3 tools</title><link>https://agentpilots.ai/blog/mcp-agent-status-monitoring</link><guid>https://agentpilots.ai/blog/mcp-agent-status-monitoring</guid><pubDate>Tue, 13 Oct 2026 10:00:00 -0700</pubDate><description>A simple pattern for monitoring AI agents over the Model Context Protocol: report_status, report_step and report_message, when to call them, and how to connect Claude Code, Cursor, Codex and Grok Bot.</description></item></channel></rss>