<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>AI agents and MCP news | SteerLatch</title><link>https://www.steerlatch.com/en/news/</link><description>News on AI agents and the Model Context Protocol: SteerLatch articles, practical guides and a dated watch of official sources (OpenAI, Google, Anthropic, EU, NIST).</description><language>en</language><atom:link href="https://www.steerlatch.com/en/news/rss.xml" rel="self" type="application/rss+xml"/><item><title>AI remote maintenance: consent, duration, STOP and traceability</title><link>https://www.steerlatch.com/en/news/ai-remote-maintenance-consent-duration-stop-traceability/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/ai-remote-maintenance-consent-duration-stop-traceability/</guid><pubDate>Fri, 13 Nov 2026 09:00:00 +0100</pubDate><category>Security and regulation</category><description>Letting an AI work remotely on a small-business computer: what the French data protection authority (CNIL) recommends, how to apply it with SteerLatch, and the limits to know.</description></item><item><title>Giving an AI the terminal: 7 security habits for a small business</title><link>https://www.steerlatch.com/en/news/ai-terminal-security-habits-small-business/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/ai-terminal-security-habits-small-business/</guid><pubDate>Fri, 13 Nov 2026 09:00:00 +0100</pubDate><category>Practical guides</category><description>The terminal is the most powerful tool an AI can use on a computer. Seven practical habits, based on CNIL recommendations, to use it without bad surprises.</description></item><item><title>Before letting an AI act on your files: back up with the 3-2-1 rule</title><link>https://www.steerlatch.com/en/news/back-up-before-ai-3-2-1-rule/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/back-up-before-ai-3-2-1-rule/</guid><pubDate>Fri, 13 Nov 2026 09:00:00 +0100</pubDate><category>Practical guides</category><description>STOP stops an AI; it does not undo what is done. A backup is the real way back: the 3-2-1 rule recommended by the CNIL, applied simply on Mac and Windows.</description></item><item><title>SteerLatch opens its free public beta on Mac and Windows</title><link>https://www.steerlatch.com/en/news/free-public-beta-mac-windows/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/free-public-beta-mac-windows/</guid><pubDate>Sun, 08 Nov 2026 09:00:00 +0100</pubDate><category>SteerLatch</category><description>Create a free account, install SteerLatch on your Mac or Windows PC and connect Claude, ChatGPT or Grok over MCP, with on-device consent and a STOP button.</description></item><item><title>Mac: allowing SteerLatch (Accessibility and Screen Recording)</title><link>https://www.steerlatch.com/en/news/allow-steerlatch-macos-accessibility-screen-recording/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/allow-steerlatch-macos-accessibility-screen-recording/</guid><pubDate>Sun, 08 Nov 2026 09:00:00 +0100</pubDate><category>Practical guides</category><description>Why macOS asks for two permissions for screenshots, mouse and keyboard, and how to grant them to SteerLatch without granting too much.</description></item><item><title>STOP, consent and revocation: staying in control of an AI that acts</title><link>https://www.steerlatch.com/en/news/stop-consent-revocation-staying-in-control/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/stop-consent-revocation-staying-in-control/</guid><pubDate>Sun, 08 Nov 2026 09:00:00 +0100</pubDate><category>Security and regulation</category><description>On-device consent, the STOP button, revoking an AI or a computer: what each SteerLatch control blocks, and what it does not.</description></item><item><title>MCP explained simply: how an AI connects to your tools</title><link>https://www.steerlatch.com/en/news/mcp-explained-simply/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/mcp-explained-simply/</guid><pubDate>Sat, 07 Nov 2026 09:00:00 +0100</pubDate><category>AI agents and MCP</category><description>The Model Context Protocol in plain words: servers, tools, OAuth sign-in and safety annotations, plus the questions to ask before connecting an AI.</description></item><item><title>A five-minute permission rehearsal before letting an agent act</title><link>https://www.steerlatch.com/en/news/five-minute-permission-rehearsal/</link><guid isPermaLink="true">https://www.steerlatch.com/en/news/five-minute-permission-rehearsal/</guid><pubDate>Fri, 06 Nov 2026 09:00:00 +0100</pubDate><category>Practical guides</category><description>Scope, review, interrupt, record: a short method to prepare any AI agent that controls a computer.</description></item></channel></rss>
