Re: WebML CG F2F TPAC 2026 (WebMCP) – meeting minutes

Hello team!

Sharing it publicly for the first time - I write (actually, a combo of write+prompt+review ) a "WebMCP book" - a comprehensive guide with technical demos included
https://webmcp.salnikov.no/

I know it's super early and spec is moving - but that's the point to have this book release-based and versioned, I set up workflows for that.

Your feedback on overall idea, content - is highly appreciated.
Please, do not share it broadly yet - it's a work in progress. Target release date - late November.

Maxim

For full transparency: this is my hobby project, not affiliated with my current employer and job responsibilities. My contribution to the webdev community.
________________________________
From: Kostiainen, Anssi <anssi.kostiainen@intel.com>
Sent: Tuesday, November 3, 2026 17:40
To: public-webmachinelearning@w3.org <public-webmachinelearning@w3.org>
Subject: [EXTERNAL] WebML CG F2F TPAC 2026 (WebMCP) – meeting minutes

Hi All,

Thanks for joining us in Dublin. It was my pleasure to chair the meeting.

Minutes for the built-in AI APIs session on Friday will be published soon.

Minutes:
https://www.w3.org/2026/10/27-webmachinelearning-minutes.html

Summary:


  1.  Intros & agenda bashing<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#7396>
  2.  WebMCP since Kobe, and lately<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#435f>
  3.  API shape & multi-stakeholder support<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#f273>
  4.  Tool discovery & cross-origin exposure<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#8b4b>
     *   Discovery beyond one traversable navigable<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#9d10>
     *   Host-controlled iframe access<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#7337>
     *   Two-sided opt-in<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#cf85>
  5.  Declarative API next steps<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#a8b3>
  6.  Developer feedback & ecosystem<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#72d8>
     *   Origin Trial feedback<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#5ab7>
  7.  Privacy & security<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#5b76>
  8.  Cross-document continuations & page context<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#5d47>
     *   Continuation designs: tokens v continueAfterNavigation<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#12df>
     *   Tool stability<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#7c16>
  9.  Human-in-the-loop & user activation<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#2159>
     *   Elicitation<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#027c>
  10. Built-in AI APIs overview<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#8029>
  11. Summary of resolutions<https://www.w3.org/2026/10/27-webmachinelearning-minutes.html#ResolutionSummary>

Thanks,

-Anssi (WebML WG/CG chair)