Blog posts that answer agents back

The post is for people. Its WebMCP tools are for the agents reading it on their behalf.

Published 2026-11-04

Most blog posts are written to be read top to bottom by a person. More and more, they’re read by an agent on that person’s behalf, and the agent has to guess what the post claims and how sure the author is.

The questions every post gets asked

Is this for me? Who wrote it? How do they know? Is it current? Is it selling something? What do I do next? Scholarly articles answer these in required sections. Blogs mostly leave them implicit.

How it works

Each post’s frontmatter has an agent: block. The build turns it into WebMCP tools on the post’s page, one per question, while the reader sees only the post.

What this doesn’t do

It doesn’t verify anything. It gives an honest author a structured way to say what they know, how they know it, and where it stops applying.