agenciaconversion

Conversion Agent Discovery helps WordPress sites expose public, read-only discovery surfaces for AI agents without pretending to support capabilities that are not implemented.
It is designed to improve measurable agent accessibility for public, read-only discovery checks without calling external scanners automatically.
Features include:
<ul>
<li>Markdown negotiation for public content with <code>Accept: text/markdown</code>.</li>
<li>Content Signals in <code>robots.txt</code>.</li>
<li>Generated <code>llms.txt</code> and <code>/.well-known/llms.txt</code>.</li>
<li>Editable AI Instructions page for official brand facts and LLM guidance, starting empty with English templates by default.</li>
<li>Runtime Front-end Placement for exact, design-safe AI Instructions link placement across matching public header, footer, or page content components.</li>
<li><code>[conversion_ai_instructions_link]</code> shortcode for manual placement.</li>
<li>Generated <code>/.well-known/api-catalog</code>.</li>
<li>Generated <code>/.well-known/agent-skills/index.json</code> and virtual <code>SKILL.md</code> files.</li>
<li>Local Agent Skills discovery schema generated by the site.</li>
<li>Read-only WebMCP tool registration with runtime diagnostics for compatible browsers.</li>
<li>Production-oriented Agentic Browsing readiness guidance for PageSpeed and Lighthouse checks.</li>
<li>Optional origin isolation headers and optional Origin-Trial header for browser WebMCP runtime support.</li>
<li>Opt-in declarative WebMCP form annotations for exact rendered form opening tags.</li>
<li>Public read-only REST endpoints for site search, recent content, single content reads, site context, and contact handoff.</li>
<li>Agentic Browsing diagnostics that separate plugin-controlled discovery surfaces from theme accessibility and layout stability checks.</li>
<li>Admin settings page with a global public switch.</li>
</ul>
Conversion Agent Discovery does not publish fake OAuth, MCP Server Card, A2A, or commerce metadata.
The AI Instructions feature was inspired by Chris Long’s public LinkedIn post and Nectiv case study showing that AI Instructions pages can be cited by ChatGPT in branded queries. It also credits the earlier Seer Interactive AI Information page by Wil Reynolds and Alisa Scharf that Chris Long identified as an important example.
References:
<ul>
<li>Chris Long’s LinkedIn post: https://www.linkedin.com/posts/chris-long-marketing_holy-moly-seos-ai-instruction-pages-influence-share-7466096236799053824-_rpb/</li>
<li>Nectiv case study: https://nectivdigital.com/blog/aeo-experiment-chatgpt-cited-an-ai-instructions-page-in-48-hours</li>
<li>Nectiv AI Instructions example: https://nectivdigital.com/ai-instructions</li>
<li>Seer Interactive AI Information example: https://www.seerinteractive.com/ai-information</li>
<li>Omniscient Digital AI Info example: https://beomniscient.com/ai-info/</li>
<li>llms.txt proposal by Jeremy Howard: https://llmstxt.org/</li>
</ul>
Important: this plugin does not guarantee LLM citations, search rankings, AI Overview visibility, or model behavior. It provides controlled public discovery surfaces for experimentation.
<h3>External Services</h3>
This plugin does not send requests to external services automatically and does not transmit visitor or site data to third parties.
The generated Agent Skills discovery document uses a local schema URL served by the WordPress site at <code>/.well-known/agent-skills/discovery-0.2.schema.json</code>; it does not depend on an external schema service.
The plugin contains optional external links that administrators or users may open manually:
<ul>
<li><code>https://conversion.ag/</code> is linked for Conversion attribution and plugin author information. The plugin does not send visitor or site data to Conversion; the link is opened only when a user clicks it manually. Conversion’s privacy policy is available at https://www.conversion.com.br/politica-de-privacidade/.</li>
<li><code>https://github.com/agencia-conversion/conversion-agent-discovery</code> is linked as the public development repository and release source. The plugin does not send visitor or site data to GitHub; the link is opened only when a user clicks it manually. GitHub’s terms are available at https://docs.github.com/en/site-policy/github-terms/github-terms-of-service and GitHub’s privacy statement is available at https://docs.github.com/en/site-policy/privacy-policies/github-general-privacy-statement.</li>
<li><code>https://wordpress.org/plugins/wp-graphql/</code> is linked from the admin screen when WPGraphQL is not detected, so administrators can optionally install it. The plugin does not send visitor or site data to WordPress.org; the link is opened only when an administrator clicks it manually. WordPress.org’s privacy policy is available at https://wordpress.org/about/privacy/ and terms are available at https://wordpress.org/about/terms-of-service/.</li>
<li><code>https://pagespeed.web.dev/</code> is linked from the Agentic Browsing admin screen so administrators can manually test their public site in PageSpeed Insights. The plugin does not send visitor or site data to Google automatically; the link is opened only when an administrator clicks it manually. Google’s privacy policy is available at https://policies.google.com/privacy and terms are available at https://policies.google.com/terms.</li>
</ul>
The developer-only <code>tools/run-agentic-lighthouse.sh</code> script is not run by WordPress and is never executed automatically. If a developer runs it manually, it uses <code>npx</code> to download and run Lighthouse from the npm registry and audits only the URL supplied by the developer.
The optional WebMCP Origin Trial token field is empty by default. If an administrator adds a public Chrome origin trial token, the plugin sends it as an <code>Origin-Trial</code> response header on public pages together with the opt-in WebMCP origin isolation headers. The plugin does not request or generate tokens and does not make a server-side request to Google or Chrome services.
Conversion’s privacy policy is available at https://www.conversion.com.br/politica-de-privacidade/.
<h3>Privacy</h3>
Conversion Agent Discovery does not track users and does not send site data to external services automatically. Outbound links in the admin screen and readme are plain manual links.
<h3>Development</h3>
Development happens at https://github.com/agencia-conversion/conversion-agent-discovery. The distributed plugin contains readable PHP, JavaScript, CSS, and SVG source files; no build step is required to run the plugin.
Local Agentic Browsing validation can be run by developers with:
<ul>
<li><code>tools/setup-agentic-fixture.sh</code></li>
<li><code>tools/run-agentic-lighthouse.sh http://localhost:8089/</code></li>
<li><code>node tools/assert-agentic-lighthouse.mjs .context/lighthouse-agentic.json</code></li>
<li><code>CAD_REQUIRE_WEBMCP=1 node tools/assert-agentic-lighthouse.mjs .context/lighthouse-agentic.json</code> when running a Chrome runtime that exposes real WebMCP APIs.</li>
</ul>
Public 0.4.0 launch materials are available in the repository:
<ul>
<li><code>docs/launch-0.4.0.md</code></li>
<li><code>docs/local-agentic-browsing-activation.md</code></li>
<li><code>docs/demo-checklist-0.4.0.md</code></li>
<li><code>docs/changelog-0.4.0.md</code></li>
</ul>

Conversion Agent Discovery

Read-only agent discovery for WordPress: llms.txt, Markdown, API Catalog, Agent Skills, WebMCP, and content signals.

Matthias Pfefferle

mrose17

This plugin enables “Well-Known URIs” support for WordPress (RFC 5785: http://tools.ietf.org/html/rfc5785).
From the RFC:
<blockquote>
It is increasingly common for Web-based protocols to require the 
 discovery of policy or other information about a host (“site-wide 
 metadata”) before making a request. For example, the Robots 
 Exclusion Protocol <a href="http://www.robotstxt.org/" rel="nofollow ugc">http://www.robotstxt.org/</a> specifies a way for 
 automated processes to obtain permission to access resources; 
 likewise, the Platform for Privacy Preferences 
 tells user-agents how to discover privacy policy beforehand.
While there are several ways to access per-resource metadata (e.g., 
 HTTP headers, WebDAV’s PROPFIND [RFC4918]), the perceived overhead 
 (either in terms of client-perceived latency and/or deployment 
 difficulties) associated with them often precludes their use in these 
 scenarios.
When this happens, it is common to designate a “well-known location” 
 for such data, so that it can be easily located. However, this 
 approach has the drawback of risking collisions, both with other such 
 designated “well-known locations” and with pre-existing resources.
To address this, this memo defines a path prefix in HTTP(S) URIs for 
 these “well-known locations”, “/.well-known/”. Future specifications 
 that need to define a resource for such site-wide metadata can 
 register their use to avoid collisions and minimise impingement upon 
 sites’ URI space.
</blockquote>
You will need ‘manage_options’ capability in order to use the Settings 
page for this plugin.

Conversion Agent Discovery vs /well-known-uris/