For WordPress site owners and agencies

See exactly what agents can use on your WordPress site

We independently compare what the owner expects, what WordPress publishes and what approved public checks observe. You receive the exact developer fix and follow-up evidence that shows whether it worked.

  • No WordPress login or passwords
  • Approved anonymous read-only checks only
  • No purchase request until the report passes quality review

One capability, three independent views

Illustrative report row—not evidence about a real site

Clear match or action

Example capability

Product search

Technical IDsearch_products

The report keeps intention, publication and observation separate, then explains the practical result.

Executive answer

The intended setup and the completed check agree.

This fictional row shows the conclusion first, then keeps every contributing source visible.

AlignmentMatches
Evidence coverageComplete
Views compared3

Owner, WordPress and public evidence.

Read-only checks1

One illustrative public check completed.

Actions needed0

No correction in this fictional row.

Evidence ledger

  1. 1
    Owner intentPublic product search should be available.
    Expected
  2. 2
    WordPress publicationThe reviewed capability is published.
    Published
  3. 3
    Public observationThe approved read-only check found the expected result.
    Observed
FindingEvidence sourceDeveloper actionFollow-up change

What you receive

A customer answer and a developer-ready evidence trail

The report starts with the useful conclusion, then preserves enough exact evidence to correct a confirmed difference and check it again.

A row for every capability

See the owner’s intention, WordPress publication, public observation, MCP and WebMCP side by side—without hiding a missing result inside one score.

An exact developer handoff

Confirmed differences name the capability ID, affected surface, likely cause, practical correction and the check that should be repeated.

HTML, PDF and evidence JSON

The same reviewed findings are prepared for the owner, a shareable document and the developer who needs the underlying evidence.

A follow-up comparison

After a focused correction, the follow-up shows what changed against the original baseline and keeps unresolved actions visible.

See the deliverable before you request it

Two fictional examples show a healthy result and a resolved follow-up

These examples demonstrate the capability table, technical evidence and developer handoff. They do not claim a result for a customer or public website.

Fictional format preview

Juniper Lane Supplies

A fictional store where the owner’s critical read-only capabilities agree across WordPress, the public site, MCP and WebMCP. These rows demonstrate the report format; they are not evidence about a real site.

CapabilityOwnerCorePublicMCPWebMCP
Search the sitesearch_siteExpectedPublishedObservedPublished; check passedPublished; check passed
Inspect a productwoo_get_productExpectedPublishedObservedPublished; no probePublished; no probe
Exact developer handoff

No correction in this fictional example. Keep the capability IDs and public surface membership stable, then repeat the same read-only comparison after a material publishing change.

See the evidence JSON shape

The delivered JSON keeps each capability comparison and follow-up action machine-readable. This shortened fictional fragment deliberately omits a final document identity.

{
  "sample": "fictional",
  "capability_id": "search_site",
  "owner_intent": "expected",
  "core_publication": "published",
  "public_observation": "observed",
  "mcp": "published; check passed",
  "webmcp": "published; check passed"
}

Healthy assessment

Technical and evidence files

JSON versions for developers or anyone who wants to inspect the report data.

Fictional format preview

Meridian Learning

A fictional baseline and follow-up where one capability was missing from the intended live MCP server, then matched after a focused publishing correction. These rows demonstrate the report format; they are not evidence about a real site.

CapabilityOwnerCorePublicMCPWebMCP
Inspect a product — baselinewoo_get_productExpectedPublishedObservedMissingPublished; no probe
Inspect a product — follow-upwoo_get_productExpectedPublishedObservedPublished; no probePublished; no probe
Exact developer handoff

Republish only woo_get_product in the live MCP server catalogue. Preserve its reviewed input, result and safety contracts, then repeat the fresh MCP publication check. No WebMCP correction is needed in this fictional example.

See the evidence JSON shape

The delivered JSON keeps each capability comparison and follow-up action machine-readable. This shortened fictional fragment deliberately omits a final document identity.

{
  "sample": "fictional",
  "capability_id": "woo_get_product",
  "owner_intent": "expected",
  "core_publication": "published",
  "public_observation": "observed",
  "mcp": "missing",
  "webmcp": "published; no probe"
}

Before correction

Technical and evidence files

JSON versions for developers or anyone who wants to inspect the report data.

Verified follow-up

Technical and evidence files

JSON versions for developers or anyone who wants to inspect the report data.

Both previews keep a working capability and a correction visible without turning the result into a score, badge or compatibility promise.

Choose the right depth

The free check reads public signals. The assessment compares them with owner intent and WordPress evidence.

Free public check

Get a fast snapshot from a public URL

  • No login or private file
  • Public page, crawler and discovery signals
  • Immediate self-serve result
Check a site for free

Verified WordPress assessment

Find the exact gap between intention and public evidence

  • The authorized owner’s critical capabilities and user journeys
  • The privacy-minimized evidence exported by Open for Agents Core
  • Fresh, anonymous read-only checks of approved public surfaces
  • A human-reviewed report, evidence JSON and developer action ledger
Request your assessment

How delivery works

Scope first, quality review before purchase, then a clear handoff

  1. Confirm fit and scope

    We confirm the site, owner authority, fixed one-site scope, total fee and delivery date in writing before asking for evidence.

  2. Record evidence and intent

    The owner reviews the privacy-minimized Core export, critical capabilities, critical journeys, expected surfaces and any approved anonymous read-only probes.

  3. We prepare and check the report

    Fresh public observations are reconciled with the owner’s choices and Core evidence. A human checks the findings and developer actions before anything is offered for sale.

  4. You decide after review

    If the report meets the quality standard, you can accept the written offer. We then issue the ordinary invoice and release the full report after payment.

Why the comparison matters

A public response alone cannot tell you whether the right capability was published in the right place

Each critical capability is reconciled separately. An unavailable source stays unavailable; it is not turned into a pass or a failure.

Owner intent

What the authorized owner expects, prioritizes and prohibits.

Core evidence

What WordPress records, classifies and publishes.

Public evidence

What approved anonymous read-only checks observed.

MCP

Model Context Protocol (MCP) is a standard way an AI application can find published tools. We check whether the capability appears on the intended MCP listing or server surface.

WebMCP

WebMCP is a browser page surface for published tools. We check whether the capability appears on the page where the owner expects it.

Technical scope and limits

The report marks each comparison as matches, does not match, unavailable or not applicable. It does not produce a universal score, badge or model certification.

Approved probes are limited to server-allowlisted anonymous read-only public requests. The assessment does not sign in, mutate the site, place orders, take payment, use private or session data, call an AI model or execute a capability on another origin.

A point-in-time HTTP or protocol observation does not guarantee compatibility with every agent, browser or model. Theme, hosting, CDN, unrelated plugin repair, SEO certification and continuous monitoring sit outside this one-site assessment.

Common questions

What the assessment checks—and what it does not need

What is checked?

We compare the owner’s important public capabilities and journeys with the privacy-minimized WordPress export and approved anonymous read-only observations. Each capability keeps its own evidence and practical action.

Why is this different from the free scanner?

The free check reads public signals from a URL. It cannot see what the owner intended or the publication evidence recorded inside WordPress, and it does not include the human-reviewed comparison, developer action ledger or follow-up.

Do you need credentials or private data?

No. Do not send passwords, cookies, customer records, form values, private pages or session data. The Core export is downloaded manually, checked before sending and not retained as a raw upload.

What if a check is unavailable?

The report records the evidence source as unavailable and explains what could not be observed. It does not turn a missing response into a pass, a failure or proof that a capability is absent.

How are the report and follow-up delivered?

After the report passes quality review, you receive a written offer and decide whether to buy. Payment releases the reviewed HTML, PDF and JSON delivery. An agreed follow-up compares fresh evidence with the original baseline and keeps unresolved actions visible.

For your site or a client site

Request your WordPress assessment

Send three details now. If the site is a fit, we will provide the written scope, total fee and delivery date before asking the authorized owner for evidence.

  • One public WordPress site per request
  • No passwords, customer records or private site content
  • No purchase commitment before the report passes review

Your request is handled under our privacy policy.