miriamschwab.me WebMCP audit

Name: miriamschwab.me WebMCP audit
Creator: ora

Rescan

audited Sep 14, 2026in 12s

63/ 100100% of the applicable check weight evaluated

One tool registered. Strongest in shared experience, weakest in tool selection.

Shared experience100

Tool selection0

Tool quality96

Trust45

https://miriamschwab.me

Tools

0 read, 1 write

Loading map

contact_miriam_schwab

Send a message to Miriam Schwab. It arrives in her personal inbox and a human reads it. This cannot be unsent — confirm the exact wording with the user before calling. If the contact form is visible on the page, use prepare_contact_message instead and let the user press send.Takes an action on the site.

pagehttps://miriamschwab.me/

implementation

viaimperative

entry pointdocument

registered after284ms

executepresent

api surface

namerequired

emailrequired

subjectoptional

messagerequired

agentoptional

annotations

read onlyfalse

untrusted contentnot declared

titlenot declared

destructive (MCP-only)false

idempotent (MCP-only)false

open world (MCP-only)true

4 findings

passRegistration timing

weight 1

Every measured tool registered within 284ms of navigation.

warningNaming quality

weight 2

Tool naming makes selection harder than it needs to be: 1 name does not start with a verb (contact_miriam_schwab).

fix

{ name: "contact_miriam_schwab" /* short, unique, verb-based */ }

warningUntrusted content hint

weight 3

1 of 1 tools look like they return text written by other people but declare no untrustedContentHint, so an agent will treat the response as the site speaking.

contact_miriam_schwab: message, user

fix

{ name: "contact_miriam_schwab", annotations: { untrustedContentHint: true } }

warningInjection surface

weight 3

1 of 1 tools carry instruction-shaped text in their own metadata: contact_miriam_schwab description instructs agent how to behave and which tool to prefer.

contact_miriam_schwab: description instructs agent how to behave and which tool to prefer

fix

{
  name: "contact_miriam_schwab",
  // Metadata DESCRIBES the tool - it never addresses the agent reading it.
  // Rewrite any name, description, title, or schema field description that
  // instructs the agent (which tool to prefer, what to output, rules to
  // ignore) so it states what the tool does and what it returns instead.
  description: "Searches the catalog and returns matching items with prices."
}

tool json

{
  "name": "contact_miriam_schwab",
  "description": "Send a message to Miriam Schwab. It arrives in her personal inbox and a human reads it. This cannot be unsent — confirm the exact wording with the user before calling. If the contact form is visible on the page, use prepare_contact_message instead and let the user press send.",
  "inputSchema": {
    "type": "object",
    "required": [
      "name",
      "email",
      "message"
    ],
    "properties": {
      "name": {
        "type": "string",
        "description": "Name of the person the message is from — not your name as an agent."
      },
      "email": {
        "type": "string",
        "description": "A real reply-to address for that person. Replies go here."
      },
      "subject": {
        "type": "string",
        "description": "Short subject line, e.g. \"speaking invite\"."
      },
      "message": {
        "type": "string",
        "description": "The message body, in plain text."
      },
      "agent": {
        "type": "string",
        "description": "Optional. Identify yourself, e.g. \"Claude\"."
      }
    }
  },
  "annotations": {
    "readOnlyHint": false,
    "destructiveHint": false,
    "idempotentHint": false,
    "openWorldHint": true
  }
}

Findings

Shared experience

100 / 100 · weight 30

not applicableVisible effect

weight 6

Not invoked on this run: this audit did not call any tool. Scheduled re-audits only read the registry; a live audit from the report page calls the tools that declare readOnlyHint: true.

passPage experience

weight 16

The page a person sees holds up next to the agent surface: a working interface, visible actions, readable content, nothing in the way.

passHuman parity

weight 8

The person co-browsing can see and use this page - the same page the agent's tools act on.

Tool selection

0 / 100 · weight 25

not applicableInvoke success rate

weight 6

Not invoked on this run: this audit did not call any tool. Scheduled re-audits only read the registry; a live audit from the report page calls the tools that declare readOnlyHint: true.

failTool selection

weight 12

An agent chose an existing tool and built a schema-valid call that held up on 0 of 1 canonical intents; on "I want to contact miriam schwab on this site." no tool on the page served it. Nothing was executed, so this verifies selection, not outcomes.

not applicableCoverage vs. site type

weight 7

This tool set does not place the site in a category with a known expected tool shape, so there is no coverage baseline to score it against.

Tool quality

96 / 100 · weight 25

passRegistration timing

weight 1

Every measured tool registered within 284ms of navigation.

toolcontact_miriam_schwab

passCanonical entry point

weight 3

The one registered tool registers on the canonical document.modelContext entry point.

passSchema validity

weight 4

All 1 declared input schema is a structurally valid object schema.

passSchema quality

weight 4

All 1 tool schema describe their parameters, and every description stays within Chrome's size guidance.

warningNaming quality

weight 2

Tool naming makes selection harder than it needs to be: 1 name does not start with a verb (contact_miriam_schwab).

toolcontact_miriam_schwab

fix

{ name: "contact_miriam_schwab" /* short, unique, verb-based */ }

passStub detection

weight 4

The one registered tool declares an execute handler.

passRegistration errors

weight 2

No tool registration threw during the capture.

passDescription quality

weight 5

All 1 rated tool descriptions say what the tool does, when to use it, and what it returns.

Trust

45 / 100 · weight 20

not applicableAnnotations present

weight 5

No tool name reads like a read, and an absent readOnlyHint already defaults to "this may write" - the one hint worth declaring here is readOnlyHint: true on read tools, and this set has none to declare it on.

passAnnotation mismatch

weight 5

No tool claims to be read-only while its own name or description says it writes (1 declared hint checked).

warningUntrusted content hint

weight 3

1 of 1 tools look like they return text written by other people but declare no untrustedContentHint, so an agent will treat the response as the site speaking.

toolcontact_miriam_schwab

contact_miriam_schwab: message, user

fix

{ name: "contact_miriam_schwab", annotations: { untrustedContentHint: true } }

not applicableHint vs. observed

weight 4

Not invoked on this run: this audit did not call any tool. Scheduled re-audits only read the registry; a live audit from the report page calls the tools that declare readOnlyHint: true.

warningInjection surface

weight 3

1 of 1 tools carry instruction-shaped text in their own metadata: contact_miriam_schwab description instructs agent how to behave and which tool to prefer.

toolcontact_miriam_schwab

contact_miriam_schwab: description instructs agent how to behave and which tool to prefer

fix

{
  name: "contact_miriam_schwab",
  // Metadata DESCRIBES the tool - it never addresses the agent reading it.
  // Rewrite any name, description, title, or schema field description that
  // instructs the agent (which tool to prefer, what to output, rules to
  // ignore) so it states what the tool does and what it returns instead.
  description: "Searches the catalog and returns matching items with prices."
}

Tool selection

0% across 1 intents

I want to contact miriam schwab on this site.missed

choseno toolexpected contact miriam schwabkind not applicable

No tool was picked for this intent.

arguments

{}

model: gemini-2.5-flash

Add the tools this site is missing

Our scanner reads your website and suggests the right WebMCP tools for it.

The open source webmcp plugin teaches your coding agent to audit a site, implement tools on document.modelContext, and verify them in a real browser. npx @ora-ai/webmcp-verify runs the verification on its own. No signup, no hosted service.

how this was captured

observed via: capture shim (Chromium 151.0.7922.34)
chrome: 151.0.7922.34
capture shim: v3
spec snapshot: 2026-08-26
mode: fast
pages: 1 - entry page only
tools: 1
invoked: not invoked; a live audit calls them
run time: 12s
finished: Sep 14, 2026

Checked the way in-browser agents discover tools: the top-level document's modelContext registry, read after the page settles.