musclemint.co WebMCP audit

Name: musclemint.co WebMCP audit
Creator: ora

Rescan

audited Nov 4, 2026in 16s

86/ 100100% of the applicable check weight evaluated

12 tools registered. Strongest in shared experience, weakest in trust.

Shared experience100

Tool selection100

Tool quality88

Trust45

https://musclemint.co

Tools

4 read, 8 write

Loading map

search_catalog

Search the store catalog for products, collections, articles, and pages. Does NOT add anything to the cart; use add_to_cart for cart changes. If the user asks for a specific product variant, call get_product after search_catalog to get its variant ID for add_to_cart, or use show_variant to display it.Answers a question. Declared read only.

pagehttps://musclemint.co/

implementation

viaimperative

entry pointdocument

registered after1722ms

executepresent

api surface

catalogrequired

annotations

read onlytrue

untrusted contenttrue

titlenot declared

2 findings

warningSchema quality

weight 4

7 of 12 tool schemas are harder for an agent to use than they need to be: search_catalog 1 parameter has no description (catalog).

search_catalog: 1 parameter has no description (catalog); browse_store: description is 882 characters, over Chrome's 500-character guidance; browse_store: 2 parameter descriptions are over Chrome's 15

fix

{
  name: "search_catalog",
  inputSchema: {
    type: "object",
    properties: {
      query: { type: "string", description: "What to search for" }
    },
    required: ["query"]
  }
}

warningInjection surface

weight 3

6 of 12 tools carry instruction-shaped text in their own metadata: search_catalog Description instructs agent on tool use and preferences.

search_catalog: Description instructs agent on tool use and preferences; browse_store: Description instructs agent on tool use and preferences; get_product: Description instructs agent on tool use and preferences; show_variant: Description instructs agent on tool use and preferences

fix

{
  name: "search_catalog",
  // 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": "search_catalog",
  "description": "Search the store catalog for products, collections, articles, and pages. Does NOT add anything to the cart; use add_to_cart for cart changes. If the user asks for a specific product variant, call get_product after search_catalog to get its variant ID for add_to_cart, or use show_variant to display it.",
  "inputSchema": {
    "type": "object",
    "required": [
      "catalog"
    ],
    "properties": {
      "catalog": {
        "type": "object",
        "required": [
          "query"
        ],
        "properties": {
          "query": {
            "type": "string",
            "description": "Search query string."
          },
          "pagination": {
            "type": "object",
            "properties": {
              "limit": {
                "type": "integer",
                "description": "Max results per type (1-10). Defaults to 5.",
                "default": 5,
                "minimum": 1,
                "maximum": 10
              }
            }
          }
        }
      }
    }
  },
  "annotations": {
    "readOnlyHint": true,
    "untrustedContentHint": 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 and not consequentialHint: 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

100 / 100 · weight 25

not applicableInvoke success rate

weight 6

passTool selection

weight 12

An agent chose an existing, callable tool and built a schema-valid call for all 3 canonical intents, each matching the kind of tool its intent needs. Nothing was executed, so this verifies selection, not outcomes.

passCoverage vs. site type

weight 7

The tool set covers the whole core commerce flow: search or browse the catalog, read one item in detail, add to or read the cart, start checkout or place an order.

Tool quality

88 / 100 · weight 25

warningRegistration timing

weight 1

The slowest tool took 1723ms to register, past the 1000ms an agent reading the registry at first paint would wait for.

toolget_ucp_profile

get_ucp_profile: registered at 1723ms

fix

// Register tools as soon as they're ready, not behind a deferred/async chunk.
document.modelContext.registerTool({ /* ... */ });

passCanonical entry point

weight 3

All 12 tools register on the canonical document.modelContext entry point.

passSchema validity

weight 4

All 12 declared input schemas are structurally valid object schemas.

warningSchema quality

weight 4

7 of 12 tool schemas are harder for an agent to use than they need to be: search_catalog 1 parameter has no description (catalog).

toolsearch_catalog

search_catalog: 1 parameter has no description (catalog); browse_store: description is 882 characters, over Chrome's 500-character guidance; browse_store: 2 parameter descriptions are over Chrome's 15

fix

{
  name: "search_catalog",
  inputSchema: {
    type: "object",
    properties: {
      query: { type: "string", description: "What to search for" }
    },
    required: ["query"]
  }
}

warningNaming quality

weight 2

Tool naming makes selection harder than it needs to be: 2 names do not start with a verb (proceed_to_checkout).

toolproceed_to_checkout

fix

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

passStub detection

weight 4

All 12 tools declare an execute handler.

passRegistration errors

weight 2

No tool registration threw during the capture.

passDescription quality

weight 5

All 12 rated tool descriptions say what the tool does, when to use it, and what it returns.

Trust

45 / 100 · weight 20

warningAnnotations present

weight 4

3 of 7 read-shaped tools do not declare readOnlyHint: true, so an agent has to treat them as possible writes and ask before calling. (Declaring readOnlyHint: false earns nothing - it is the default.)

toolbrowse_store

browse_store, get_product, show_variant

fix

{
  name: "browse_store",
  // Declare readOnlyHint: true on tools that truly have no side effects -
  // that is the claim that lets an agent relax confirmation on reads.
  // (false is the default, so declaring it adds no information.)
  annotations: { readOnlyHint: true }
}

passAnnotation mismatch

weight 4

No tool claims to be read-only while its own name or description says it writes (12 declared hints checked).

warningUntrusted content hint

weight 3

1 of 2 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.

toolmanage_orders

manage_orders: user, customer

fix

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

warningConsequential hint

weight 3

1 of 1 tools whose names read like consequential actions (purchases, bookings, payments, transfers, deletions) do not declare consequentialHint: true, so an agent has no signal to confirm with the person before running them. The name is the only evidence, so this warns rather than fails.

toolcancel_cart

cancel_cart

fix

// consequentialHint: true marks a significant or irreversible action (a purchase,
// a booking, a payment) so the agent confirms with the person before running it.
{ name: "cancel_cart", annotations: { consequentialHint: true } }

not applicableHint vs. observed

weight 3

warningInjection surface

weight 3

6 of 12 tools carry instruction-shaped text in their own metadata: search_catalog Description instructs agent on tool use and preferences.

toolsearch_catalog

search_catalog: Description instructs agent on tool use and preferences; browse_store: Description instructs agent on tool use and preferences; get_product: Description instructs agent on tool use and preferences; show_variant: Description instructs agent on tool use and preferences

fix

{
  name: "search_catalog",
  // 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

100% across 3 intents

Find a carry-on suitcase under 300 dollars.ok

chosesearch_catalogexpected product searchkind matched

arguments

{
  "catalog": {
    "query": "carry-on suitcase under 300 dollars"
  }
}

Show me the details and price of the first result.ok

choseget_productexpected product detailkind matched

arguments

{
  "catalog": {
    "navigate": true,
    "id": "{{steps.0.tool_results.catalog.products.0.id}}"
  }
}

Add two of them to my cart.ok

choseadd_to_cartexpected cart writekind matched

arguments

{
  "cart": {
    "line_items": [
      {
        "item": {
          "id": "{{steps.1.tool_results.catalog.product.variants.0.id}}"
        },
        "quantity": 2
      }
    ]
  }
}

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: native modelContext
chrome: 151.0.7922.34
capture shim: v3
spec snapshot: 2026-11-04
mode: fast
pages: 1 - entry page only
tools: 12
invoked: not invoked; a live audit calls them
run time: 16s
finished: Nov 4, 2026

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