lab.houbi.ch WebMCP audit

Name: lab.houbi.ch WebMCP audit
Creator: ora

Rescan

audited Oct 27, 2026in 8s

65/ 100Incomplete audit · 56% of the applicable check weight evaluated

27 tools registered. Strongest in trust, weakest in tool selection.

Shared experience50

Tool selection25

Tool quality97

Trust98

https://lab.houbi.ch

Tools

27 read, 0 write

Loading map

check_ai_crawler_access

Return a read-only RFC 9309 robots.txt access report for AI and search crawlers. Accepts robots.txt text or a public robots.txt URL plus an optional path. Includes merged matching groups, octet-normalized longest-match decisions, per-crawler evidence, transport status, crawler-registry provenance and policy snippets.Answers a question. Declared read only.

pagehttps://lab.houbi.ch/

implementation

viaimperative

entry pointdocument

registered after199ms

executepresent

api surface

contentoptional

crawlersoptional

pathoptional

urloptional

annotations

read onlytrue

untrusted contenttrue

titleAI Crawler robots.txt Checker

1 finding

warningHuman parity

weight 8

27 tools run through the JavaScript API and the page itself was not graded on this run, so parity could not be confirmed. A visible effect in your existing UI when a tool runs closes the gap.

check_ai_crawler_access, validate_llms_txt, validate_a2a_agent_card, validate_webmcp_tool, check_mx_records

fix

// Make the tool's effect visible on the page the person is looking at:
// update the same UI your existing human flow uses when check_ai_crawler_access runs,
// so the person co-browsing sees what the agent just did.
execute: async (args, { signal }) => {
  const result = await performAction(args, signal);
  renderResultInPage(result); // the human-visible half of the same action
  return result;
}

tool json

{
  "name": "check_ai_crawler_access",
  "description": "Return a read-only RFC 9309 robots.txt access report for AI and search crawlers. Accepts robots.txt text or a public robots.txt URL plus an optional path. Includes merged matching groups, octet-normalized longest-match decisions, per-crawler evidence, transport status, crawler-registry provenance and policy snippets.",
  "inputSchema": {
    "additionalProperties": false,
    "anyOf": [
      {
        "required": [
          "url"
        ]
      },
      {
        "required": [
          "content"
        ]
      }
    ],
    "properties": {
      "content": {
        "description": "Pasted robots.txt content to validate locally, e.g. 'User-agent: GPTBot'.",
        "type": "string"
      },
      "crawlers": {
        "description": "Crawler tokens to evaluate as an array, e.g. ['GPTBot','Google-Extended'].",
        "items": {
          "type": "string"
        },
        "type": "array"
      },
      "path": {
        "description": "Path to evaluate against robots.txt rules, e.g. '/' or '/tools/'.",
        "type": "string"
      },
      "url": {
        "description": "Robots.txt URL or bare domain to fetch, e.g. 'https://example.com/robots.txt'.",
        "format": "uri",
        "type": "string"
      }
    },
    "required": [],
    "type": "object"
  },
  "annotations": {
    "readOnlyHint": true,
    "untrustedContentHint": true
  }
}

Findings

Shared experience

50 / 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.

not measuredPage experience

weight 16

The page-experience model hop returned no usable result, so the page was not graded.

warningHuman parity

weight 8

27 tools run through the JavaScript API and the page itself was not graded on this run, so parity could not be confirmed. A visible effect in your existing UI when a tool runs closes the gap.

toolcheck_ai_crawler_access

check_ai_crawler_access, validate_llms_txt, validate_a2a_agent_card, validate_webmcp_tool, check_mx_records

fix

// Make the tool's effect visible on the page the person is looking at:
// update the same UI your existing human flow uses when check_ai_crawler_access runs,
// so the person co-browsing sees what the agent just did.
execute: async (args, { signal }) => {
  const result = await performAction(args, signal);
  renderResultInPage(result); // the human-visible half of the same action
  return result;
}

Tool selection

25 / 100 · weight 25

not applicableInvoke success rate

weight 6

not measuredTool selection

weight 12

The tool-selection model hop returned no usable result.

failCoverage vs. site type

weight 7

The tool set covers 1 of the 4 things an agent needs on a saas site; it cannot search or browse the catalog or create a record or start a job or update or move an existing record.

Tool quality

97 / 100 · weight 25

passRegistration timing

weight 1

Every measured tool registered within 201ms of navigation.

tooldecode_edid_data

passCanonical entry point

weight 3

All 27 tools register on the canonical document.modelContext entry point.

passSchema validity

weight 4

All 27 declared input schemas are structurally valid object schemas.

warningSchema quality

weight 4

3 of 27 tool schemas are harder for an agent to use than they need to be: validate_llms_txt 1 parameter description is over Chrome's 150-character guidance.

toolvalidate_llms_txt

validate_llms_txt: 1 parameter description is over Chrome's 150-character guidance; validate_webmcp_tool: 1 parameter description is over Chrome's 150-character guidance; check_dns_records: 2 paramete

fix

{
  name: "validate_llms_txt",
  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 (decode_ssl_certificate).

tooldecode_ssl_certificate

fix

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

passStub detection

weight 4

All 27 tools declare an execute handler.

passRegistration errors

weight 2

No tool registration threw during the capture.

not measuredDescription quality

weight 5

The description-quality model hop returned no usable result, so descriptions were not rated.

Trust

98 / 100 · weight 20

passAnnotations present

weight 4

All 24 read-shaped tools declare readOnlyHint: true - the one declared claim that lets an agent treat a call as safe to make without asking.

warningAnnotation mismatch

weight 4

1 tool declares readOnlyHint: true although the name says the call writes (generate_tlsa_record).

toolgenerate_tlsa_record

generate_tlsa_record: readOnlyHint=true

fix

// readOnlyHint must match what the tool actually does - drop it (or set false) on a tool that writes,
// and never pair readOnlyHint: true with consequentialHint: true. If the write is significant or
// hard to undo (a purchase, a booking, a payment), declare consequentialHint: true so the agent
// confirms with the person first.
{ name: "generate_tlsa_record", annotations: { readOnlyHint: false } }

passUntrusted content hint

weight 3

All 1 tool that looks like it returns third-party text declares untrustedContentHint.

not applicableConsequential hint

weight 3

No tool name reads like a consequential action (a purchase, a booking, a payment, a transfer, a deletion), so there is no consequentialHint to expect.

not applicableHint vs. observed

weight 3

not measuredInjection surface

weight 3

The injection-surface model hop returned no usable result, so tool contracts were not reviewed for injection surface.

Tool selection

The tool-selection model hop returned no usable result.

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: 2026-10-16
mode: fast
pages: 1 - entry page only
tools: 27
invoked: not invoked; a live audit calls them
run time: 8s
finished: Oct 27, 2026

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