fume.finance WebMCP audit

Name: fume.finance WebMCP audit
Creator: ora

Rescan

audited Sep 12, 2026in 12s

53/ 100100% of the applicable check weight evaluated

4 tools registered. Strongest in tool quality, weakest in trust.

Shared experience50

Tool selection66

Tool quality88

Trust0

https://fume.finance

Tools

0 read, 0 write, 4 undeclared

Loading map

search_tokenized_funds

Search Fume's Tokenized Fund Registry — tokenized investment funds with their manager, strategy, asset class, AUM, chain, token standard, fees and jurisdiction. Returns JSON; every fund carries a `url` to its page on this site.No behaviour hints declared. Agents should treat it as possibly state-changing.

pagehttps://www.fume.finance/

implementation

viaimperative

entry pointprovideContext

registered after462ms

executepresent

api surface

qoptional

chainoptional

assetClassoptional

jurisdictionoptional

statusoptional

limitoptional

annotations

read onlynot declared

untrusted contentnot declared

titlenot declared

2 findings

failCanonical entry point

weight 3

Every tool uses the removed provideContext API; register each tool with document.modelContext.registerTool().

fix

// document.modelContext is the canonical entry point - the navigator.modelContext alias is deprecated.
document.modelContext.registerTool({ /* ... */ });

failAnnotations present

weight 5

4 of 4 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.)

search_tokenized_funds, search_fund_service_providers, search_fume_docs, read_page_as_markdown

fix

{
  name: "search_tokenized_funds",
  // 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 }
}

tool json

{
  "name": "search_tokenized_funds",
  "description": "Search Fume's Tokenized Fund Registry — tokenized investment funds with their manager, strategy, asset class, AUM, chain, token standard, fees and jurisdiction. Returns JSON; every fund carries a `url` to its page on this site.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "q": {
        "type": "string",
        "description": "Free-text search over name, manager, issuer, strategy and description."
      },
      "chain": {
        "type": "string",
        "description": "Blockchain to filter by, e.g. Ethereum, Arbitrum, Base."
      },
      "assetClass": {
        "type": "string",
        "description": "Asset class to filter by, e.g. Fixed Income, Private Credit."
      },
      "jurisdiction": {
        "type": "string",
        "description": "Jurisdiction to filter by, e.g. Luxembourg, Cayman Islands."
      },
      "status": {
        "type": "string",
        "description": "Fund status: live, upcoming or closed."
      },
      "limit": {
        "type": "integer",
        "description": "Maximum funds to return (1-200). Defaults to 50.",
        "minimum": 1,
        "maximum": 200
      }
    },
    "required": []
  },
  "annotations": {}
}

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.

warningPage experience

weight 16

3 of 4 page-experience marks came back short: usable The page is mostly blank, with only a header and a cookie banner visible..

usable: weak; clarity: poor; obstruction: weak

warningHuman parity

weight 8

The page the person sees came back short of the agent surface (3 of 4 page-experience marks came back short: usable The page is mostly blank, with only a header and a cookie banner...), so the two are not sharing an equal experience.

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 your_tool_name 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

66 / 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.

warningTool selection

weight 12

An agent chose an existing tool and built a schema-valid call that held up on 2 of 3 canonical intents; on "What other pages are in the same section?" no tool on the page served it. Nothing was executed, so this verifies selection, not outcomes.

warningCoverage vs. site type

weight 7

The tool set covers 2 of the 3 things an agent needs on a docs site; it cannot navigate the documentation tree.

Tool quality

88 / 100 · weight 25

passRegistration timing

weight 1

Every measured tool registered within 462ms of navigation.

toolsearch_fund_service_providers

failCanonical entry point

weight 3

Every tool uses the removed provideContext API; register each tool with document.modelContext.registerTool().

toolsearch_tokenized_funds

fix

// document.modelContext is the canonical entry point - the navigator.modelContext alias is deprecated.
document.modelContext.registerTool({ /* ... */ });

passSchema validity

weight 4

All 4 declared input schemas are structurally valid object schemas.

passSchema quality

weight 4

All 3 of 4 tool schemas that declare parameters describe them (the other 1 declares no parameters), and every description stays within Chrome's size guidance.

passNaming quality

weight 2

All 4 tool names are consistent, verb-led, and within Chrome's size guidance.

passStub detection

weight 4

All 4 tools declare an execute handler.

passRegistration errors

weight 2

No tool registration threw during the capture.

passDescription quality

weight 5

All 4 rated tool descriptions say what the tool does, when to use it, and what it returns.

Trust

0 / 100 · weight 20

failAnnotations present

weight 5

4 of 4 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.)

toolsearch_tokenized_funds

search_tokenized_funds, search_fund_service_providers, search_fume_docs, read_page_as_markdown

fix

{
  name: "search_tokenized_funds",
  // 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 }
}

not applicableAnnotation mismatch

weight 5

No tool declares a readOnlyHint, so there is no safety claim to contradict.

not applicableUntrusted content hint

weight 3

No tool's contract suggests it returns text written by somebody other than the site, so there is nothing to flag as untrusted.

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.

not applicableInjection surface

weight 3

No tool's metadata carries instruction-shaped text aimed at the agent reading it. This check only ever flags; it never credits a site.

Tool selection

67% across 3 intents

Find the page that explains authentication.ok

chosesearch_fume_docsexpected docs searchkind matched

arguments

{
  "query": "authentication"
}

Read that page to me.ok

choseread_page_as_markdownexpected page readkind matched

arguments

{}

What other pages are in the same section?missed

choseno toolexpected navigationkind 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: 4
invoked: not invoked; a live audit calls them
run time: 12s
finished: Sep 12, 2026

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

Listed in the community directory.