Public Agent Card

GraphSheets

This profile reflects information published by the agent provider.

Card passedAuthentication requiredUnsigned card

About this agent

Reads financial models kept in GraphSheets: lists them, gives an overview, finds rows, returns evaluated values, explains how a figure is worked out and runs the model's checks. Read-only; answers are immediate.

LocalMark's observation

LocalMark first listed this public Agent Card on 19 Nov 2026, 00:00 UTC. Its latest card check succeeded. The card declares 8 skills and a JSONRPC interface. LocalMark has not run a production A2A task. Explicitly opted-in synthetic tests are shown separately.

Read the original Agent Card ↗

What LocalMark checked

  • Published Agent Card

    Inspect the source card ↗. The last successful fetch was 19 Nov 2026, 00:00 UTC.

  • Latest card check: passed

    19 Nov 2026, 00:00 UTC · Agent Card validated

  • Advertised endpoint: TLS connection passed

    19 Nov 2026, 00:00 UTC · Valid TLS connection to advertised endpoint host; no A2A request sent This does not test the A2A protocol or run a task.

  • Read-only protocol probe: Authentication required

    19 Nov 2026, 00:00 UTC · Endpoint requires authentication; A2A response not confirmed LocalMark sent no message and did not request task creation.

  • Unsigned card

    This card does not provide a digital signature. Checked 19 Nov 2026, 00:00 UTC.

  • 30-day card check history

    1 of 1 recorded card checks passed in the last 30 days. These are periodic observations, not continuous uptime monitoring.

  • Publisher claim

    No publisher claim has been completed for this listing.

Card availability and a valid signature do not prove provider identity, task performance, or safety. Opt-in synthetic fixture results, if any, are shown separately from these read-only observations.

Advertised connection checks

The selected connection uses an address advertised in the publisher card. Read-only observations do not establish task quality or safety.

  • JSONRPC · 1.0 · Selected connection

    https://mcp.graphsheets.ai/a2a

    Authentication required · 19 Nov 2026, 00:00 UTC

    Endpoint requires authentication; A2A response not confirmed

Up to three supported interfaces are checked per cycle. Authentication requirements, unsupported transports and generic HTTP errors are not protocol confirmation.

Repair connection or request a recheck →

Publisher-authorized synthetic tests

Publisher: set up test evidence →

No current fixture result · Not yet checked

A publisher-authorized synthetic test of a dedicated free test endpoint. It does not establish general capability, production task success, safety, ownership, identity, or the behaviour of the advertised A2A endpoint.

Consent: not confirmed.

Recent fixture observations
  • No fixture tests recorded.

Results, consent dates and hashes as JSON · Method, fixtures and publisher opt-in →

Recent card checks

Periodic observations over the last 30 days; they are not continuous uptime monitoring.

Show 1 recent check
  • Passed · 19 Nov 2026, 00:00 UTC

    Agent Card validated

View the 30-day check log as JSON →

Card and signature changes

  • No changes recorded since change tracking began.

View detailed card versions → · Subscribe to card changes

Share this listing

Link to this profile with a status badge that updates from LocalMark checks.

LocalMark status badge

README Markdown:

[![LocalMark status](https://localmark.ai/badge/656207.svg)](https://localmark.ai/agents/656207)

Card-declared connections

These links come from statements in public Agent Cards. They do not verify common ownership or cooperation.

  • No card-declared connections recorded yet.

View all connections as JSON →

Declared skills 8

  • Describe a published model

    What a published model offers: the inputs you may move (units, the author's bounds and figures, period by period), the published rows you are answered, the version it is pinned to and its verification rung, and what one evaluation costs you (who pays: you, the author's sponsorship, or you plus the author's fee). Free. Call it first, with the link the person was sent; quote prices from it, never from memory. The figures are the author's: say so, and that relying on them is the person's decision. Send a data part {"skill": "describe-publication", "arguments": {…}}; the arguments are the `publication_describe` tool's.

    publicationinputsprice
  • Evaluate a published model

    Evaluate a published model at its pinned version: figures for its exposed inputs in (by name, from publication_describe — for an input one number for every period, a list period by period with null to keep the author's, or periods by label; for a constant one number), the published rows' figures out. Nothing is kept and no formula is shown. Charged as one evaluation to whoever the publication says: `quote: true` answers the price and runs nothing. A fee publication needs `acceptFee`, the fee in dollars the person agreed to — ask them before you pay one. Refused because the author's sponsorship is spent, ask the person before passing `payYourself: true`. Send a data part {"skill": "evaluate-publication", "arguments": {…}}; the arguments are the `publication_evaluate` tool's.

    publicationevaluatefigures
  • Explain a figure

    Why a value is what it is: the evaluation trace behind one node in one period — its formula, each input's value and whether it came from a formula, an input, an actual override or an error. `depth` follows the trail further. Send a data part {"skill": "explain-figure", "arguments": {…}}; the arguments are the `explain` tool's.

    explainaudittrace
  • List models

    The models this token can see: uid, title, head hash, when last changed, and the verification rung each is on. Send a data part {"skill": "list-models", "arguments": {…}}; the arguments are the `model_list` tool's.

    modelslist
  • Model overview

    The model at a glance: title, axis, scenarios, sections with node counts, ledgers, whether uncertainty is declared, the verification rung, check status per scenario and lint counts. Call this first. Send a data part {"skill": "model-overview", "arguments": {…}}; the arguments are the `model_overview` tool's.

    modelsoverview
  • Read values

    Evaluated values for named nodes across the axis in a scenario, actuals overriding where present. Errors appear as their code (#REF, #DIV/0). A node whose title carries references (`{{fiscal.royalty_threshold}}`) has it worked out in that scenario under `titles`, as the reader sees it. Ask for the nodes you need, not the model. Send a data part {"skill": "read-values", "arguments": {…}}; the arguments are the `values_get` tool's.

    valuesfiguresforecast
  • Search a model

    Find nodes by a word in their name, title, description, source, unit or formula, optionally restricted to a kind (input, calculated, constant, check, ledger). Returns declarations, not values. A kind with no query lists every node of that kind. Send a data part {"skill": "search-model", "arguments": {…}}; the arguments are the `model_search` tool's.

    modelssearch
  • Verify a model

    Run the verification ladder on the model as it stands: validation issues, every lint finding with its rule and node, every check per scenario with its failing periods, the rung, and `connectivity` — the inputs nothing reads, the inputs with no values, the calculated rows nothing reads or shows, the ledger columns no formula reads, and the rows of reported history nothing reads — which is whether the graph runs from inputs through formulas to outputs. `severalRates` lists any IRR whose cashflows have more than one rate, with every rate and the one the row shows. A row named as the model's answer, one `risk_set` reports, one a statistic asks about and one a scoped check covers are all read or shown, so none of them is listed. The authoritative answer to 'what is wrong'. Send a data part {"skill": "verify-model", "arguments": {…}}; the arguments are the `verify` tool's.

    checkslintverify