Arquestra Audit

io.github.pablo-ramossanchez由 GitHub 帳號 pablo-ramossanchez 發布v0.8.1更新於 Nov 3, 2026

發布前審核資料檔案：每個圖表一盞燈。你的資料永遠不會到達伺服器。

AI 翻譯由 AI 模型根據下方的原文描述翻譯，可能有誤。

原文描述

Audit a data file before publishing: one light per figure. Your data never reaches the server.

發布方自己的服務：audit.domains - Domain Valuation

已驗證Streamable HTTP可網頁執行

服務詳細資訊

識別碼: io.github.pablo-ramossanchez/arquestra-audit
傳輸: Streamable HTTP
執行環境: Web executable
信任: Verified
發布於: Nov 3, 2026

安全提示

MCP 服務會接收對話中的工具呼叫請求，並可能執行外部操作。啟用前請先了解該服務能做什麼，並只授予它所需的權限。

註冊中心已驗證此條目所屬的 GitHub 帳號 pablo-ramossanchez。這並不表明該帳號就是產品方。

SourceWeft 的檢查只涵蓋條目是否完整、端點是否回應；不審查伺服器的程式碼，也不審查其工具的行為。

在 README 中新增徽章

向讀者展示該 MCP 服務已收錄於 SourceWeft，徽章連結到本頁。

[![Listed on SourceWeft](https://sourceweft.com/badge/mcp/io.github.pablo-ramossanchez%2Farquestra-audit.svg)](https://sourceweft.com/mcp/io.github.pablo-ramossanchez%2Farquestra-audit?utm_source=readme&utm_medium=badge)

概覽

Arquestra Audit 的託管 MCP 服務（arquestra.dev）；無需登入；Streamable HTTP；5 個工具。

暫無 AI 概覽：以下資訊來自註冊中心條目和服務本身。

端點: arquestra.dev
最近檢查: 2026年11月3日

安裝

在 SourceWeft 中

開啟儀表板中的 Arquestra Audit，將其新增到工作區。
為需要使用其工具的對話啟用該服務。

Web executable，透過 Streamable HTTP。遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/io.github.pablo-ramossanchez%2Farquestra-audit/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "arquestra-audit": {
      "type": "http",
      "url": "https://arquestra.dev/api/mcp/audit"
    }
  }
}

README

Arquestra Audit · MCP server

Check your own data before you publish a result from it. Arquestra Audit is a remote MCP server by Arquestra that walks your AI assistant through an audit of a data file: it agrees with you how to read each column, measures the file where it is (in the chat's audit panel or with a small script on your computer), and shows one light per figure, red, amber or green, with the reasons.

Your data never reaches the server and nothing is stored. The measuring runs in your browser or on your computer; only counts come back.

Endpoint: https://arquestra.dev/api/mcp/audit (Streamable HTTP, no sign-in)
Page: arquestra.dev/assets/audit

Connect it

Claude (claude.ai, desktop, mobile): open Claude's connectors, search Arquestra Audit and press Connect. Not in the list yet? Add it directly.

Claude Code:

or install this repository as a plugin (Manage plugins → Install by GitHub URL → https://github.com/pablo-ramossanchez/arquestra-audit-mcp).

VS Code: install

Cursor: install

ChatGPT: settings → connectors (developer mode) → add https://arquestra.dev/api/mcp/audit.

Any other MCP client: add a remote server with the URL above.

Tools

Tool	What it does
`start_audit`	Proposes how to read your file (the role of each column, each result written as a claim) from its column names and the results you want to publish.
`measure_in_app`	Opens the audit panel in the chat; the panel reads your file in your browser and measures it with the confirmed reading.
`prepare_script`	Returns `measure.py`, a fixed Python script (standard library, no network) that counts what each check needs and prints counts, never rows.
`show_audit`	Shows the lights: one per figure, with the reasons and a small chart.
`example_audits`	How the audit works, with worked examples.

About Arquestra

Arquestra builds open, checkable data assets on public procurement and AI governance: the Consulting Market Radar (public consulting tenders in Europe and who wins them), the AI Procurement Observatory, research papers, and this audit.

Earlier versions of this repository held a plugin for a scientific-computing product; that product is now listed as Arquestra Chem on arquestra.dev.

來源：README.md，提交 dcde607

工具

start_audit

Start an audit of the person's data

Proposes how to read a data file for an audit, from its column names (optionally a type and one example value each) and the results the person wants to publish. Returns a proposed spec (the role of each column, each result written as a claim), how claims are written, the catalog of checks and the points the person has to confirm. It receives no rows.

輸入參數columns必填figuresdataset_name

measure_in_app

Measure the person's file inside the audit panel

Opens the audit panel in the chat. Without a spec, the panel receives the person's file, reads it in their browser and returns a profile of it: per column its kind, empty cells, distinct values, the values of categories with their counts, the range of numbers and dates, and the first rows. With a confirmed spec, the panel measures that same file in the browser with measure.py and shows one light per result. The file is not sent to this server.

輸入參數speccolumns

prepare_script

Get the script that measures the data

Returns measure.py with a confirmed spec written in: a fixed Python script (standard library, no network, fixed seeds) that counts what each check needs from a data file and prints those counts with its own fingerprint and no rows. Reads .csv or .json. If the spec has errors, returns the errors instead.

輸入參數spec必填columns

show_audit

Show the lights of the person's audit

Turns the output of measure.py into one light per result (red, amber or green, with its reasons), after checking that the output comes from the unchanged script and from the given spec. Shows the lights in the chat and returns them, with a link to the same view on arquestra.dev that carries only codes. Stores nothing.

輸入參數spec必填output必填

example_audits

What the audit is, with worked examples

Published example audits, such as the AI Project Radar, and how the audit works. With no arguments: the overview and the list of examples. With a name: one example, result by result (level filters by light). With a name and an id: one result with every reason and what to change. They are not the person's data.

輸入參數idlangnamelevel