Skip to main content
Glama

wire-build-mcp — 柏讯产物构建 MCP

Part of the Wire line (寻数·柏讯) of FindData — artifact build surface: create and publish data-entry forms and record views from specs, with draft-first write discipline.

柏讯(wire 线)的产物构建面 MCP:让 AI agent 能按 spec 创建/更新/发布数据录入 表单(form)与单记录扫码页(record-view),并取数据集列事实、按需建集。草稿优先的 写入纪律:创建/更新一律落 draft,发布必须显式单独调用。

  • 契约:openspec/changes/wire-ai-artifact-build/specs/wire-build-mcp/spec.md (finddata 工作区;后续治理面见 wire-ai-data-steward)

  • 接入规约:LINES-MCP-CONFORMANCE.md

与 wire-ops-mcp 的分工

入口

定位

工具面

wire-ops-mcp

客户面读操作

需求单 / 数据集只读查询 / 市场订阅

wire-build-mcp

产物构建(写)

产物 CRUD + 列事实 + 建集

两入口并列,各自独立部署与凭据;wire-ops-mcp 十工具面不受本服务影响。

Related MCP server: Deriva MCP Server

工具面

组

工具

写

事实

wire_dataset_schema(列名+类型)、wire_artifact_list、wire_artifact_get

全读

建集

wire_dataset_create(CSV 内容 → 导入建集)

✅

产物写

wire_artifact_form_create、wire_artifact_record_view_create、wire_artifact_update

✅(落草稿)

状态

wire_artifact_publish、wire_artifact_archive、wire_artifact_restore

✅

两条硬纪律

  1. 写入落草稿,发布独立。上游 publish 端点自身不做产物结构校验(仅状态迁移), 因此本服务不提供「创建即发布」的开关(多余参数会被 MCP 层直接拒绝)。 发布前请人工在控制台目视一次,或确认 spec 已过校验。

  2. 列名核对。spec 引用了数据集中不存在的列时,错误响应会带上该数据集的真实 列名清单——生成 spec 前请先用 wire_dataset_schema 取事实,别编列名。

鉴权

进站 Bearer 原样转发,wire 侧判轨(Logto ID token console 轨 / wdk_ 数据面 key):

  • 无凭据:本服务 401,不触上游

  • 坏凭据:上游 401 原文透传

  • 跨企业产物/数据集:404 与「不存在」不可区分

环境

变量

默认

WIRE_BUILD_BASE_URL

https://wire.finddatatech.cloud

WIRE_BUILD_TIMEOUT_SECONDS

20

WIRE_BUILD_AUTH_CACHE_TTL

60

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI systems to manage Verodat accounts, workspaces, and datasets, including creating datasets, querying data, and executing AI-powered queries through natural language.
      9 npm
      Apache 2.0
    • A
      license
      Not graded
      quality
      B
      maintenance
      Turns Google Sheets, CSV, Excel, and Parquet files into durable, queryable data for AI agents. Lets clients securely discover datasets, inspect semantic models, run bounded analytics, and manage approved semantic overlays.
      Apache 2.0
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables AI agents to ingest CSV/TSV files, profile and categorize tables, generate SQL DDL for multiple engines, and launch background data workflows whose completion and artifacts can be independently verified.
      MIT