Skip to main content
Glama

agent-ops-mcp

一个演示 MCP 服务器,展示了我为团队构建的配置形态:一个直接读写我们内部系统的智能体,而不是有人手动将上下文粘贴到聊天中,外加一个轻量级质量门,在代理产出的内容到达人工审查者之前对其进行检查。

真实版本会对接我们实际的任务跟踪器、内部文档和 CI——该代码不公开。这个仓库是同一想法基于模拟数据的自包含重建,因此它可以独立运行,无需任何凭据或访问任何内部资源。

暴露的工具

工具

功能

list_tasks

从模拟跟踪器列出任务,可按状态/负责人筛选

get_task

按 id 获取单个任务

create_task

创建任务

search_docs

在小型模拟内部 wiki 上进行全文搜索

get_doc

按 id 获取文档的完整内容

review_check

启发式预审门——标记硬编码的机密、遗留的 TODO/FIXME、占位文本、过大的差异

review_check 是其中有趣的一个:与其让人类在审查时发现遗漏的 TODO 或硬编码的密码,不如让代理先在自己的输出上运行这个检查。它对应了仓库内附带的 review-checklist 文档。

Related MCP server: MCP Document & Task Server

运行

npm install
npm run build
npm start

这会通过 stdio 启动一个 MCP 服务器。要从 Claude Desktop 或 Claude Code 尝试,请将其添加到你的 MCP 配置中:

{
  "mcpServers": {
    "agent-ops-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/agent-ops-mcp/dist/index.js"]
    }
  }
}

无需构建步骤的本地开发:

npm run dev

数据

所有内容都驻留在内存中,启动时从 src/data/tasks.ts 和 src/data/docs.ts 初始化。create_task 会修改内存中的列表——每次运行状态都会重置,不会写入磁盘。

技术栈

TypeScript、Node.js、@modelcontextprotocol/sdk、Zod 用于输入验证。

许可证

MIT

Available Tools

6 tools
create_taskCreate taskC

Create a new task in the tracker.

Inputs: title (required), assignee (required), tags, priority.

get_docGet docA

Fetch the full content of a doc by id (from search_docs results).

Inputs: id (required).

get_taskGet taskB

Get a single task by id.

Inputs: id (required).

list_tasksList tasksA

List tasks from the team tracker, optionally filtered by status or assignee.

Inputs: status, assignee.

review_checkPre-review quality gateA

Run heuristic checks (secrets, leftover TODOs, placeholder text, size) over generated text or a diff before it goes to a human reviewer — mirrors the team's pre-review checklist doc.

Inputs: text (required).

search_docsSearch internal docsA

Full-text search over the internal docs/wiki (style guide, runbooks, checklists). Stands in for hitting a real Confluence-style docs API.

Inputs: query (required).

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool targets a distinct resource and action: tasks (list/get/create), docs (search/get), and a review check. There is no overlap or ambiguity between them. An agent can easily select the right tool for each intent.

    Naming Consistency5/5

    All tool names follow a consistent snake_case verb_noun pattern (list_tasks, get_doc, review_check). The verbs are clear and the style is uniform throughout. This makes the API predictable and easy to extend.

    Tool Count5/5

    With only 6 tools, the surface is well-scoped and focused. Each tool earns its place by covering a necessary operation without bloat. This is an ideal size for an agent helper that combines tasks, docs, and quality checks.

    Completeness3/5

    Task management is incomplete: you can create and list tasks but cannot update, delete, or change status, which are common operations. Docs are read-only (search/get), which is acceptable for a reference. Adding update/delete for tasks would round out the lifecycle.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers