Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@cislunarspace-mcp 什么是轴向共振轨道？"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
地月空间入门指南术语词典的 MCP server。给 AI 客户端提供一个可调用的词典检索工具，检索对象是 cislunarspace.cn 的词典工件。
数据来自站点构建期生成的 /glossary-dictionary.json，包含 720 条术语，分 8 个类别，每条带中文名、英文名、关键词、定义正文与页面链接。server 只读，不发任何写请求。
search 按关键词检索，返回标题、英文名、分类、页面链接与定义摘要。参数 query 必填，limit 默认 5，最大 20。
get 按分类与 slug 取单条词条，返回完整定义与参考文献题录。参数 slug 可写成 orbits/aro 形式，也可以分开写 slug 与 category。
Related MCP server: tech-glossary-mcp
需要 Node 20 以上。四种客户端都用 npx 拉取本仓库的标签版本，不需要预先安装。
npm 12 起 git 依赖默认关闭，所以要带 --allow-git=all。npm 11.10 之前的版本不认识这个参数，但也只是被当作未知配置忽略，同样可用。等发布到 npm 官方仓库之后，这一串可以缩成 npx -y cislunarspace-mcp，进度见 issue #1。
Claude Desktop，写入 claude_desktop_config.json：
{
"mcpServers": {
"cislunarspace": {
"command": "npx",
"args": ["--allow-git=all", "-y", "github:cislunarspace/cislunarspace-mcp#v0.1.1"]
}
}
}Claude Code：
claude mcp add cislunarspace -- npx --allow-git=all -y github:cislunarspace/cislunarspace-mcp#v0.1.1Cursor，写入 ~/.cursor/mcp.json，形状与 Claude Desktop 相同：
{
"mcpServers": {
"cislunarspace": {
"command": "npx",
"args": ["--allow-git=all", "-y", "github:cislunarspace/cislunarspace-mcp#v0.1.1"]
}
}
}VS Code，写入 .vscode/mcp.json：
{
"servers": {
"cislunarspace": {
"type": "stdio",
"command": "npx",
"args": ["--allow-git=all", "-y", "github:cislunarspace/cislunarspace-mcp#v0.1.1"]
}
}
}配置完成后在客户端里问“什么是轴向共振轨道”，模型会调用 search 取到站内定义与链接。
启动时对词典工件发一次条件请求。数据没变时服务端返回 304，直接复用本地缓存。网络不可用时退回上一次成功拉取的缓存，只有从未拉取成功才报错退出。
缓存位置默认是 $XDG_CACHE_HOME/cislunarspace-mcp，没有设置该变量时用 ~/.cache/cislunarspace-mcp。
环境变量：
CISLUNAR_DICTIONARY_URL 覆盖词典工件地址，本地调试时可指向 file:// 或本地服务。
CISLUNAR_BIBLIOGRAPHY_URL 覆盖题录地址。
CISLUNAR_CACHE_DIR 覆盖缓存目录。
npm install
npm run build
npm test # 单元测试，不联网
npm run smoke # 联网拉线上词典并打印几组检索结果在本机用源码调试的客户端配置：
{
"mcpServers": {
"cislunarspace": {
"command": "node",
"args": ["/绝对路径/cislunarspace-mcp/dist/index.js"]
}
}
}词典工件的字段与兼容承诺记录在站点仓库的 ADR-0009，需求背景见 issue #77。
Apache License 2.0。
按分类与 slug 取一条词条的完整定义与参考文献。slug 可写成 分类/slug 形式，例如 orbits/aro。
| Name | Required | Description | Default |
|---|---|---|---|
| slug | Yes | 词条 slug，例如 aro 或 orbits/aro | |
| category | No | 分类 slug，例如 orbits。slug 已含分类时可省略 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the burden. It usefully discloses that the response includes the full definition plus references, but says nothing about missing entries, error behavior, or whether the single-slug form short-circuits the category parameter.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two tight sentences with the retrieval scope front-loaded and the slug format example last; nothing is wasted.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple two-parameter read tool with a fully documented schema, the description covers what is returned (definition + references) and the accepted slug shapes. Only the not-found/error behavior is left implicit.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% and the schema already documents both parameters, so baseline would be 3; the description's note that slug may be given as `category/slug` (with an example) clarifies the interaction between the two parameters beyond the schema text.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a concrete verb+resource: fetch one entry's full definition and references, keyed by category and slug. This distinguishes it from a listing/search operation, though it never names the sibling `search` tool explicitly.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Usage is implied—single-entry retrieval by slug—but there is no explicit statement of when to prefer this over `search`, nor any note about what happens if the slug is unknown or the category is ambiguous.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
按关键词检索地月空间入门指南的术语词典，返回词条标题、英文名、分类、页面链接与定义摘要。适合回答“什么是 X”这类问题。
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | 返回条数，默认 5 | |
| query | Yes | 检索词，中文或英文均可，例如 晕轨道、NRHO、Halo Orbit |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations exist, so the description must carry behavioral disclosure. It does useful work by enumerating the returned fields (标题、英文名、分类、链接、定义摘要) and constraining the corpus to the guide's glossary, but it says nothing about match behavior (fuzzy vs exact, ranking) or paging beyond the schema's limit field.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two short sentences: the retrieval behavior and return shape come first, then the use case. No filler, no repetition of the title.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With no output schema, the description compensates by listing the exact fields returned, and it scopes the corpus. The remaining gap is sibling routing (`search` vs `get`), which matters for a two-tool server.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100% and both params are documented there, including Chinese/English examples for `query` and a default of 5 for `limit`. The description only restates 'keyword' without adding syntax, matching rules, or limit trade-offs, so baseline 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb (按关键词检索) and a scoped resource (地月空间入门指南的术语词典), plus what is returned. It is clear what the tool does, but it never contrasts itself with the sibling `get`, so the agent must infer that `get` is for direct lookup rather than keyword search.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The closing line gives one positive usage cue (适合回答“什么是 X”这类问题), which implies the retrieval use case. There is no when-not guidance, no mention of the `get` alternative for known identifiers or full entries, and no note on how many results to request.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.1getsearchScored across 2 tools
The two tools have clearly distinct roles: search performs keyword lookup and returns summaries, while get retrieves a full entry by category and slug. There is no meaningful overlap or ambiguity between them.
Both tool names are short lowercase verbs in a consistent bare-verb style. The naming pattern is predictable and easy to distinguish.
The server only exposes two tools, which is borderline thin for a glossary server. Search and get cover the core lookup flow, but the surface feels minimal.
For a read-only terminology dictionary, search plus full-entry retrieval covers the main workflow. Minor gaps remain, such as browsing all categories or listing terms without a keyword search.
Public read-only MCP access to digital learning concepts, semantic relationships, and citations.
Search and read the aicoolies catalog of AI and developer tools via a remote MCP knowledge graph.
TermsBird: the site's own MCP server — dataset; every answer cites the site.
Read-only MCP tools for AI agent discovery, structured resources, and NIULAI information.