clify
Verified websites as JSON commands for AI agents over plain curl or MCP.
我该使用它吗
质量与安全性
发现（2）
- LOWTool 'crates__get-crate' doesn't follow camelCase/snake_case在 crates__get-crate 中
- LOWTool 'pypi__get-package' doesn't follow camelCase/snake_case在 pypi__get-package 中
基于对工具定义和协议合规性的自动分析。
上下文开销
这是每次将服务器的工具加载到模型上下文窗口时所消耗的大致 token 数。数值越高，可用于其他任务的注意力就越少。
安装
一键安装
将以下内容添加到你的 `claude_desktop_config.json` 文件中：
{
"mcpServers": {
"clify": {
"url": "https://bold-cloud-h4u6s.run.mcp-use.com/mcp"
}
}
}远程端点
https://bold-cloud-h4u6s.run.mcp-use.com/mcpstreamable-http它能做什么
工具清单
工具（7）
🟢clify_sites(query, category, limit, offset)
Search the catalogue of 135 real websites clify can call (package registries, government data, science, news, jobs, weather, culture…). Pass a keyword query to get the best sites with their matching commands. Without a query it pages through the catalogue, one line per site ("name (category, commands): about"); filter with category. Start here.
输入模式
{
"type": "object",
"properties": {
"query": {
"description": "what you need, e.g. \"rust crate versions\" or \"weather forecast\"",
"type": "string"
},
"category": {
"type": "string",
"enum": [
"culture",
"dev",
"gov-eu",
"gov-nordic",
"gov-us",
"jobs-market",
"maps-weather",
"news",
"sandbox",
"science",
"sports-ent"
]
},
"limit": {
"description": "max sites (default 6 with a query, 40 per page without)",
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 9007199254740991
},
"offset": {
"description": "skip this many sites when paging the catalogue (see nextPage)",
"type": "integer",
"minimum": 0,
"maximum": 9007199254740991
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢clify_help(site, command)
A site's commands with params ("name*" = required, ":a|b" = allowed values, "=x" = default). Pass command for full param docs, default fields and a ready clify_run call.
输入模式
{
"type": "object",
"properties": {
"site": {
"type": "string",
"description": "site name from clify_sites, e.g. crates, weather"
},
"command": {
"description": "one command name for full details",
"type": "string"
}
},
"required": [
"site"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🔴clify_run(site, command, args, fields, limit, ...)
Call one command of a clify site and get JSON back. args holds the command's params (see clify_help). Lists return count, total and nextPage (args for the next page). Write commands return a dry-run preview unless confirm:true.
输入模式
{
"type": "object",
"properties": {
"site": {
"type": "string",
"description": "site name, e.g. crates"
},
"command": {
"type": "string",
"description": "command name, e.g. get-crate"
},
"args": {
"description": "params, e.g. {\"name\":\"serde\"}",
"type": "object",
"propertyNames": {
"type": "string"
},
"additionalProperties": {
"type": [
"string",
"number",
"boolean"
]
}
},
"fields": {
"description": "comma-separated keys to keep (dot paths ok); \"*\" for all",
"type": "string"
},
"limit": {
"description": "first N items of a list",
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 9007199254740991
},
"confirm": {
"description": "true sends a write; otherwise writes return a preview",
"type": "boolean"
}
},
"required": [
"site",
"command"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢crates__get-crate(name, fields, limit)
crates.io: Get a crate's details: description, links, download totals, keywords, categories and default version Returns name, description, default_version, downloads, recent_downloads, num_versions, keywords, categories, repository, documentation, homepage, updated_at by default. (Same as clify_run site=crates command=get-crate.)
输入模式
{
"type": "object",
"properties": {
"name": {
"type": "string",
"description": "Crate name"
},
"fields": {
"description": "keys to keep; \"*\" for all",
"type": "string"
},
"limit": {
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 9007199254740991
}
},
"required": [
"name"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢pypi__get-package(name, fields, limit)
pypi.org: Get a package's PyPI page for its latest version: name and version, summary, pip install command, project links and metadata sidebar, and README text. Returns name_version, summary, pip_command, sidebar, description by default. (Same as clify_run site=pypi command=get-package.)
输入模式
{
"type": "object",
"properties": {
"name": {
"type": "string",
"description": "Package name as it appears on PyPI, e.g. requests"
},
"fields": {
"description": "keys to keep; \"*\" for all",
"type": "string"
},
"limit": {
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 9007199254740991
}
},
"required": [
"name"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢wikipedia__summary(title, fields, limit)
en.wikipedia.org: Short summary of an article (title, description, extract, url) Returns title, description, extract, content_urls.desktop.page by default. (Same as clify_run site=wikipedia command=summary.)
输入模式
{
"type": "object",
"properties": {
"title": {
"type": "string",
"description": "article title, spaces as underscores, e.g. Y_Combinator"
},
"fields": {
"description": "keys to keep; \"*\" for all",
"type": "string"
},
"limit": {
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 9007199254740991
}
},
"required": [
"title"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢hnsearch__search(query, tags, page, hitsPerPage, fields, ...)
hn.algolia.com: Search stories/comments by relevance Returns title, points, num_comments, author, created_at, url, objectID by default. (Same as clify_run site=hnsearch command=search.)
输入模式
{
"type": "object",
"properties": {
"query": {
"type": "string"
},
"tags": {
"type": "string",
"enum": [
"story",
"comment",
"ask_hn",
"show_hn",
"poll"
],
"description": "default story"
},
"page": {
"type": "number"
},
"hitsPerPage": {
"type": "number"
},
"fields": {
"description": "keys to keep; \"*\" for all",
"type": "string"
},
"limit": {
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 9007199254740991
}
},
"required": [
"query"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}社区
证据