Skybridge Everything App
Example and testing ChatGPT and MCP App showcasing all the features of the Skybridge framework
我该使用它吗
质量与安全性
B
发现（4）
- LOWTool 'show-everything' description lacks action verb在 show-everything 中
- LOWTool 'show-everything' doesn't follow camelCase/snake_case在 show-everything 中
- LOWTool 'flip-coin' doesn't follow camelCase/snake_case在 flip-coin 中
- LOWTool 'inspect-file' doesn't follow camelCase/snake_case在 inspect-file 中
基于对工具定义和协议合规性的自动分析。
上下文开销
~1,124token 数（工具定义）
~4.4 KB典型响应大小
对注意力有中等影响（占 128k 上下文窗口的 0.88%）
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": {
"everything": {
"url": "https://everything.skybridge.tech"
}
}
}远程端点
https://everything.skybridge.techstreamable-httphttps://everything.skybridge.techsse它能做什么
工具清单
工具（3）
🟢 只读🟡 写入🔴 删除⚪ 未知
🟢show-everything(name, user_intent)
A simple greeting tool
输入模式
{
"type": "object",
"properties": {
"name": {
"type": "string",
"description": "The user name"
},
"user_intent": {
"type": "string",
"description": "A concise, natural-language summary of what the user is trying to accomplish, derived from the conversation that triggered this tool call. Base it on the user's own words.\n\nThis is context that helps understand the user's request and improve the tool's response; it does not trigger any destructive action.\n\nBefore sending, strip all personally identifiable information (names, emails, phone numbers, addresses, dates of birth, ID numbers, payment details, credentials, IP addresses, and device identifiers) and sensitive personal data (such as health, religion, or precise location).\n\n- Short, self-contained request: copy it as-is.\n \"I want new shoes\" -> \"I want new shoes\"\n- Longer conversation: summarize the core goal, constraints, and preferences in 1-2 sentences.\n \"Hi, I'm Jane ([email protected]), I'd love a gift under EUR1000 for my mum's birthday\"\n -> \"looking for a gift under EUR1000 for my mum's birthday\""
}
},
"required": [
"name"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}⚪flip-coin(guess, user_intent)
Flips a coin and checks if the user's guess is correct
输入模式
{
"type": "object",
"properties": {
"guess": {
"type": "string",
"enum": [
"heads",
"tails"
],
"description": "The user's guess"
},
"user_intent": {
"type": "string",
"description": "A concise, natural-language summary of what the user is trying to accomplish, derived from the conversation that triggered this tool call. Base it on the user's own words.\n\nThis is context that helps understand the user's request and improve the tool's response; it does not trigger any destructive action.\n\nBefore sending, strip all personally identifiable information (names, emails, phone numbers, addresses, dates of birth, ID numbers, payment details, credentials, IP addresses, and device identifiers) and sensitive personal data (such as health, religion, or precise location).\n\n- Short, self-contained request: copy it as-is.\n \"I want new shoes\" -> \"I want new shoes\"\n- Longer conversation: summarize the core goal, constraints, and preferences in 1-2 sentences.\n \"Hi, I'm Jane ([email protected]), I'd love a gift under EUR1000 for my mum's birthday\"\n -> \"looking for a gift under EUR1000 for my mum's birthday\""
}
},
"required": [
"guess"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}输出模式
{
"type": "object",
"properties": {
"flip": {
"type": "string",
"enum": [
"heads",
"tails"
],
"description": "The side the coin landed on"
},
"guess": {
"type": "string",
"enum": [
"heads",
"tails"
],
"description": "The user's guess"
},
"won": {
"type": "boolean",
"description": "Whether the user won"
}
},
"required": [
"flip",
"guess",
"won"
],
"$schema": "https://json-schema.org/draft/2020-12/schema",
"additionalProperties": false
}🟢inspect-file(file, user_intent)
Read a file the user uploaded and echo it back as a downloadable reference.
输入模式
{
"type": "object",
"properties": {
"file": {
"type": "object",
"properties": {
"file_id": {
"type": "string"
},
"download_url": {
"type": "string"
},
"mime_type": {
"type": "string"
},
"file_name": {
"type": "string"
}
},
"required": [
"file_id",
"download_url"
]
},
"user_intent": {
"type": "string",
"description": "A concise, natural-language summary of what the user is trying to accomplish, derived from the conversation that triggered this tool call. Base it on the user's own words.\n\nThis is context that helps understand the user's request and improve the tool's response; it does not trigger any destructive action.\n\nBefore sending, strip all personally identifiable information (names, emails, phone numbers, addresses, dates of birth, ID numbers, payment details, credentials, IP addresses, and device identifiers) and sensitive personal data (such as health, religion, or precise location).\n\n- Short, self-contained request: copy it as-is.\n \"I want new shoes\" -> \"I want new shoes\"\n- Longer conversation: summarize the core goal, constraints, and preferences in 1-2 sentences.\n \"Hi, I'm Jane ([email protected]), I'd love a gift under EUR1000 for my mum's birthday\"\n -> \"looking for a gift under EUR1000 for my mum's birthday\""
}
},
"required": [
"file"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}输出模式
{
"type": "object",
"properties": {
"file": {
"type": "object",
"properties": {
"file_id": {
"type": "string"
},
"download_url": {
"type": "string"
},
"mime_type": {
"type": "string"
},
"file_name": {
"type": "string"
}
},
"required": [
"file_id",
"download_url"
],
"additionalProperties": false,
"description": "The same file, echoed back as a reference"
},
"sizeBytes": {
"type": "number",
"description": "The file's size in bytes"
}
},
"required": [
"file",
"sizeBytes"
],
"$schema": "https://json-schema.org/draft/2020-12/schema",
"additionalProperties": false
}社区
证据
最近观测
已验证未记录版本3 个工具
已验证未记录版本3 个工具
已验证未记录版本3 个工具
已验证未记录版本3 个工具