Invokera Discover
Search, inspect and invoke every public tool on Invokera through one MCP connection.
我该使用它吗
质量与安全性
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": {
"discover": {
"url": "https://invokera.com/discover"
}
}
}远程端点
https://invokera.com/discoverstreamable-http它能做什么
工具清单
工具（3）
🟢search_tools(query, category, limit)
Search the Invokera catalog for published tools across all projects. Returns a ranked list of tools (name, project, summary, category) WITHOUT their input schemas. Call describe_tool next to get the full schema and calling instructions for a specific tool.
输入模式
{
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "Keywords, e.g. \"housing price\" or \"stock signal\". Empty lists popular tools."
},
"category": {
"type": "string",
"description": "Optional category filter.",
"enum": [
"other",
"data",
"commerce",
"productivity",
"social",
"finance"
]
},
"limit": {
"type": "integer",
"description": "Max results, 1-25 (default 10).",
"minimum": 1,
"maximum": 25
}
}
}🟢describe_tool(project_slug, tool_name)
Get the full input schema and calling instructions for one tool found via search_tools. Returns the JSON Schema plus the exact REST/MCP endpoint to call it on. Calling a tool requires that project's own endpoint token, obtained from its owner.
输入模式
{
"type": "object",
"properties": {
"project_slug": {
"type": "string",
"description": "The projectSlug from a search_tools result."
},
"tool_name": {
"type": "string",
"description": "The toolName from a search_tools result."
}
},
"required": [
"project_slug",
"tool_name"
]
}🟢invoke_tool(project_slug, tool_name, arguments)
Call a tool whose owner has enabled public invocation, without needing that project's token. Only works for tools opted in by their owner; others return not-found (use describe_tool for how to call them with a token). Usage is metered to the tool owner. Get project_slug/tool_name and the argument schema from search_tools + describe_tool first.
输入模式
{
"type": "object",
"properties": {
"project_slug": {
"type": "string",
"description": "The projectSlug from a search_tools result."
},
"tool_name": {
"type": "string",
"description": "The toolName from a search_tools result."
},
"arguments": {
"type": "object",
"description": "Arguments matching the tool inputSchema from describe_tool."
}
},
"required": [
"project_slug",
"tool_name"
]
}社区
证据