BrunoSan Intelligence Concierge
Discovery MCP for 11 intelligence domains plus LENS with live tools, routing and endpoints.
我该使用它吗
质量与安全性
发现(1)
- LOW在 get_connection_info 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": {
"intelligence": {
"url": "https://brunosan.de/mcp"
}
}
}远程端点
https://brunosan.de/mcpstreamable-http它能做什么
工具清单
工具(9)
⚪discover_brunosan
Start here: understand the complete BrunoSan Intelligence network. Returns every intelligence domain, the LENS cross-domain layer, canonical MCP endpoints, human documentation, evidence doctrine, current access summary and concrete next actions. This is intentionally richer than a directory so an agent can decide whether BrunoSan can satisfy its goal.
输入模式
{
"type": "object",
"properties": {},
"title": "discover_brunosanArguments"
}🟡list_intelligence_domains(include_live_tool_counts)
List all BrunoSan intelligence authorities. By default this is instant and SSOT-only. Set include_live_tool_counts=true when the agent needs proof that the underlying MCP servers are reachable now.
输入模式
{
"type": "object",
"properties": {
"include_live_tool_counts": {
"default": false,
"description": "If true, probe each canonical vertical MCP and include current tool counts.",
"title": "Include Live Tool Counts",
"type": "boolean"
}
},
"title": "list_intelligence_domainsArguments"
}🟢inspect_domain(domain, include_tools)
Open one BrunoSan vertical as a complete capability object. Returns purpose, canonical endpoint, human page, related routing vocabulary, disclaimer and—by default—the LIVE tools/list inventory from that vertical.
输入模式
{
"type": "object",
"properties": {
"domain": {
"description": "Canonical domain id or name, e.g. cyber, arxiv, AI News, regulatory, lens.",
"title": "Domain",
"type": "string"
},
"include_tools": {
"default": true,
"description": "Include the live tools/list inventory and input schemas from the vertical MCP.",
"title": "Include Tools",
"type": "boolean"
}
},
"required": [
"domain"
],
"title": "inspect_domainArguments"
}🟢search_capabilities(query, limit)
Search the whole BrunoSan network for what an agent is trying to do. Ranking is deterministic over SSOT descriptions, a documented routing vocabulary and LIVE tool names/descriptions/schemas. Returns concrete tool candidates, not merely domain names.
输入模式
{
"type": "object",
"properties": {
"query": {
"description": "Natural-language goal or capability, e.g. 'critical CVEs', 'citation network', 'EU AI Act sanctions'.",
"minLength": 2,
"title": "Query",
"type": "string"
},
"limit": {
"default": 8,
"maximum": 12,
"minimum": 1,
"title": "Limit",
"type": "integer"
}
},
"required": [
"query"
],
"title": "search_capabilitiesArguments"
}🟢inspect_tool(tool_name, domain)
Find and inspect one underlying vertical MCP tool. Returns the owning domain, exact live tool description and input schema. This does not execute the tool and never requests or forwards an API key.
输入模式
{
"type": "object",
"properties": {
"tool_name": {
"description": "Exact or partial MCP tool name to inspect.",
"minLength": 1,
"title": "Tool Name",
"type": "string"
},
"domain": {
"default": "",
"description": "Optional BrunoSan domain id/name to restrict the lookup.",
"title": "Domain",
"type": "string"
}
},
"required": [
"tool_name"
],
"title": "inspect_toolArguments"
}⚪plan_intelligence_route(goal, max_domains)
Plan a source-authority route for a complex intelligence goal. The plan is deterministic and names concrete live tools where possible. If several specialized domains are useful, LENS is recommended as the cross-domain entity layer. V1 plans and connects; it does not proxy protected vertical calls.
输入模式
{
"type": "object",
"properties": {
"goal": {
"description": "Research or monitoring goal that may require one or several BrunoSan verticals.",
"minLength": 3,
"title": "Goal",
"type": "string"
},
"max_domains": {
"default": 3,
"maximum": 5,
"minimum": 1,
"title": "Max Domains",
"type": "integer"
}
},
"required": [
"goal"
],
"title": "plan_intelligence_routeArguments"
}🟢get_connection_info(domain)
Return canonical connection details without guessing client-specific config.
输入模式
{
"type": "object",
"properties": {
"domain": {
"description": "BrunoSan domain id/name whose canonical MCP connection should be returned.",
"title": "Domain",
"type": "string"
}
},
"required": [
"domain"
],
"title": "get_connection_infoArguments"
}🟢get_intelligence_status(refresh)
Check whether each canonical BrunoSan MCP currently initializes and lists tools. This is a capability-plane health check only; it does not execute protected intelligence tools or make claims about underlying data freshness.
输入模式
{
"type": "object",
"properties": {
"refresh": {
"default": false,
"description": "Request a fresh tools/list probe. Normal calls use the short-lived cache to protect vertical MCPs.",
"title": "Refresh",
"type": "boolean"
}
},
"title": "get_intelligence_statusArguments"
}🟢get_access_options
Return current BrunoSan access and pricing from the shared pricing manifest. No price is embedded in this server. Unknown or absent manifest values stay absent rather than being estimated.
输入模式
{
"type": "object",
"properties": {},
"title": "get_access_optionsArguments"
}社区
证据