Matchmaking rating for agent selection
Calibrated Glicko-2 per-dimension agent rating with deviation, from verified receipts only.
我该使用它吗
质量与安全性
B
发现（2）
- LOWTool 'rank_agents' description lacks action verb在 rank_agents 中
- LOWTool 'ingest_receipt' description lacks action verb在 ingest_receipt 中
基于对工具定义和协议合规性的自动分析。
上下文开销
~166token 数（工具定义）
~394 B典型响应大小
对注意力的影响极小（占 128k 上下文窗口的 0.13%）
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": {
"matchmaking-rating-for-agent-selection": {
"url": "https://textforagents.com/products/matchmaking-rating-for-agent-selection/mcp"
}
}
}远程端点
https://textforagents.com/products/matchmaking-rating-for-agent-selection/mcpstreamable-http它能做什么
工具清单
工具（3）
🟢 只读🟡 写入🔴 删除⚪ 未知
🟢query_rating(agent, dimension, task_rating)
Pre-delegation query $0.01 (trial applies)
输入模式
{
"type": "object",
"properties": {
"agent": {
"type": "string"
},
"dimension": {
"type": "string"
},
"task_rating": {
"type": "number"
}
}
}⚪rank_agents(dimension, candidates, mode, consequence, prov_budget, ...)
Rank candidates $0.01 (trial applies)
输入模式
{
"type": "object",
"properties": {
"dimension": {
"type": "string"
},
"candidates": {
"type": "array"
},
"mode": {
"type": "string"
},
"consequence": {
"type": "string"
},
"prov_budget": {
"type": "integer"
},
"task_rating": {
"type": "number"
}
}
}⚪ingest_receipt(receipt)
Ingest verified receipt (verifier-signed)
输入模式
{
"type": "object",
"properties": {
"receipt": {
"type": "object"
}
}
}社区
证据
最近观测
已验证未记录版本3 个工具