Intent Cloak: query-intent privacy for agents
Query-intent privacy for agents: threshold cloaking with signed exposure receipts.
我该使用它吗
质量与安全性
A
基于对工具定义和协议合规性的自动分析。
上下文开销
~183token 数(工具定义)
~348 B典型响应大小
对注意力的影响极小(占 128k 上下文窗口的 0.14%)
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": {
"intent-cloak-query-privacy": {
"url": "https://textforagents.com/products/intent-cloak-query-privacy/mcp"
}
}
}远程端点
https://textforagents.com/products/intent-cloak-query-privacy/mcpstreamable-http它能做什么
工具清单
工具(3)
🟢 只读🟡 写入🔴 删除⚪ 未知
🟡cloak_run(task_id, task, provider_kind, cloaked)
Run a task through the Intent Cloak and return answers, utility, prices and the signed exposure receipt. Same billing as POST /cloak.
输入模式
{
"type": "object",
"properties": {
"task_id": {
"type": "string"
},
"task": {
"type": "object"
},
"provider_kind": {
"type": "string",
"enum": [
"exploitative",
"neutral"
]
},
"cloaked": {
"type": "boolean"
}
}
}⚪exposure_receipt(task_id)
Return the signed exposure receipt of your own last cloak_run for a task.
输入模式
{
"type": "object",
"properties": {
"task_id": {
"type": "string"
}
},
"required": [
"task_id"
]
}⚪verify_receipt(receipt)
Verify a signed receipt's HMAC signature (free).
输入模式
{
"type": "object",
"properties": {
"receipt": {
"type": "object"
}
},
"required": [
"receipt"
]
}社区
证据
最近观测
已验证未记录版本3 个工具