Agentic Diaries Audit
Checks what a support agent knew but did not say against its policies. Free tier, no login.
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
~182token 数（工具定义）
~981 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": {
"behavioral-audit": {
"url": "https://receipts-mcp.vercel.app/api/mcp"
}
}
}远程端点
https://receipts-mcp.vercel.app/api/mcpstreamable-http它能做什么
工具清单
工具（1）
🟢 只读🟡 写入🔴 删除⚪ 未知
🟢disclosure_check(customer_message, draft_reply)
Check a draft support reply for disclosure gaps before sending it. Given what the customer said and the reply you are about to send, it flags any should-know policy the customer's situation made relevant that the draft failed to proactively surface (the knew-but-did-not-say gap). Returns a verdict (pass or gap); on a gap, each missed policy and the line that should have been surfaced. It detects and suggests, it does not rewrite.
输入模式
{
"type": "object",
"properties": {
"customer_message": {
"type": "string",
"description": "What the customer said, verbatim."
},
"draft_reply": {
"type": "string",
"description": "The reply you are about to send."
}
},
"required": [
"customer_message",
"draft_reply"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}社区
证据
最近观测
已验证未记录版本1 个工具
已验证未记录版本1 个工具
已验证未记录版本1 个工具