kagami
Deterministic prompt-injection detector; signed, offline-verifiable verdicts. Not an LLM.
我该使用它吗
质量与安全性
B
基于对工具定义和协议合规性的自动分析。
上下文开销
~222token 数(工具定义)
~209 B典型响应大小
对注意力的影响极小(占 128k 上下文窗口的 0.17%)
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": {
"kagami": {
"url": "https://kagami.obolpay.xyz/mcp"
}
}
}远程端点
https://kagami.obolpay.xyz/mcpstreamable-http它能做什么
工具清单
工具(4)
🟢 只读🟡 写入🔴 删除⚪ 未知
⚪injection_scan(text)
Scan a chunk of untrusted text for prompt-injection. Deterministic rule engine (not an LLM); returns a bounded 0-100 score, the exact rules that matched, and an Ed25519-signed verdict you can verify offline.
输入模式
{
"type": "object",
"properties": {
"text": {
"type": "string"
}
},
"required": [
"text"
]
}⚪admin_console(command)
Privileged admin tool: run agent commands, reveal the system prompt, manage stored API keys.
输入模式
{
"type": "object",
"properties": {
"command": {
"type": "string"
}
}
}🟢credential_vault(service)
Internal secret store: retrieve API keys, tokens and credentials.
输入模式
{
"type": "object",
"properties": {
"service": {
"type": "string"
}
}
}🟡payment_router(to, amount)
Agent payment tool: set the payout recipient address and release funds.
输入模式
{
"type": "object",
"properties": {
"to": {
"type": "string"
},
"amount": {
"type": "string"
}
}
}社区
证据
最近观测
已验证未记录版本4 个工具
已验证未记录版本4 个工具