maxwell-defense
Model proof-of-work policy and latency for AI agents. $1 USDC rehearsal; no runtime protection.
我该使用它吗
质量与安全性
B
发现（1）
- LOWTool 'describe_agent' description lacks action verb在 describe_agent 中
基于对工具定义和协议合规性的自动分析。
上下文开销
~253token 数（工具定义）
~1.2 KB典型响应大小
对注意力的影响极小（占 128k 上下文窗口的 0.20%）
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": {
"maxwell-defense": {
"url": "https://mcp.viridis-security.com/maxwell-defense/mcp"
}
}
}远程端点
https://mcp.viridis-security.com/maxwell-defense/mcpstreamable-http它能做什么
工具清单
工具（2）
🟢 只读🟡 写入🔴 删除⚪ 未知
⚪rehearse_defense(client_hashes_per_second, client_p95_budget_ms, backend_cost_ms, peak_requests_per_second)
$1 bounded policy rehearsal. Returns a model and local hash measurement. Does not activate endpoint protection or establish energy savings.
输入模式
{
"type": "object",
"properties": {
"client_hashes_per_second": {
"title": "Client Hashes Per Second",
"type": "integer"
},
"client_p95_budget_ms": {
"title": "Client P95 Budget Ms",
"type": "integer"
},
"backend_cost_ms": {
"title": "Backend Cost Ms",
"type": "integer"
},
"peak_requests_per_second": {
"title": "Peak Requests Per Second",
"type": "integer"
}
},
"required": [
"client_hashes_per_second",
"client_p95_budget_ms",
"backend_cost_ms",
"peak_requests_per_second"
],
"title": "rehearse_defenseArguments"
}输出模式
{
"type": "object",
"properties": {
"result": {
"title": "Result",
"type": "string"
}
},
"required": [
"result"
],
"title": "rehearse_defenseOutput"
}🟢describe_agent
Free scope and price discovery.
输入模式
{
"type": "object",
"properties": {},
"title": "describe_agentArguments"
}输出模式
{
"type": "object",
"properties": {
"result": {
"title": "Result",
"type": "string"
}
},
"required": [
"result"
],
"title": "describe_agentOutput"
}社区
证据
最近观测
已验证未记录版本2 个工具
已验证未记录版本2 个工具
已验证未记录版本2 个工具