MCP MeasureMCP Measure

索引已观测

agent-trace-auditor

Cloudflare Workers MCP server: agent-trace-auditor

我该使用它吗

质量与安全性

A
描述质量
100%
模式完整度
94%
命名质量
87%
投毒风险
100%
权限匹配度
100%
协议合规性
100%

基于对工具定义和协议合规性的自动分析。

上下文开销

~364token 数（工具定义）
~932 B典型响应大小
对注意力的影响极小（占 128k 上下文窗口的 0.28%）

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": {
    "agent-trace-auditor": {
      "url": "https://api.lazy-mac.com/agent-trace-auditor/mcp"
    }
  }
}

远程端点

https://api.lazy-mac.com/agent-trace-auditor/mcpstreamable-http

它能做什么

工具清单

工具（3）

🟢 只读🟡 写入🔴 删除⚪ 未知
⚪audit_trace(trace, budget_usd, tool_schemas)

Full audit of an agent execution trace. Detects infinite loops, duplicate calls, tool errors, cost overruns, schema violations, and latency bottlenecks. Returns verdict (pass/review/block), score 0-100, and detailed findings.

输入模式

{
  "type": "object",
  "properties": {
    "trace": {
      "type": "array",
      "description": "Agent execution steps. Each step: {tool, input, output, model, ts, latency_ms, error}",
      "items": {
        "type": "object",
        "properties": {
          "tool": {
            "type": "string"
          },
          "input": {},
          "output": {},
          "model": {
            "type": "string"
          },
          "ts": {
            "type": "string"
          },
          "latency_ms": {
            "type": "number"
          },
          "error": {}
        }
      }
    },
    "budget_usd": {
      "type": "number",
      "description": "Cost cap in USD (default 5.0)"
    },
    "tool_schemas": {
      "type": "object",
      "description": "Per-tool required field schemas for input validation"
    }
  },
  "required": [
    "trace"
  ]
}
⚪quick_verdict(trace, budget_usd)

Fast pass/review/block verdict with score and issue counts. Use in CI gates.

输入模式

{
  "type": "object",
  "properties": {
    "trace": {
      "type": "array",
      "description": "Agent execution steps"
    },
    "budget_usd": {
      "type": "number"
    }
  },
  "required": [
    "trace"
  ]
}
⚪diff_traces(trace_a, trace_b, budget_usd)

Compare two agent traces to detect regressions. Returns score delta, cost delta, added/removed tools, new/resolved issue types.

输入模式

{
  "type": "object",
  "properties": {
    "trace_a": {
      "type": "array",
      "description": "Baseline (previous) trace"
    },
    "trace_b": {
      "type": "array",
      "description": "New trace to compare against baseline"
    },
    "budget_usd": {
      "type": "number"
    }
  },
  "required": [
    "trace_a",
    "trace_b"
  ]
}

社区

评价此服务器

证据

最近观测

已验证未记录版本3 个工具
已验证未记录版本3 个工具