irrational

Adversarial behavioural-bias engine — audits your decisions for cognitive biases via your own AI.

我该使用它吗

质量与安全性

A
描述质量
93%
模式完整度
83%
命名质量
100%
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

~348token 数(工具定义)
~1.0 KB典型响应大小
对注意力的影响极小(占 128k 上下文窗口的 0.27%)

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": {
    "irrational": {
      "url": "https://irrational.pages.dev/mcp/server"
    }
  }
}

远程端点

https://irrational.pages.dev/mcp/serverstreamable-http

它能做什么

工具清单

工具(3)

🟢 只读🟡 写入🔴 删除⚪ 未知
🟢analyze_decision(judgment, reasoning, mode, language, structured)

Adversarially audit a decision for cognitive biases. Returns a directive YOUR model executes to produce the composed audit (verdict-first). Provide reasoning, not just the conclusion.

输入模式

{
  "type": "object",
  "properties": {
    "judgment": {
      "type": "string",
      "description": "The decision/judgment in one line."
    },
    "reasoning": {
      "type": "string",
      "description": "How you arrived at it (required to audit)."
    },
    "mode": {
      "type": "string",
      "enum": [
        "forward",
        "retrospective"
      ],
      "description": "forward = a decision you are about to make; retrospective = reviewing a past decision/outcome."
    },
    "language": {
      "type": "string",
      "description": "Optional. Natural language for the audit prose (e.g. \"Tamil\", \"Spanish\"). Bias ids stay canonical English so the result is still parseable. Defaults to English."
    },
    "structured": {
      "type": "boolean",
      "description": "Optional. Default false → the audit comes back as readable prose. Set true to get a machine-parseable JSON object (bias ids/keys in English) for pipelines that store or compare audits."
    }
  },
  "required": [
    "judgment"
  ]
}
🟢list_biases(family)

List the 22-bias catalogue, optionally filtered by family.

输入模式

{
  "type": "object",
  "properties": {
    "family": {
      "type": "string",
      "enum": [
        "too-much-information",
        "not-enough-meaning",
        "need-to-act-fast",
        "what-we-remember"
      ]
    }
  }
}
🟢get_bias(id)

Get the full entry for one bias by id.

输入模式

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

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