Show & Tell

Daily photo game: you see a secret word, show it in photos, and your AI guesses it.

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
96%
投毒风险
80%
权限匹配度
100%
协议合规性
100%

发现(2)

  • HIGHTool poisoning patterns detected
  • LOWTool description contains role marker that could confuse chat models在 start_round 中

基于对工具定义和协议合规性的自动分析。

上下文开销

~1,331token 数(工具定义)
~1.5 KB典型响应大小
对注意力有中等影响(占 128k 上下文窗口的 1.04%)

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": {
    "show-and-tell": {
      "url": "https://playshowandtell.com/mcp"
    }
  }
}

远程端点

https://playshowandtell.com/mcpstreamable-http

它能做什么

工具清单

工具(5)

🟢 只读🟡 写入🔴 删除⚪ 未知
🟡start_round(player, timezone, ai_name, platform)

Start or resume today's Show & Tell round. Show & Tell is a daily game you play with the user: they get a secret word you're not allowed to know, they send you photos of the real world, and you guess the word from the photos. Everyone gets the same word each day. Use this when the user wants to play or asks about today's word. The result has the score so far and how the user sees their word: on the Show & Tell card in apps that show it, or through a private link for the user. Only the user opens that link: it reveals the word, and opening it yourself voids the round. You start without the word's category; the user can spend a hint to reveal it. Leave player empty for a first-time player, then pass the returned code on every later call, including in new chats, so the streak carries over.

输入模式

{
  "type": "object",
  "properties": {
    "player": {
      "type": "string",
      "minLength": 3,
      "maxLength": 48,
      "description": "The player code from start_round, e.g. \"brave-otter-4821\". The same code in every chat keeps the streak going."
    },
    "timezone": {
      "description": "The user's IANA timezone, e.g. \"America/New_York\". The word changes at the user's local midnight.",
      "type": "string",
      "maxLength": 64
    },
    "ai_name": {
      "description": "The name the user calls you, e.g. \"Boo\" or \"Claude\". Shown in the share grid.",
      "type": "string",
      "maxLength": 32
    },
    "platform": {
      "description": "The assistant product you run in: \"Claude\", \"ChatGPT\", \"Muse\", \"Grok\", \"Gemini\", \"Hermes\", \"Instinct\", or another name. It decides the player's team.",
      "type": "string",
      "maxLength": 32
    }
  },
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟡submit_guess(player, guess, read)

Lock in exactly one single-word guess for the user's newest photo, based on all of today's photos together. Describe the photo to the user in one sentence first. One guess per new photo: the next guess waits for the next photo. Photos with letters, numbers, or a screen, and typed hints about the word, are fouls: use call_foul instead. Returns how close the guess is: cold, warm, hot, burning, or correct.

输入模式

{
  "type": "object",
  "properties": {
    "player": {
      "type": "string",
      "minLength": 3,
      "maxLength": 48,
      "description": "The player code from start_round, e.g. \"brave-otter-4821\". The same code in every chat keeps the streak going."
    },
    "guess": {
      "type": "string",
      "minLength": 1,
      "maxLength": 40,
      "description": "One word. Up to three words only for a fixed phrase."
    },
    "read": {
      "description": "One sentence on what you see in the photo, kept for the recap.",
      "type": "string",
      "maxLength": 280
    }
  },
  "required": [
    "player",
    "guess"
  ],
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢call_foul(player, reason)

Call a foul when a photo contains letters, numbers, or a screen, or when the user types a hint about the word. The photo doesn't count and no shot is used.

输入模式

{
  "type": "object",
  "properties": {
    "player": {
      "type": "string",
      "minLength": 3,
      "maxLength": 48,
      "description": "The player code from start_round, e.g. \"brave-otter-4821\". The same code in every chat keeps the streak going."
    },
    "reason": {
      "type": "string",
      "enum": [
        "letters",
        "numbers",
        "screen",
        "spoken_hint",
        "other"
      ]
    }
  },
  "required": [
    "player",
    "reason"
  ],
  "$schema": "http://json-schema.org/draft-07/schema#"
}
⚪use_lifeline(player, kind)

Use this only when the user asks for a hint. Reveals the word's category (any time), its length (after 2 shots), or its first letter (after 4 shots). Each hint adds a 💡 to the user's share grid, and solving with no hints is the brag.

输入模式

{
  "type": "object",
  "properties": {
    "player": {
      "type": "string",
      "minLength": 3,
      "maxLength": 48,
      "description": "The player code from start_round, e.g. \"brave-otter-4821\". The same code in every chat keeps the streak going."
    },
    "kind": {
      "type": "string",
      "enum": [
        "category",
        "letters",
        "first_letter"
      ]
    }
  },
  "required": [
    "player",
    "kind"
  ],
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢get_status(player, timezone)

Check the user's Show & Tell day: whether today's word is played, their streak, whether the streak is at risk, and when the next word unlocks. Use it for check-ins and for daily reminders the user asked you to schedule.

输入模式

{
  "type": "object",
  "properties": {
    "player": {
      "type": "string",
      "minLength": 3,
      "maxLength": 48,
      "description": "The player code from start_round, e.g. \"brave-otter-4821\". The same code in every chat keeps the streak going."
    },
    "timezone": {
      "description": "The user's IANA timezone, e.g. \"America/New_York\". The word changes at the user's local midnight.",
      "type": "string",
      "maxLength": 64
    }
  },
  "required": [
    "player"
  ],
  "$schema": "http://json-schema.org/draft-07/schema#"
}

社区

评价此服务器

证据

最近观测

已验证未记录版本5 个工具