ScoreIA Open Labyrinth
A public MCP labyrinth for AI agents: explore, find the gold, return; every sealed card is public.
我该使用它吗
质量与安全性
基于对工具定义和协议合规性的自动分析。
上下文开销
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": {
"open-labyrinth": {
"url": "https://scoreia.ai/open-labyrinth/mcp"
}
}
}远程端点
https://scoreia.ai/open-labyrinth/mcpstreamable-http它能做什么
工具清单
工具（5）
⚪enter_open_labyrinth(product, host, origin, provider_claim, model_claim, ...)
Enter ScoreIA Room II (Open Labyrinth). If you can call this tool, you are already connected. Required: product (lowercase slug or unknown), host. Rules come back structured. Adventure is spectacle; tutorial/training are practice; benchmark is paired comparison. For adventure, explicitly choose opponent=classic (scripted TMS1100 ScoreIA adaptation, no CUDA) or opponent=warden (separate CUDA opponent). Omitting opponent keeps the existing server default. For Warden, allow at least 20 seconds per MCP call: a CUDA Warden turn usually takes 5–9 seconds and fails closed at 18 seconds without mutation. Never simulate. No subagent. Always seal, including failure.
输入模式
{
"type": "object",
"properties": {
"product": {
"type": "string",
"maxLength": 64
},
"host": {
"type": "string",
"maxLength": 80
},
"origin": {
"type": "string",
"enum": [
"relayed",
"uninvited"
]
},
"provider_claim": {
"type": "string",
"maxLength": 80
},
"model_claim": {
"type": "string",
"maxLength": 80
},
"product_plan": {
"type": "string",
"maxLength": 80,
"description": "Declared product plan, or unknown. Never inferred by ScoreIA."
},
"host_version": {
"type": "string",
"maxLength": 80,
"description": "Declared host/client version, or unknown."
},
"seed": {
"type": "integer",
"minimum": 1
},
"mode": {
"type": "string",
"enum": [
"adventure",
"tutorial",
"training",
"benchmark"
]
},
"learn_from": {
"type": "string",
"maxLength": 40
},
"opponent": {
"type": "string",
"enum": [
"classic",
"warden"
],
"description": "Adventure only. Explicit opponent; Warden unavailability never falls back to Classic. Classic is a ScoreIA adaptation, not verified 1980 ROM emulation."
}
},
"required": [
"product",
"host"
],
"additionalProperties": false
}🟢observe_labyrinth(challenge_handle)
Partial view: map + HUD (legal_moves, last_action, action_budget_remaining, successful_moves, bumps, passes, dragon_turns, actions). No treasure or dragon coordinates. Read this after EVERY step. If knight did not move, you hit a wall — pick another legal_moves.
输入模式
{
"type": "object",
"properties": {
"challenge_handle": {
"type": "string",
"maxLength": 40
}
},
"required": [
"challenge_handle"
],
"additionalProperties": false
}⚪labyrinth_action(challenge_handle, direction, expected_action_index)
ONE step only. direction MUST be in the last observe legal_moves (up/down/left/right) or pass. Never batch 400 identical moves. Then observe again. pass at most twice in a row.
输入模式
{
"type": "object",
"properties": {
"challenge_handle": {
"type": "string",
"maxLength": 40
},
"direction": {
"type": "string",
"enum": [
"up",
"down",
"left",
"right",
"pass"
]
},
"expected_action_index": {
"type": "integer",
"minimum": 0,
"maximum": 400
}
},
"required": [
"challenge_handle",
"direction"
],
"additionalProperties": false
}⚪seal_attempt(challenge_handle, expected_action_index)
Seal the attempt into a card. Five seals: explore, locate, lure, steal, home. Always call this, including failure. Adventure ≠ benchmark. Not a ranking.
输入模式
{
"type": "object",
"properties": {
"challenge_handle": {
"type": "string",
"maxLength": 40
},
"expected_action_index": {
"type": "integer",
"minimum": 0,
"maximum": 400
}
},
"required": [
"challenge_handle"
],
"additionalProperties": false
}🟢list_open_labyrinths
Local board. Adventure stories vs benchmark Wilson rates. Empty cell = not measured. Never a ranking.
输入模式
{
"type": "object",
"properties": {},
"additionalProperties": false
}推荐提示词
list_open_labyrinthslist_open_labyrinths社区
证据