Nankai Trough

io.github.mrslbt由 GitHub 帳號 mrslbt 發布v0.2.1更新於 Oct 24, 2026

Nankai Trough earthquake hazard & building-safety, from official Japan data. No verdicts.

已驗證STDIO僅桌面位置與生活資料與分析

概覽

AI 產生的概覽

本機 MCP 伺服器,讓助理取得附出處的日本官方南海海槽地震災害資料,並提供建築耐震標準與防災準備建議。

可以做什麼
  • 概述 2025 年官方南海海槽情境:發生機率、傷亡、震度 7 範圍與海嘯,皆附來源與日期
  • 依日本地址或 7 位郵遞區號查詢官方逐地址災害風險圖與 J-SHIS 30 年震度機率
  • 依建築年份與結構判定耐震標準世代,並導向補助的耐震診斷與補強計畫
  • 說明日本氣象廳震度等級,並產出有出處的家庭防災準備計畫
  • 對日本地址進行地理編碼,並讀取來源清單、主要數據、震度等級與建築標準等參考資源
適用情境
適合有人詢問南海海槽地震的規模與波及範圍、某個日本地址的官方災害風險圖,或某年建造的房屋屬於舊耐震還是新耐震標準時使用。對象是希望取得有出處數據、而非安心保證的居民與屋主。
執行需求
透過 npx(需 Node.js)或從原始碼建置後在本機以 stdio 執行;不需 API 金鑰或環境變數。地址工具最好使用日文地址或 7 位郵遞區號,並需連網存取官方地理編碼與資料來源。
安裝前請注意
它刻意只報告官方數據,不判定任何人是否安全;建築分類是依你提供的建造年份與結構推算,並非逐棟建築紀錄。此伺服器為非官方專案,與任何政府機關無關,不能取代官方避難指引或專業耐震診斷。郵遞區號查詢定位到郵遞區中心,而非特定建築。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Nankai Trough,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Desktop only,透過 STDIO。 STDIO 服務會啟動本機處理程序,因此需要 SourceWeft 桌面主機。

其他 MCP 客戶端

參照 儲存庫 中的啟動說明。

README

nankai-trough-mcp

[License: MIT]

Nankai Trough (南海トラフ地震) earthquake hazard and building-safety engine for AI assistants. Ask "how bad is the Nankai Trough quake, how far does it reach, and what does that mean for a house built in 1978?" and get official Japanese government figures, every one cited and dated, with a route to the official maps for your exact address. It never tells you you're "safe."

It reports what 内閣府 (Cabinet Office), 地震本部, 気象庁, and 国土交通省 actually publish, and where a figure isn't verifiable, it points you at the source instead of inventing one. Bilingual EN/JA. Runs locally, no API keys.

Live companion map: https://nankaitrough.bymarsel.me (this is the open data engine behind it, the same official data as an explorable map).

⚠️ Not a prediction, not a verdict. This explains official data and routes you to official guidance and a professional seismic diagnosis (耐震診断). Always confirm with your municipality and official hazard maps. In an emergency, follow official evacuation instructions.

Why it exists

Most people underestimate the reach. The 30-year probability was revised by 地震本部 in September 2025 away from the old single "80%" to a two-model range (60–90%+ (higher model) / 20–50%). Intensity-7 shaking is projected across 10 prefectures, and 764 municipalities in 31 prefectures face major shaking or a 3 m+ tsunami. "I'm inland, so I'm fine" is exactly the assumption this corrects.

The authoritative data is real, but it's scattered across five agencies and mostly buried in PDFs. This server makes it cited and usable from an AI assistant, without ever inventing a number.

The attitude: assume the worst

The earthquake is unknown, and claiming we know is an illusion. So the refusal here is one-directional: the tools will alarm you with official numbers, and they will never reassure you.

Where the official model publishes several cases, the severe one leads. A low probability is reported as a floor, not a ceiling, and is never relayed as good news. location_probability will tell you the government puts your mesh at a 12% chance of 震度6強 in 30 years; it will not tell you that means you are fine.

If you cannot write Japanese

The official geocoder only reads Japanese addresses, which would have excluded the people this server is for. Give a 7-digit postal code instead and every address-taking tool resolves it through Japan Post data:

232-0063        2320063        〒232-0063

That lands on the centre of your postal district rather than your building, and the tools say so in the response rather than pretending to more precision than they have. All output is available in English with language: "en".

Data rules

  • No verdicts. It reports official figures + plain-language meaning, never "your home is safe/unsafe."
  • Every figure cites its official source + date. If a value isn't verifiable in a primary source, it points to the source instead of guessing.
  • Probabilistic ≠ scenario. J-SHIS (all-source probability) is never presented as the Nankai scenario (Cabinet Office).
  • Building safety can't be looked up. It's classified from the build year + structure you provide. No fabricated per-building data.

Install

bash
claude mcp add nankai -- npx -y nankai-trough-mcp

JSON config

The server config is the same across clients; only the file path differs.

Claude Code: ~/.claude/.mcp.json
json
{  "mcpServers": {    "nankai": { "command": "npx", "args": ["-y", "nankai-trough-mcp"] }  }}
Cursor: ~/.cursor/mcp.json
json
{  "mcpServers": {    "nankai": { "command": "npx", "args": ["-y", "nankai-trough-mcp"] }  }}
VS Code (GitHub Copilot): .vscode/mcp.json
json
{  "servers": {    "nankai": { "type": "stdio", "command": "npx", "args": ["-y", "nankai-trough-mcp"] }  }}
Claude Desktop: claude_desktop_config.json
json
{  "mcpServers": {    "nankai": { "command": "npx", "args": ["-y", "nankai-trough-mcp"] }  }}

Config path:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json

From source

bash
git clone https://github.com/mrslbt/nankai-trough-mcp.gitcd nankai-trough-mcpnpm install && npm run build

Point the client config at the built entry:

json
{ "command": "node", "args": ["/absolute/path/to/nankai-trough-mcp/dist/index.js"] }

Tools

ToolDescription
nankai_overviewThe official 2025 scenario's scale and reach (probability, casualties, intensity-7 reach, tsunami), each with its source. Start here.
official_hazard_mapsAddress → links to the official per-address hazard maps, where the exact predicted intensity and tsunami live. This server bridges; it doesn't invent those numbers.
building_seismic_checkBuild year + structure → 旧耐震 / 新耐震 / 2000-standard classification with context. Not a verdict.
taishin_subsidy_guideRoutes to subsidised, often-free 耐震診断 / 耐震補強 programs and the national support framework.
shindo_meaningWhat a JMA intensity (震度 5弱–7) means for people and buildings, on the official 気象庁 scale.
location_probabilityAddress → the official J-SHIS 30-year probability of reaching each JMA intensity at that ~250m mesh, severe case first. All-source probabilistic (地震本部), not the Nankai scenario. Reported, never computed.
preparedness_planHousehold details → a sourced plan: stockpile sized to the official one-week guidance, what to do now, what to do when it shakes, and the structural route. An officially recommended floor, not a verdict on survival.
geocode_addressAddress → coordinates via the GSI geocoder. Utility.

Prompts: assess_home_earthquake_risk, nankai_briefing.

Resources

Read-only reference documents a client can list and read without a tool call. Each is rendered from the same data the tools use, so they never drift out of sync.

ResourceContent
nankai://sourcesThe official source registry: every agency cited, with URL, plus the standing disclaimer.
nankai://headline-figuresThe 2025 scenario's headline figures, each with source, as-of date, and notes.
nankai://shindo-scaleThe JMA intensity scale (震度 5弱–7): what each level means for people and buildings.
nankai://building-standardsThe 1981 新耐震 / 2000 wooden-house boundaries + 2016 Kumamoto field-damage data.

Example prompts

Brief me on the Nankai Trough earthquake: the scale, the reach, and whatintensity 7 actually means for a building.
My house is in 高知市本町. Give me the official hazard maps for my address'spredicted shaking and tsunami.
I live in a wooden house built in 1978. What seismic standard is that under,and what should I do about it? Don't tell me whether it's safe.
Walk me through assessing my home's earthquake risk:静岡市葵区追手町9-6, built 1990, reinforced concrete.

Scope

This server does not compute its own per-address values. Two different things follow from that:

  • It reports official published per-mesh probabilities (location_probability, J-SHIS Y2024). Surfacing a government figure is not inventing one.
  • It bridges to the official maps for the Nankai scenario at an address (official_hazard_maps). That data lives in bulk Cabinet Office GIS files needing an ingestion pipeline, and faking it is not on the table.

Headline figures (and where they come from)

Every figure below is verified against the primary government source and carries an as-of date in the tool output. See src/data/nankai.ts for the full verification log.

FigureValueSource
30-year probabilityTwo models since the 2025-09-26 revision: 60–90%+ / 20–50%. Act on the higher.地震本部
Worst-case deaths~298,000 (upper bound, flood defenses functioning)内閣府 2025-03
Economic loss~¥270兆 (¥224.9兆 direct + ¥45.4兆 production); ~¥292兆 incl. transport disruption内閣府 2025-03
Intensity 7Across 10 prefectures内閣府 2025-03
Reach764 municipalities in 31 prefectures at 震度6弱以上 or 津波3m以上内閣府 2025-03
Tsunami5 m+ within minutes on the fastest coasts; max ~34 m (高知), ~31 m (静岡)内閣府 / 気象庁

Sources & license

Official data is used as derived output with attribution; raw government files are never re-hosted.

出典:内閣府 南海トラフの巨大地震モデル・被害想定検討会 / 地震調査研究推進本部 / 気象庁 / 国土交通省(住まいの耐震化)/ 国土地理院(地理院タイル・ジオコーディング)/ 住宅リフォーム推進協議会。

日本語

南海トラフ地震のハザードと建物の耐震性を、公的データだけで扱うMCPサーバーです。内閣府・地震本部・気象庁・国土交通省の公表値を、出典と日付つきで提示し、あなたの住所の正確な想定震度・津波は公式ハザードマップへ橋渡しします。建物の安全性は「安全/危険」と判定せず、建築年と構造から耐震基準の世代を示し、耐震診断・補助制度へ案内します。ローカル動作、APIキー不要。

Disclaimer

This server surfaces official Japanese government data and explanations. It is not a prediction, not a verdict on whether you or your home are safe, and not a substitute for official evacuation guidance or a professional seismic diagnosis (耐震診断). Figures are approximate and as reported by their official sources; always confirm at the source. Unofficial, not affiliated with any government agency.

License

MIT


More MCPs

MCPWhat it does
Japan DesignEnforces Japanese UI + UX conventions on real CSS/markup
ripprYouTube transcript ripper for humans and AI agents
RakutenSearch Rakuten's marketplace, books, and hotels

Built by Marsel Bait in Tokyo

來源:README.md,提交 cd1cea6

工具

0
SourceWeft 尚未在沙箱中啟動過該軟體包,因此其工具未知。

版本歷史

2
  1. v0.2.1最新Oct 6, 2026
  2. v0.1.1Sep 16, 2026

評分

已滿足適用檢查項的 50%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通25 分

    請保持端點能夠回應 MCP 握手;SourceWeft 每 30 天探測一次。

  • 已宣告認證方式10 分

    請在註冊中心條目中宣告憑證請求標頭,或保持端點能夠回應探測,以便判定其認證方式。

  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list,或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分