Fetch the full text / 条文 / chapter structure of a specific Japanese law from e-Gov 法令検索 API v2 (デジタル庁). Use after search-egov-laws to retrieve actual statute body, article-by-article. Returns structured law content (chapters / articles / paragraphs) with public 公式 citation. 日本語 keyword: 条文取得 / 法令本文 / e-Gov 法令データ / 法律全文取得 / law_id で条文取得.
Installs into .claude/skills of the current project.
Are you the author of Get Egov Laws Data?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/kokai-data-get-egov-laws-data)
---
name: get-egov-laws-data
description: Fetch the full text / 条文 / chapter structure of a specific Japanese law from e-Gov 法令検索 API v2 (デジタル庁). Use after search-egov-laws to retrieve actual statute body, article-by-article. Returns structured law content (chapters / articles / paragraphs) with public 公式 citation. 日本語 keyword: 条文取得 / 法令本文 / e-Gov 法令データ / 法律全文取得 / law_id で条文取得.
---
## When to use
Use this skill when:
- You have a `law_id` from a prior `search_egov_laws` call and need the actual statute text.
- The user asks "what does Article X of the Y Act say?" (e.g., "個人情報保護法 第15条").
- DD / legal research requires verbatim citation of statute text.
- You need to quote specific provisions in a due diligence brief or subsidy-fit analysis.
- Regulatory mapping requires the full chapter / article structure (法令の章立て).
## How to invoke
Call the kokai MCP server's `get_egov_laws_data` tool:
```json
{
"name": "get_egov_laws_data",
"arguments": {
"law_id": "<law_id from search_egov_laws, e.g., 338AC0000000154>"
}
}
```
`law_id` is required (obtained from `search_egov_laws` results).
## Output structure
- `law_id` (input echoed)
- `law_num` (法令番号)
- `title` (法令名)
- `enforcement_date` / `promulgation_date`
- `chapters` (array of 章): each with `chapter_num`, `chapter_title`, `articles` (array of 条)
- Each article (条): `article_num`, `article_title`, `paragraphs` (array of 項), `items` (号)
- `revision_info` (改正履歴 summary if available)
- `source_authority`: `"official"` (公式 cite_required)
- `attribution_text`: `"出典: e-Gov 法令検索 (デジタル庁) https://laws.e-gov.go.jp/law/<law_id>"`
- `source_url`: direct e-Gov URL for the law's HTML view
## Workflow pattern (paired with search_egov_laws)
```
1. User asks about a Japanese law topic
↓
2. search_egov_laws (keyword search) → list of candidate laws (10-20)
↓
3. User / agent selects one law_id
↓
4. get_egov_laws_data (law_id) → full statute text + structure
↓
5. Quote specific 条 / 項 / 号 with attribution_text + source_url
```
## Differentiation from third-party sources
e-Gov 法令検索 is the **only official statutory database** for Japanese law. Third-party services (jurist sites, commentary platforms) provide interpretation but not 公式 text. When AI agent needs cite_required statute text:
- ✅ e-Gov (this skill) — 公式, current, デジタル庁 maintained
- ❌ Wikipedia / commentary sites — not cite_required for legal opinion
## Boundary
- Output is verbatim statute text + structure — **NOT legal advice or interpretation**.
- 法令解釈 / 適用判定 / 法的助言 requires a certified Japanese 弁護士 / 行政書士 / 司法書士.
- Cite `attribution_text` + `law_id` + 条番号 + `enforcement_date` for any quoted provisions.
- Laws are amended frequently — always show the `enforcement_date` of the version retrieved.
- Never fictionalize statute text. If a law has been repealed (廃止), the API may not return it; in that case, note this and recommend lawyer consultation for historical research.