srj-wikidata-mcp
MCP server for maintaining specific Wikidata items (person, company, publications) with read tools and a two-step write workflow that enforces sourced, human-approved batch edits.
README
srj-wikidata-mcp
MCP server for maintaining the Wikidata items belonging to Stephen R. Jordan, SRJ Consulting & Services LLC, and the published volumes of The Operating Discipline for AI Library.
| Subject | Item |
|---|---|
| Stephen R. Jordan (person) | Q140622333 |
| SRJ Consulting & Services LLC | Q140622666 |
| Published volumes | separate items, linked from the person via P800 notable work |
Built to mirror srj-mcp: FastMCP, OAuth 2.1 + PKCE for Claude.ai's custom
connector flow, Docker on Render, public /healthz. Deployed as a separate
service on purpose — a different credential and a different blast radius, so a
bug here cannot reach the audit database.
Tools
Reads — unrestricted.
| Tool | Purpose |
|---|---|
health_check |
API reachable; reports whether writes are enabled and as which user |
search_entity |
Find Q-ids or P-ids by name |
get_entity |
Labels, descriptions, statements with claim GUIDs, and a count of unsourced statements |
sparql |
Read-only query against query.wikidata.org |
Writes — two-step, no single-claim write tool exists.
| Tool | Purpose |
|---|---|
plan_edits |
Validate and render a batch. Touches nothing. Returns a plan_id |
apply_plan |
Execute a plan. Requires confirm="I-CONFIRM-WIKIDATA-EDIT" |
list_plans |
Pending and applied plans in this process |
Why writes are shaped this way
Wikidata's bot policy requires community approval before automated editing, and an LLM writing statements via the API at volume is a bot whatever it is called. This server is built for the other thing: a human approving a small, sourced, reviewable batch that is attributable in the revision history.
Enforced, not advisory:
- Every
add_claimmust carry at least one reference. Refused otherwise. Both items currently have zero references on every statement, and with no sitelinks their notability rests on being supported by serious publicly available sources — so adding unsourced claims makes the deletion case stronger, not weaker. - 30 operations per plan. Split larger work.
- 2 seconds between writes,
maxlag=5,bot=0, and every edit summarised asvia srj-wikidata-mcp (assistant-prepared, human-approved batch). - Plans live in memory. A restart discards pending plans, which is correct: an unapproved plan should not survive a deploy and get applied later against a changed item.
Setup
1. Register a Wikimedia OAuth consumer
At Special:OAuthConsumerRegistration:
- Choose OAuth 2.0
- Tick "This consumer is for use only by <your username>" (owner-only — no third-party authorisation dance, and the token acts solely as you)
- Grants: Edit existing pages, and Create, edit, and move pages
- Copy the access token it issues
Leave WIKIDATA_OAUTH_TOKEN blank to run read-only. Every read tool still works.
2. Deploy
git init && git add -A && git commit -m "srj-wikidata-mcp initial"
git remote add origin <your remote>
git push -u origin main
Render → New → Blueprint → point at this repo. render.yaml provisions it.
Then set the two secrets in the dashboard (both are sync: false):
WIKIDATA_OAUTH_TOKENWIKIDATA_MCP_TOKEN—python3 -c "import secrets; print(secrets.token_urlsafe(48))"
3. Verify
curl https://srj-wikidata-mcp.onrender.com/healthz
Then add https://srj-wikidata-mcp.onrender.com/mcp as a custom connector in
Claude settings and call health_check. If it returns anonymous: true, the
OAuth token is invalid, expired, or missing the edit grant.
Local
cp .env.example .env # fill in
pip install -r requirements.txt
uvicorn src.server:app --host 0.0.0.0 --port 8000
Known defects in the live items, as at 2026-07-29
Recorded in srj_audit.career_narratives under slug wikidata-items.
- Every statement on both items has zero references. Highest priority.
- Person item: LinkedIn personal profile ID is
stepherjordan— a typo forstephenrjordan. Currently a dead external link. - Person item: The AI Efficiency & Process Optimization™ appears twice under notable work.
- Person item: no employer statements — Citi, Intel, McAfee, Optiv all absent.
- Company item: does not record the Nevada → Texas conversion of 22 May 2026.
- Neither item has sitelinks on any Wikimedia project.
Sourcing
Statements should be sourced from material that already exists and is checkable:
srj_audit.press_booksandpress_book_isbns— titles, page counts, ISBNs, publication dates- Open Library editions —
OL62330768M,OL62330769M,OL62330772M,OL62371435M; authorOL16525496A - OpenCorporates
us_tx/0806615653— Texas registration - Library of Congress and Dewey classifications on Volumes I and III
Use P854 (reference URL) or P248 (stated in). Note that
press_books.published_on and Open Library disagree on Volumes I and III —
resolve that before citing either as a publication date.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。