Licensed to be used in conjunction with basebox, only.
// integration
SharePoint
Applies to
Product: Cloud · Server · Audience: Administrator
The SharePoint connector lets the basebox assistant search files and documents in SharePoint, browse sites and libraries and read documents. It changes nothing and signs in with a Windows (Active Directory) account. Search covers the whole farm, trimmed to what the signed-in account may see.
What does the connector do?
This and the following sections are generated from the information the connector's image carries. What applies to every connector is described under What applies to every connector.
In the catalog it describes itself as:
Search files and documents across a SharePoint site collection.
It gives the assistant these instructions:
Read-only SharePoint document store, showing only what the user's own credential may see: absent results can mean 'not permitted' as well as 'not there'. Find documents with sp_search, or browse with sp_list_webs, sp_list_libraries and sp_list_items, each reporting one level at a time. Read one with sp_get_document. Its text for Office documents and PDFs can come from SharePoint's search index and then be an extract, or be truncated; the result says which, so never imply you read a whole document when you did not. Always cite the source document's URL.
Prepare SharePoint
- The connector works with on-premises SharePoint Server farms. SharePoint Online (Microsoft 365) is not supported: it accepts only Entra ID sign-in, which the connector does not offer.
- The connector signs in with a Windows account over Basic authentication and HTTPS. The SharePoint web application must accept Basic authentication; NTLM, Kerberos, forms-based sign-in and ADFS do not work.
Settings
The administrator fills in these fields in the connector's settings. They apply to the whole organisation.
| Field | Required | Meaning |
|---|---|---|
| Site collection URL | no | The site collection the connector starts from, over HTTPS. Search covers the whole farm, trimmed to what each user may see. Example: https://sp.acme.local/sites/docs |
| Help link for signing in | no | Where users can read how to sign in, e.g. a page on your intranet. |
Sign-in
The administrator decides how the connector signs in: with Personal sign-in each person signs in with their own account and sees only what they may. With Company account everyone uses one account the administrator enters.
Personal sign-in
- Windows account – asks for a username and password. Username: Your Windows/AD account, e.g. ACME\jane.doe or jane.doe@acme.local
Company account
- Windows service account – asks for a username and password. Username: A dedicated AD account, e.g. ACME\svc-basebox
A company account gives everyone the same rights
Every basebox user reaches the system with the rights of this one account. Results are not trimmed to what each person may see.
The connector adds:
SharePoint trims search results to what the signed-in account may see. A shared account flattens that: every basebox user will see whatever that one account can see.
Tools
The assistant can call 5 tools.
| Tool | Effect | Description |
|---|---|---|
sp_get_document |
reads | Read the text of one document, identified by the ServerRelativeUrl from sp_list_items or the Path from sp_search. Office documents and PDFs are read from SharePoint's search index where it holds their text, which is as fresh as the last crawl and may be an extract rather than the whole document; otherwise, and for plain-text files, the file itself is read, scanned documents through OCR. The result names the document's URL to cite, which path produced the text, and whether it was truncated. |
sp_list_items |
reads | Browse the folders and files directly inside a document library folder. Start from a library's RootFolder.ServerRelativeUrl (from sp_list_libraries), then call again with a returned subfolder's ServerRelativeUrl to go deeper. Each file's ServerRelativeUrl identifies it for retrieval and citation. |
sp_list_libraries |
reads | List document libraries in a site. Returns that site's own lists only: omit the argument for the configured site collection, or pass a site URL from sp_list_webs to list that site's libraries instead. |
sp_list_webs |
reads | List the sites/subsites the user can reach. Returns a site's direct children only: start with no argument for the configured site collection, then call again passing a returned site's URL to explore further down the tree. |
sp_search |
reads | Search across sites and libraries, returning title, path and snippets. The query accepts SharePoint KQL, so filters such as 'filetype:pdf budget' narrow it. |
Generated from mcp-sharepoint:v0.3.1.