get_template_fields Leave feedback

get_template_fields reads a template and tells the agent what data it needs: every <<…>> tag, the data names (contract, customer, items), the fields it reads, and a JSON skeleton to fill in and pass to assemble_document as dataJson. Nothing is generated. Example prompt: “What data does contract-template.docx need?”

Tool description (as the AI agent sees it):

Lists what a template needs before generating a document from it: every «…» tag, the names it uses for its data, the fields it reads, and a JSON skeleton of the data to fill in and pass to assemble_document as ‘dataJson’. Use it when the user asks what data a template expects, or before calling assemble_document with data you have to build. Reads DOCX, DOTX, DOCM, RTF, ODT, TXT, Markdown, HTML, XLSX, XLTX, ODS, PPTX, POTX, ODP and EML templates; binary DOC, XLS, PPT and MSG templates cannot be read without generating them (readable=false). Processes no data and does not run the engine. Do NOT pre-check whether files exist — just pass the filename the user provided; a file that cannot be found returns the list of available files instead. Returns JSON: { fileName, family, readable, tagCount, dataSourceNames[], fields[], dataSkeleton, tags: [{ tag, kind, part }], tagsTruncated, notes[] }. fields use ‘[]’ for lists (‘customers[].Name’). The skeleton is inferred from the tags; check it against your data. On failure isError is set and the text starts with ‘Template inspection failed for’.

{
  "name": "get_template_fields",
  "arguments": {
    "template": {
      "filePath": "contract-template.docx"
    }
  }
}

JSON — captured from server 26.11.0 (the tags array is trimmed to its first entries):

{
  "fileName": "contract-template.docx",
  "family": "WordProcessing",
  "readable": true,
  "tagCount": 17,
  "dataSourceNames": [ "contract", "customer", "items" ],
  "fields": [ "contract.Number", "contract.Date", "contract.Discount",
              "customer.Name", "customer.Address", "customer.Contact",
              "customer.Email",
              "items[].Title", "items[].Quantity", "items[].UnitPrice" ],
  "dataSkeleton": {
    "contract": { "Number": "", "Date": "", "Discount": "" },
    "customer": { "Name": "", "Address": "", "Contact": "", "Email": "" },
    "items": [ { "Title": "", "Quantity": 0, "UnitPrice": 0 } ]
  },
  "tags": [
    { "tag": "<<[contract.Number]>>", "kind": "expression", "part": "body" },
    { "tag": "<<foreach [item in items]>>", "kind": "foreach", "part": "body" },
    ...
  ],
  "tagsTruncated": false,
  "notes": [ "fields and dataSkeleton are inferred from the tags, not from
              the engine; values computed by Select/GroupBy projections are
              not followed." ]
}

On a document with {{…}} placeholders, tagCount is 0 and a note says “The template contains 2 ‘{{…}}’ placeholder(s). GroupDocs.Assembly does not process that syntax and leaves it as text; write «[name]» instead.” It inspects tags statically: an unclosed <<foreach>> is still listed — the engine reports it when you generate.

• “What data does contract-template.docx need?”
• “Give me the JSON I have to fill for this template.”
• “Why does this template not fill in? Check its placeholders.”

See it used end-to-end: Inspect and fix a template.