TaskQueryResponse

What the query answers with.

Type object

Shape

Field	Type	Description
rowsrequired	array of TaskRow	The matching tasks, one row each, in the requested order.
definitionsrequired	array of AttributeDefinition	Every attribute definition in play: the system registry always, plus the scoped workspaces' custom definitions - so a surface draws a value from its type rather than guessing.
tagsrequired	array of AttributeTag	The scoped workspaces' tag tables (live tags only, merge tombstones resolved away), so a `koru.tag` value resolves to a name and a chip color without another request.
groups	array of TaskGroup	Corpus-wide counts per group - per list, or per bucket - when the query asked for them with `group_by`. Empty otherwise. Names are not repeated here. A group carries its key - a `ListId`, or a bucket token - with the count and the overdue share; the name, the swatch and the rail order live in the workspace's list table, which a client reads from its own workspace rather than from this response. Two spellings of one name in one answer is a drift waiting to happen - the same rule `tags` follows above. Default `[]`
projectedrequired	boolean	Whether every workspace in scope has reached the read model at all. `false` is "still indexing" - the projector is debounced, so a task created a moment ago may not be a row yet.
updating_workspaces	array of WorkspaceId	The workspaces the answer may be behind on, by the same rule search and the attribute query use: named rather than waited for. Default `[]`
limitrequired	integer (uint, >= 0)	The page size actually applied, after the store's cap.
next_after	string or null	The `task:` reference to pass as the next request's [`TaskQuery::after`], or absent when this page is the end of the list. Default `null`
restarted	boolean	The `after` this request carried named a row the filter no longer matches - completed, removed or binned between pages - so the answer is page one again. A client can tell a fresh list from a continuation instead of silently drawing page one twice. Default `false`
covered	boolean or null	Whether the rows and counts include every write in [`TaskQuery::covering`]. `None` means this server cannot tell - an older store that never read the requirement. A client must read `None` as unknown, never as coverage: it records the capability as absent, falls back to its ordinary schedule and stops sending requirements, so an old server is never asked again quickly for an answer it cannot give. A server that understands the field always answers `Some`, `true` when nothing was required. A different question from [`projected`](Self::projected), which says whether the workspace has been indexed at all; both are worth having. Default `null`

Example

{
  "rows": [
    {
      "workspace": "00000000-0000-4000-8000-000000000000",
      "document": "00000000-0000-4000-8000-000000000000",
      "node": "string",
      "task": "string",
      "node_kind": "string",
      "lifecycle": "string",
      "document_trashed": false,
      "in_trash": false,
      "breadcrumb": "string",
      "title": "string",
      "attributes": [
        {
          "id": "string",
          "role": "string",
          "kind": "string",
          "value": "string",
          "zone": "string",
          "label": "string"
        }
      ]
    }
  ],
  "definitions": [
    {
      "id": "string",
      "key": "string",
      "label": "string",
      "value_type": "string",
      "cardinality": "string",
      "scope": "string",
      "system": false,
      "options": [],
      "filterable": false,
      "sortable": false,
      "groupable": false
    }
  ],
  "tags": [
    {
      "id": "string",
      "name": "string",
      "color": "string",
      "workspace": {}
    }
  ],
  "groups": [],
  "projected": false,
  "updating_workspaces": [],
  "limit": 0,
  "next_after": "string",
  "restarted": false,
  "covered": false
}

The contract has no example values, so example bodies are generated from each schema: every field is shown with a placeholder of its type.

Used by

POST/v1/tasks/query
What is outstanding, across workspaces, from the read model

Generated from contract version 1. Last updated November 12, 2026.