github/github-mcp-server · error

updated_field must set either id or name, not both

Error message

updated_field must set either id or name, not both

What it means

Raised by parseBatchFieldSpec (pkg/github/projects_batch.go:518) when the updated_field object sets both 'id' and 'name'. The library refuses ambiguity about which field identifier wins and fails the whole request before any item is processed — this is a top-level argument error, not a per-item one.

Source

Thrown at pkg/github/projects_batch.go:518

func parseBatchFieldSpec(raw any) (batchFieldSpec, error) {
	var spec batchFieldSpec
	input, ok := raw.(map[string]any)
	if !ok || input == nil {
		return spec, fmt.Errorf("updated_field must be an object")
	}

	value, hasValue := input["value"]
	if !hasValue {
		return spec, fmt.Errorf("updated_field.value is required")
	}
	spec.value = value

	idField, hasID := input["id"]
	nameField, hasName := input["name"]
	switch {
	case hasID && hasName:
		return spec, fmt.Errorf("updated_field must set either id or name, not both")
	case !hasID && !hasName:
		return spec, fmt.Errorf("updated_field requires either id or name")
	case hasID:
		id, err := validatePositiveInt64(idField)
		if err != nil {
			return spec, fmt.Errorf("updated_field.id: %w", err)
		}
		spec.id = id
	default:
		name, ok := nameField.(string)
		if !ok || name == "" {
			return spec, fmt.Errorf("updated_field.name must be a non-empty string")
		}
		spec.name = name
	}
	return spec, nil
}

View on GitHub (pinned to 0ea1f775a7)

Solutions

1. Delete one of the two keys — keep name for readability or id to skip the field-by-name resolution
2. If your source has both, pick id when it came from a recent list_project_fields call
3. Strip null-valued keys before sending; null counts as present for this check
4. Remember the choice applies to the entire batch, not one item

Example fix

// before
{"updated_field": {"id": 123456, "name": "Status", "value": "Done"}}
// after
{"updated_field": {"name": "Status", "value": "Done"}}

_, hasID := spec["id"]
_, hasName := spec["name"]
if hasID && hasName {
	return fmt.Errorf("updated_field: set only one of id or name")
}

func hasExactlyOneFieldID(spec map[string]any) bool {
	_, id := spec["id"]
	_, name := spec["name"]
	return id != name // exactly one true
}

When it happens

Trigger: {"updated_field": {"id": 123456, "name": "Status", "value": "Done"}} — both identifier keys present. Both keys are looked up with Go's two-value map access, so "id": null plus "name": "Status" still counts as both set.

Common situations: Merging a cached field record (has id) with a human-readable override (has name); serializers that emit null for unset keys, which still counts as present here; defensive clients that include every known property.

Related errors

• each item must set exactly one of node_id, item_id, or item_
• updated_field must be an object
• updated_field.value is required
• updated_field requires either id or name
• updated_field.id: %w

AI-assisted analysis of github/github-mcp-server@0ea1f775a7 (2026-08-15). Data as JSON: /api/errors/52a50ca5fefe36eb. Report an issue: GitHub.