feat(import): harden workbook parser boundaries

packages/engine/src/__tests__/estimate-export-serializer.test.ts

@@ -1,4 +1,3 @@
-import * as XLSX from "xlsx";
 import {
   EstimateExportFormat,
   EstimateStatus,
@@ -144,8 +143,8 @@ function buildSource(): EstimateExportSource {
 }
 
 describe("estimate export serializer", () => {
-  it("creates a structured JSON export payload", () => {
-    const payload = serializeEstimateExport(buildSource(), EstimateExportFormat.JSON);
+  it("creates a structured JSON export payload", async () => {
+    const payload = await serializeEstimateExport(buildSource(), EstimateExportFormat.JSON);
 
     expect(payload.encoding).toBe("utf8");
     expect(payload.mimeType).toBe("application/json; charset=utf-8");
@@ -154,9 +153,16 @@ describe("estimate export serializer", () => {
     expect(payload.previewText).toContain('"schemaVersion": 1');
   });
 
-  it("creates a multi-sheet xlsx export payload", () => {
-    const payload = serializeEstimateExport(buildSource(), EstimateExportFormat.XLSX);
-    const workbook = XLSX.read(payload.content, { type: "base64" });
+  it("creates a multi-sheet xlsx export payload", async () => {
+    const payload = await serializeEstimateExport(buildSource(), EstimateExportFormat.XLSX);
+    const ExcelJS = await import("exceljs");
+    const workbook = new ExcelJS.Workbook();
+    const workbookBytes = Uint8Array.from(Buffer.from(payload.content, "base64"));
+    const workbookBuffer = workbookBytes.buffer.slice(
+      workbookBytes.byteOffset,
+      workbookBytes.byteOffset + workbookBytes.byteLength,
+    );
+    await workbook.xlsx.load(workbookBuffer);
 
     expect(payload.encoding).toBe("base64");
     expect(payload.sheetNames).toEqual([
@@ -167,7 +173,7 @@ describe("estimate export serializer", () => {
       "Resources",
       "Metrics",
     ]);
-    expect(workbook.SheetNames).toContain("DemandLines");
+    expect(workbook.getWorksheet("DemandLines")).toBeDefined();
     expect(payload.byteLength).toBeGreaterThan(100);
   });
 });

packages/engine/src/estimate/export-serializer.ts

@@ -1,4 +1,3 @@
-import * as XLSX from "xlsx";
 import {
   EstimateExportFormat,
   type EstimateExportArtifactPayload,
@@ -8,6 +7,8 @@ import {
 } from "@capakraken/shared";
 import { summarizeEstimateDemandLines } from "./metrics.js";
 
+type ExcelJsModule = typeof import("exceljs");
+
 type ExportProjectRef = {
   id: string;
   name: string;
@@ -109,6 +110,18 @@ type ExportMetric = {
   updatedAt: Date;
 };
 
+type ExportSheetRow = Record<string, unknown>;
+
+let _excelJs: ExcelJsModule | null = null;
+
+async function getExcelJS() {
+  if (!_excelJs) {
+    _excelJs = await import("exceljs");
+  }
+
+  return _excelJs;
+}
+
 export interface EstimateExportSource {
   estimate: {
     id: string;
@@ -508,6 +521,52 @@ function base64ByteLength(content: string) {
   return Math.floor((content.length * 3) / 4) - padding;
 }
 
+function buildSheetColumns(rows: ExportSheetRow[]) {
+  return Array.from(
+    rows.reduce((keys, row) => {
+      for (const key of Object.keys(row)) {
+        keys.add(key);
+      }
+      return keys;
+    }, new Set<string>()),
+  );
+}
+
+function toWorksheetCellValue(value: unknown): boolean | Date | number | string {
+  if (value == null) {
+    return "";
+  }
+
+  if (value instanceof Date) {
+    return value;
+  }
+
+  if (typeof value === "string" || typeof value === "number" || typeof value === "boolean") {
+    return value;
+  }
+
+  return stringifyValue(value);
+}
+
+function appendWorksheetFromRows(
+  workbook: InstanceType<ExcelJsModule["Workbook"]>,
+  sheetName: string,
+  rows: ExportSheetRow[],
+): void {
+  const worksheet = workbook.addWorksheet(sheetName);
+  const columns = buildSheetColumns(rows);
+
+  if (columns.length === 0) {
+    return;
+  }
+
+  worksheet.addRow(columns);
+
+  for (const row of rows) {
+    worksheet.addRow(columns.map((column) => toWorksheetCellValue(row[column])));
+  }
+}
+
 function buildTextPayload(
   format: EstimateExportFormat,
   content: string,
@@ -536,17 +595,18 @@ function buildTextPayload(
   };
 }
 
-function buildXlsxPayload(
+async function buildXlsxPayload(
   source: EstimateExportSource,
   summary: EstimateExportSummary,
-): EstimateExportArtifactPayload {
+): Promise<EstimateExportArtifactPayload> {
   const overviewRows = buildOverviewRows(source, summary);
   const assumptionRows = buildAssumptionRows(source.version.assumptions);
   const scopeRows = buildScopeRows(source.version.scopeItems);
   const demandRows = buildDemandRows(source);
   const resourceRows = buildResourceRows(source.version.resourceSnapshots);
   const metricRows = buildMetricRows(source.version.metrics);
-  const workbook = XLSX.utils.book_new();
+  const ExcelJS = await getExcelJS();
+  const workbook = new ExcelJS.Workbook();
   const sheets = [
     { name: "Overview", rows: overviewRows },
     { name: "Assumptions", rows: assumptionRows },
@@ -557,17 +617,11 @@ function buildXlsxPayload(
   ] as const;
 
   for (const sheet of sheets) {
-    XLSX.utils.book_append_sheet(
-      workbook,
-      XLSX.utils.json_to_sheet(sheet.rows),
-      sheet.name,
-    );
+    appendWorksheetFromRows(workbook, sheet.name, sheet.rows);
   }
 
-  const content = XLSX.write(workbook, {
-    type: "base64",
-    bookType: "xlsx",
-  });
+  const buffer = await workbook.xlsx.writeBuffer();
+  const content = Buffer.from(buffer).toString("base64");
 
   return {
     schemaVersion: 1,
@@ -593,10 +647,10 @@ function buildXlsxPayload(
   };
 }
 
-export function serializeEstimateExport(
+export async function serializeEstimateExport(
   source: EstimateExportSource,
   format: EstimateExportFormat,
-): EstimateExportArtifactPayload {
+): Promise<EstimateExportArtifactPayload> {
   const summary = buildSummary(source);
 
   if (format === EstimateExportFormat.JSON) {