diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 48bb16a..b59f9fc 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -7,16 +7,28 @@ on:
 
 jobs:
   gate:
-    runs-on: ubuntu-latest
+    # Local-first + cross-platform is a core promise, and Windows is the
+    # divergent platform (backslash paths, native better-sqlite3 build, no
+    # symlinks). Run the full gate on both Linux and Windows so path/native
+    # regressions are caught in CI, not on a user's machine.
+    strategy:
+      fail-fast: false
+      matrix:
+        # windows-2022 (not windows-latest): the newer windows-latest image ships
+        # a Visual Studio that node-gyp 10 (bundled with Node 20) can't detect
+        # ("unknown version undefined at ...\Visual Studio\18"), so better-sqlite3's
+        # source-build fallback fails. VS 2022 on windows-2022 is detected fine.
+        os: [ubuntu-latest, windows-2022]
+    runs-on: ${{ matrix.os }}
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-node@v4
         with:
           node-version: 20
           cache: npm
-      # npm install (not ci) so the Linux runner resolves its own platform-
-      # specific optional deps (e.g. @tailwindcss/oxide native/wasm variants),
-      # which a Windows-generated lockfile can't fully pin cross-platform.
+      # npm install (not ci) so each runner resolves its own platform-specific
+      # optional deps (e.g. @tailwindcss/oxide native/wasm variants), which a
+      # single committed lockfile can't pin across Linux + Windows.
       - run: npm install --no-audit --no-fund
       - run: npm run lint
       - run: npm run typecheck
@@ -25,4 +37,6 @@ jobs:
       # Supply-chain gate: fail the build on CRITICAL advisories in prod deps.
       # High/moderate transitive advisories that can't be fixed until an
       # upstream release (e.g. next's bundled postcss) are tracked in SECURITY.md.
+      # Advisory data is platform-independent, so run it once (Linux).
       - run: npm audit --omit=dev --audit-level=critical
+        if: matrix.os == 'ubuntu-latest'
diff --git a/docs/MEGASPRINT_ROADMAP.md b/docs/MEGASPRINT_ROADMAP.md
index 45d04c7..3f56d89 100644
--- a/docs/MEGASPRINT_ROADMAP.md
+++ b/docs/MEGASPRINT_ROADMAP.md
@@ -20,7 +20,7 @@ free + BYOK path (`ANTHROPIC_API_KEY` optional; offline analyzer is the default)
 | **MS2** ✅ | Security | The app ingests users' private source code — it must be provably safe | **DONE** — symlink-confined walk, shell-free git, realpath+credential-dir blocks, prod CSP no `unsafe-eval`, `npm audit` critical-gate; self-scan security lens **A/100** |
 | **MS3** ✅ | Architecture & Data Integrity | Kill load-bearing hacks (description-as-path, no backup) | **DONE** — `sourcePath` column (migration 0008) retires the "Imported from:" hack; DB durability (`synchronous=NORMAL`, `busy_timeout`, boot integrity check, WAL checkpoint on exit) + downloadable backup; workspace error/loading boundaries; graph reconciliation already handled (`graphJson ?? buildKnowledgeGraph`). **#27/#28 module FKs deferred to MS4** (no analyzer-produced associations to populate them — REVIEW-012). Gate green (19/19) |
 | **MS4** ✅ | Algorithms & Analysis Depth | The analysis *is* the product — make it rigorous and calibrated | **DONE** — ✅ codeHealth calibrated to code *density* not project size (self-scan F→**B(76)**, overall **A(95)**); `computeCodeHealthScore` + relativity test. ✅ incremental re-analysis — `computeSourceFingerprint` (path\|size\|mtime sha256) + `learning_maps.source_fingerprint` (0009); reanalyze short-circuits `{unchanged:true}`. ✅ [benchmarks](BENCHMARKS.md) — 300-file analysis <200ms, flat ~4ms lens pass, incremental ~8× cheaper. ✅ #27/#28 module associations — soft `module_key` on questions/concepts (0010), deterministic tagging on both AI + offline paths, module cards link to their questions/concepts |
-| **MS5** | Quality & Reliability (QA) | 16 tests is a foundation, not production confidence | Critical path e2e-covered; CI green on Win/mac/Linux from a clean clone |
+| **MS5** 🔄 | Quality & Reliability (QA) | 16 tests is a foundation, not production confidence | **In progress** — ✅ E2E smoke of the critical path (import→analyze→deep-review→export), which also closed a real gap: the deep review now reaches the export. ✅ broadened coverage — BM25 search (indexing/ranking/rebuild). ✅ cross-platform CI — gate now runs on Windows + Linux (macOS deferred: fs/path-equivalent to Linux, 10× runner cost; Windows is the divergent platform). Suite 24/24. Remaining: watch first Windows CI run land green |
 | **MS6** | Framework & Performance | Production Next.js build quality | Standalone build runs; perf/a11y budgets met; no UI-blocking analysis |
 | **MS7** | Distribution & Release | Turn the repo into installable, versioned software — the actual "deploy" | `npx nirmiqcodesensei@latest` runs on a fresh machine; tagged v1.0.0 GitHub Release; CHANGELOG current; scaling-N/A ADR recorded |
 | MS8 | Launch (stretch) | Discoverability & polish | Onboarding, docs, MCP-directory submission, desktop-installer spike |
diff --git a/lib/services/export.service.ts b/lib/services/export.service.ts
index 5ded34c..3554690 100644
--- a/lib/services/export.service.ts
+++ b/lib/services/export.service.ts
@@ -6,6 +6,7 @@ import { getConceptLinksByWorkspaceId } from "@/lib/services/concept-link.servic
 import { getDailyLogsByWorkspaceId } from "@/lib/services/daily-log.service";
 import { formatDate, parseExpectedPoints } from "@/lib/utils";
 import type { ServiceResult } from "@/lib/types";
+import type { SeniorReview } from "@/lib/services/senior-review.service";
 
 export type ExportPayload = {
   filename: string;
@@ -42,6 +43,78 @@ function hr(): string {
   return "\n\n---\n\n";
 }
 
+const SEV_EMOJI: Record<string, string> = {
+  critical: "🔴",
+  high: "🟠",
+  medium: "🟡",
+  low: "🔵",
+  info: "⚪",
+};
+
+// Minimal structural shape shared by every senior-review lens, so we can render
+// them uniformly without importing eight distinct lens types.
+type LensLike = {
+  score: { grade: string; score: number };
+  findings: Array<{ severity: string; title: string }>;
+  present?: boolean;
+};
+
+// Render the stored Senior Review (deep review) as a compact export section:
+// overall grade, a per-lens grade table, and the worst findings. Returns [] when
+// there is no review (manual maps) or the JSON can't be parsed.
+function renderSeniorReview(json: string | null): string[] {
+  if (!json) return [];
+  let review: SeniorReview;
+  try {
+    review = JSON.parse(json) as SeniorReview;
+  } catch {
+    return [];
+  }
+
+  const lines: string[] = [hr(), "## 🔍 Senior Review", ""];
+  lines.push(`**Overall: ${review.overall.grade} (${review.overall.score}/100)**`);
+  lines.push("");
+  if (review.overall.summary) {
+    lines.push(`> ${review.overall.summary}`);
+    lines.push("");
+  }
+
+  const lenses: Array<[string, LensLike]> = [
+    ["Security", review.security],
+    ["Testing", review.testing],
+    ["Code Health", review.codeHealth],
+    ["Architecture", review.architecture],
+    ["Frontend", review.frontend],
+    ["Backend", review.backend],
+    ["Dependencies", review.dependencies],
+    ["Feasibility", review.feasibility],
+  ];
+
+  lines.push("| Lens | Grade |");
+  lines.push("|------|-------|");
+  for (const [label, lens] of lenses) {
+    if (lens.present === false) continue; // frontend/backend absent for this project
+    lines.push(`| ${label} | ${lens.score.grade} (${lens.score.score}) |`);
+  }
+  lines.push("");
+
+  // Worst findings across every lens, most severe first.
+  const order = ["critical", "high", "medium", "low", "info"];
+  const findings = lenses
+    .flatMap(([, lens]) => (lens.present === false ? [] : lens.findings))
+    .sort((a, b) => order.indexOf(a.severity) - order.indexOf(b.severity))
+    .slice(0, 5);
+  if (findings.length > 0) {
+    lines.push("**Top findings:**");
+    for (const f of findings) {
+      lines.push(`- ${SEV_EMOJI[f.severity] ?? "•"} ${f.title}`);
+    }
+    lines.push("");
+  }
+
+  return lines;
+}
+
 export async function generateWorkspaceMarkdown(
   workspaceId: string
 ): Promise<ServiceResult<ExportPayload>> {
@@ -141,6 +214,11 @@ export async function generateWorkspaceMarkdown(
     }
   }
 
+  // ── Senior Review (deep review) ───────────────────────────────────────────
+  if (map?.seniorReviewJson) {
+    lines.push(...renderSeniorReview(map.seniorReviewJson));
+  }
+
   // ── Explain-Back ──────────────────────────────────────────────────────────
   lines.push(hr());
   lines.push("## 💬 Explain-Back");
diff --git a/tests/import-pipeline.test.mts b/tests/import-pipeline.test.mts
index 599e5a3..9731524 100644
--- a/tests/import-pipeline.test.mts
+++ b/tests/import-pipeline.test.mts
@@ -520,6 +520,79 @@ test("createConceptLinkSchema: conceptType enum-enforced on the form path", asyn
   );
 });
 
+// ── BM25 search ranking (MS5 coverage) ───────────────────────────────────────
+test("search: BM25 indexes, ranks by relevance, and rebuilds cleanly", async () => {
+  const { createWorkspace } = await import("@/lib/services/workspace.service");
+  const { buildSearchIndex, searchWorkspace, hasSearchIndex } = await import(
+    "@/lib/services/search.service"
+  );
+
+  const wsRes = await createWorkspace({ title: "search-fixture", type: "project" });
+  assert.ok(wsRes.ok);
+  if (!wsRes.ok) return;
+  const wsId = wsRes.data.id;
+
+  assert.equal(await hasSearchIndex(wsId), false, "no index before build");
+
+  const built = await buildSearchIndex(wsId, [
+    { filePath: "auth/login.ts", chunkType: "file", layer: "Services / Logic",
+      chunkText: "authenticate user login password session authentication token" },
+    { filePath: "db/schema.ts", chunkType: "file", layer: "Data Layer",
+      chunkText: "database table column migration schema drizzle sqlite" },
+    { filePath: "ui/Button.tsx", chunkType: "file", layer: "UI Components",
+      chunkText: "render button click handler component props state" },
+  ]);
+  assert.ok(built.ok && built.data === 3, "index built with 3 chunks");
+  assert.equal(await hasSearchIndex(wsId), true, "index present after build");
+
+  const res = await searchWorkspace(wsId, "authentication login");
+  assert.ok(res.ok);
+  if (!res.ok) return;
+  assert.ok(res.data.length > 0, "auth query matches");
+  assert.equal(res.data[0].filePath, "auth/login.ts", "most relevant chunk ranks first");
+  assert.ok(
+    res.data.every((r, i) => i === 0 || r.score <= res.data[i - 1].score),
+    "results sorted by descending BM25 score"
+  );
+
+  // Stopwords + sub-3-char tokens are dropped entirely.
+  const empty = await searchWorkspace(wsId, "the a is to");
+  assert.ok(empty.ok && empty.data.length === 0, "stopword-only query returns nothing");
+
+  // Rebuild replaces the index rather than accumulating.
+  const rebuilt = await buildSearchIndex(wsId, [
+    { filePath: "solo.ts", chunkType: "file", layer: null, chunkText: "unrelated solo content" },
+  ]);
+  assert.ok(rebuilt.ok && rebuilt.data === 1);
+  const after = await searchWorkspace(wsId, "authentication");
+  assert.ok(after.ok && after.data.length === 0, "old chunks gone after rebuild");
+});
+
+// ── E2E smoke: import → analyze → deep-review → export (MS5) ──────────────────
+test("export: markdown captures the full pipeline (map, review, questions, concepts)", async () => {
+  const { generateWorkspaceMarkdown } = await import("@/lib/services/export.service");
+
+  const res = await generateWorkspaceMarkdown(workspaceId);
+  assert.ok(res.ok, res.ok ? "" : res.error);
+  if (!res.ok) return;
+
+  const { filename, markdown } = res.data;
+  assert.match(filename, /^nirmiqcodesensei-.+-\d+\.md$/, "filename = slug + timestamp");
+
+  // The chain: each pipeline stage's output must reach the exported artifact.
+  assert.match(markdown, /^# .+/m, "title header");
+  assert.ok(markdown.includes("Exported from NirmiqCodeSensei"), "export header line");
+  assert.ok(markdown.includes("## 🗺️ Learning Map"), "learning-map section");
+  assert.ok(markdown.includes("## 🔍 Senior Review"), "deep-review section reaches export");
+  assert.match(markdown, /\*\*Overall: [A-F] \(\d+\/100\)\*\*/, "overall grade line");
+  assert.ok(markdown.includes("| Lens | Grade |"), "per-lens grade table");
+  assert.ok(markdown.includes("## 💬 Explain-Back"), "explain-back section");
+  assert.match(markdown, /### Q1:/, "at least one generated question");
+  assert.ok(markdown.includes("## ⚙️ DSA Bridge"), "DSA-bridge section");
+  // The green answer submitted by the progress test flows through to the export.
+  assert.ok(markdown.includes("test answer"), "user answer captured in export");
+});
+
 // ── deleteWorkspace cascade ───────────────────────────────────────────────────
 test("deleteWorkspace: cascade leaves zero orphaned child rows", async () => {
   const res = await deleteWorkspace(workspaceId);