f3e2bacd499ddd0a46a8999f16d03de4bec529e0
diff --git a/src/c21_app.ts b/src/c21_app.ts
index 6c80556aa8e28d7942d91bcce4fd8ba0362a7b67..8f94a9a0e8e4ac01b2a5005f8cc87be4e185b6fd 100644
--- a/src/c21_app.ts
+++ b/src/c21_app.ts
@@ -78,15 +78,15 @@ const demoContext = () => ({
   scopeLabel: `${DEMO_REPOS} repos · ${DEMO_ORG}`,
   bannerHtml: DEMO_BANNER_HTML,
   narrative: {
-    changedHeading: "wat veranderde dit kwartaal",
+    changedHeading: "what changed this quarter",
     changedBody:
-      "Cursor's score zakte 15 punten nadat agent-mode in maart default werd; test-deletion-incidenten stegen van 2% naar 14% van refactor-commits, geconcentreerd in de `api-gateway` repo. Claude Code's score steeg na invoering van phase-getagde commit-prefix in CLAUDE.md aan het einde van januari. Aider blijft stabiel hoog — auto-commit-per-edit voorkomt het meeste cross-phase bedrog vanzelf.",
-    doingHeading: "wat we doen",
+      "Cursor's score dropped 15 points after agent-mode became default in March; test-deletion incidents climbed from 2% to 14% of refactor commits, concentrated in the `api-gateway` repo. Claude Code's score rose after a phase-tagged commit prefix was added to CLAUDE.md at the end of January. Aider stays steadily high — auto-commit-per-edit prevents most cross-phase cheating on its own.",
+    doingHeading: "what we're doing",
     doingBody:
-      "- **Cursor in `api-gateway`**: agent-mode gedeactiveerd voor refactor-prompts, CONVENTIONS-regel \"never delete a test in a refactor commit\" gepind ([details →](/reports/demo/agents/cursor)).\n- **Claude Code uitrollen**: het CLAUDE.md-template dat in `billing-service` werkte naar de andere drie repos kopiëren.\n- **Volgende meting**: 2026-04-30, mid-Q2, om te zien of de Cursor-fix vasthoudt.",
+      "- **Cursor in `api-gateway`**: agent-mode disabled for refactor prompts, CONVENTIONS rule \"never delete a test in a refactor commit\" pinned ([details →](/reports/demo/agents/cursor)).\n- **Roll out Claude Code**: copy the CLAUDE.md template that worked in `billing-service` to the other three repos.\n- **Next reading**: 2026-04-30, mid-Q2, to check whether the Cursor fix holds.",
   },
   footerLinks:
-    "[per-agent drill-down: Claude Code](/reports/demo/agents/claude-code) · [Cursor](/reports/demo/agents/cursor) · [Aider](/reports/demo/agents/aider) · [tests overzicht](/reports/demo/tests) · [back to /reports](/reports)",
+    "[per-agent drill-down: Claude Code](/reports/demo/agents/claude-code) · [Cursor](/reports/demo/agents/cursor) · [Aider](/reports/demo/agents/aider) · [tests overview](/reports/demo/tests) · [back to /reports](/reports)",
 });
 
 const liveContext = async () => {
@@ -102,7 +102,7 @@ const liveContext = async () => {
     period,
     scopeLabel: `${LIVE_REPO_OWNER}/${LIVE_REPO_NAME} · ${live.totalCommits} commits sampled${live.unknownCount > 0 ? ` (${live.unknownCount} unattributed, excluded)` : ""}`,
     bannerHtml: LIVE_BANNER_HTML,
-    footerLinks: `${drillLinks ? drillLinks + " · " : ""}[tests overzicht](/reports/live/tests) · [demo preview](/reports/demo) · [back to /reports](/reports)`,
+    footerLinks: `${drillLinks ? drillLinks + " · " : ""}[tests overview](/reports/live/tests) · [demo preview](/reports/demo) · [back to /reports](/reports)`,
   };
 };
 
@@ -476,7 +476,7 @@ ${rows}
   "/reports/demo": async () => {
     const ctx = demoContext();
     const html = await renderPage({
-      title: "TDD-discipline rapport · Q1 2026 (demo) — tdd.md",
+      title: "TDD-discipline report · Q1 2026 (demo) — tdd.md",
       description: "Mockup of the management-level TDD-discipline report — single page, three agents, with trend and narrative.",
       bodyMarkdown: execSummaryMd(ctx),
       ogPath: "https://tdd.md/reports/demo",
@@ -487,7 +487,7 @@ ${rows}
 
   "/reports/demo/tests": async () => {
     const html = await renderPage({
-      title: "Tests overzicht (demo) — tdd.md",
+      title: "Tests overview (demo) — tdd.md",
       description: "Mockup of the per-test overview: current pass/fail snapshot per repo plus test stability over the quarter.",
       bodyMarkdown: testsOverviewMd({
         period: DEMO_PERIOD,
@@ -523,8 +523,8 @@ ${rows}
   "/reports/live": async () => {
     const ctx = await liveContext();
     const html = await renderPage({
-      title: "TDD-discipline rapport · live — tdd.md",
-      description: `Live discipline rapport gebouwd uit de echte commit-historie van syntaxai/tdd.md (laatste ${LIVE_FETCH_COUNT} commits, 5-min cache).`,
+      title: "TDD-discipline report · live — tdd.md",
+      description: `Live discipline report built from the real commit history of syntaxai/tdd.md (last ${LIVE_FETCH_COUNT} commits, 5-min cache).`,
       bodyMarkdown: execSummaryMd(ctx),
       ogPath: "https://tdd.md/reports/live",
       noindex: true,
@@ -536,14 +536,14 @@ ${rows}
     const data = await buildLiveTestData(LIVE_REPO_OWNER, LIVE_REPO_NAME);
     const ranOn = data.ranAt ? new Date(data.ranAt).toISOString().slice(0, 10) : null;
     const period = data.runsCount === 0
-      ? "geen runs in bundle"
-      : `last run ${ranOn} · ${data.runsCount} run${data.runsCount === 1 ? "" : "s"} cumulatief`;
+      ? "no runs in bundle"
+      : `last run ${ranOn} · ${data.runsCount} run${data.runsCount === 1 ? "" : "s"} cumulative`;
     const unavailableNote = data.runsCount === 0
-      ? "Nog geen test-runs gebundeld. De volgende deploy draait `bun test --reporter=junit` op de huidige HEAD en publiceert het resultaat hier. Stabiliteit (flaky %, deletion) bouwt zich op naarmate er meer runs in de bundle staan — de demo op [/reports/demo/tests](/reports/demo/tests) toont waar het naartoe groeit."
+      ? "No test runs bundled yet. The next deploy will run `bun test --reporter=junit` on the current HEAD and publish the result here. Stability (flaky %, deletion) builds up as more runs land in the bundle — the demo at [/reports/demo/tests](/reports/demo/tests) shows where this is heading."
       : undefined;
     const html = await renderPage({
-      title: "Tests overzicht · live — tdd.md",
-      description: `Live test-snapshot van ${LIVE_REPO_OWNER}/${LIVE_REPO_NAME} — ${data.runsCount} run${data.runsCount === 1 ? "" : "s"} gebundeld.`,
+      title: "Tests overview · live — tdd.md",
+      description: `Live test snapshot of ${LIVE_REPO_OWNER}/${LIVE_REPO_NAME} — ${data.runsCount} run${data.runsCount === 1 ? "" : "s"} bundled.`,
       bodyMarkdown: testsOverviewMd({
         period,
         bannerHtml: LIVE_BANNER_HTML,
@@ -567,7 +567,7 @@ ${rows}
     const entry = ctx.reports.find((r) => r.slug === slug)!;
     const html = await renderPage({
       title: `${entry.name} drill-down · live — tdd.md`,
-      description: `Live drill-down voor ${entry.name} op syntaxai/tdd.md — trend, failure-mode breakdown, recent commits.`,
+      description: `Live drill-down for ${entry.name} on syntaxai/tdd.md — trend, failure-mode breakdown, recent commits.`,
       bodyMarkdown: md,
       ogPath: `https://tdd.md/reports/live/agents/${slug}`,
       noindex: true,
diff --git a/src/c51_render_reports.ts b/src/c51_render_reports.ts
index 126d8b61b14d419aaff7f3484ffa1bba765ffb1d..7f1be09ab39a6fc45fe3dc4e69964b53f76b0f16 100644
--- a/src/c51_render_reports.ts
+++ b/src/c51_render_reports.ts
@@ -121,7 +121,7 @@ const agentTagHtml = (slug: AgentReport["slug"]): string => {
 
 const stabilityRow = (s: TestStability): string => {
   const cls = s.flagged ? "test-stab-row flagged" : "test-stab-row";
-  const warn = s.flagged ? ` <span class="test-stab-warn" title="test-deletion of weakening dit kwartaal">⚠</span>` : "";
+  const warn = s.flagged ? ` <span class="test-stab-warn" title="test-deletion or weakening this quarter">⚠</span>` : "";
   return `<tr class="${cls}">
   <td class="test-stab-name">${escape(s.test)}<div class="test-stab-repo">${escape(s.repo)}</div></td>
   <td class="test-stab-num green">${s.pass}</td>
@@ -141,8 +141,8 @@ Two views of the same shape:
 - **[/reports/demo](/reports/demo)** — the polished design preview with synthetic data for three agents and four repos. Useful for screenshots and showing the full failure-mode breakdown the live view can't compute yet.
 
 Drill-downs:
-- [live drill-down per agent](/reports/live/agents/claude-code) · [tests overzicht (live: placeholder)](/reports/live/tests)
-- [demo drill-down per agent](/reports/demo/agents/cursor) · [tests overzicht (demo)](/reports/demo/tests)
+- [live drill-down per agent](/reports/live/agents/claude-code) · [tests overview (live)](/reports/live/tests)
+- [demo drill-down per agent](/reports/demo/agents/cursor) · [tests overview (demo)](/reports/demo/tests)
 
 Want a real repo on this layer? [Register a project →](/projects) — drops \`.tdd-md.json\` at the repo root, onboards in seconds. Per-commit judging on tracked branches lands in a follow-up sliver; live reporting from the GitHub API already works for the dogfood case (the tdd.md repo itself).
 
@@ -186,19 +186,19 @@ ${ctx.narrative.doingBody}
 
 `
     : "";
-  return `# tdd-discipline rapport · ${ctx.period}
+  return `# tdd-discipline report · ${ctx.period}
 
 ${ctx.bannerHtml}
 
-> **Periode** ${ctx.period} · **Scope** ${escape(ctx.scopeLabel)} · ${totalCommits.toLocaleString()} AI-toegeschreven commits.
+> **Period** ${ctx.period} · **Scope** ${escape(ctx.scopeLabel)} · ${totalCommits.toLocaleString()} AI-attributed commits.
 
 <div class="report-tiles">
 ${tiles}
 </div>
 
-${narrativeBlock}## wat dit getal *niet* meet
+${narrativeBlock}## what this number does *not* measure
 
-Discipline, niet code-kwaliteit. Hidden tests (zoals op de katas) bestaan niet voor productie-repos, dus *tautologische* tests en *zwak-geformuleerde* asserties blijven onzichtbaar voor de judge. Dit cijfer zegt: "de agent volgt de TDD-cyclus eerlijk". Het zegt niets over of de tests die hij schrijft het juiste beweren. Voor dat tweede signaal blijft kata-performance ([leaderboard](/leaderboard)) de proxy.
+Discipline, not code quality. Hidden tests (like the ones on the katas) don't exist for production repos, so *tautological* tests and *weakly-asserted* checks stay invisible to the judge. This number says: "the agent honours the TDD cycle". It says nothing about whether the tests it writes assert the right thing. For that second signal, kata performance ([leaderboard](/leaderboard)) remains the proxy.
 
 ---
 
@@ -226,9 +226,9 @@ export const agentDrilldownMd = (
 
 ${ctx.bannerHtml}
 
-> Discipline-score **${a.score} / 100** <span class="report-tile-trend ${arr.cls}">${arr.glyph} ${deltaStr}</span> over ${ctx.period}. ${a.commits.toLocaleString()} commits geanalyseerd, phase-coverage **${a.phaseCoveragePct}%**.
+> Discipline score **${a.score} / 100** <span class="report-tile-trend ${arr.cls}">${arr.glyph} ${deltaStr}</span> over ${ctx.period}. ${a.commits.toLocaleString()} commits analysed, phase coverage **${a.phaseCoveragePct}%**.
 
-## trend (30 dagen)
+## trend (30 days)
 
 <div class="${arr.cls === "down" ? "red" : arr.cls === "up" ? "green" : "muted"}">
 ${sparkline(a.trend)}
@@ -240,7 +240,7 @@ ${streakBox(a)}
 
 ${bars(a.failureMix)}
 
-Top issue dit kwartaal: **${escape(a.topIssueLabel)}** (${a.topIssuePct}% van commits).
+Top issue this quarter: **${escape(a.topIssueLabel)}** (${a.topIssuePct}% of commits).
 
 ## recent flagged
 
@@ -261,7 +261,7 @@ ${ctx.footerLinks}
 
 export const testsOverviewMd = (ctx: TestsOverviewContext): string => {
   if (ctx.unavailableNote) {
-    return `# tests overzicht
+    return `# tests overview
 
 ${ctx.bannerHtml}
 
@@ -275,23 +275,23 @@ ${ctx.bannerHtml}
   const failing = ctx.snapshots.reduce((s, r) => s + r.failing, 0);
   const snapshots = ctx.snapshots.map(snapshotBlock).join("\n");
   const stabRows = ctx.stability.map(stabilityRow).join("\n");
-  return `# tests overzicht
+  return `# tests overview
 
 ${ctx.bannerHtml}
 
-> Snapshot van de huidige test-stand per repo + stabiliteit van individuele tests over ${ctx.period}. Een hoge fail-count zonder deletion betekent dat de test echte regressies vangt; hoge fail+deletion is het signaal dat een test onder druk komt te staan — vaak het spoor van een agent die het makkelijker maakt zichzelf te laten "winnen".
+> Snapshot of the current test state per repo + stability of individual tests over ${ctx.period}. A high fail count with zero deletions means the test is actively catching regressions; high fail + deletion is the signal that a test is being squeezed — often the trace of an agent making it easier to "win".
 
-## huidige stand · per repo
+## current state · per repo
 
 <div class="test-snapshots">
 ${snapshots}
 </div>
 
-**Totaal**: ${total.toLocaleString()} tests · <span class="green">${passing.toLocaleString()} passing</span> · <span class="${failing > 0 ? "red" : "muted"}">${failing.toLocaleString()} failing</span>.
+**Total**: ${total.toLocaleString()} tests · <span class="green">${passing.toLocaleString()} passing</span> · <span class="${failing > 0 ? "red" : "muted"}">${failing.toLocaleString()} failing</span>.
 
-## test-stabiliteit · q1 2026
+## test stability · ${ctx.period}
 
-Top 12 meest-flappende tests dit kwartaal, met aantal pass/fail/deleted-events en de agent die de test het laatst heeft gebroken.
+Top tests by failure activity this period, with pass/fail/deleted counts and the agent who last broke the test.
 
 <table class="test-stability">
 <thead>
@@ -300,7 +300,7 @@ Top 12 meest-flappende tests dit kwartaal, met aantal pass/fail/deleted-events e
     <th class="num">pass</th>
     <th class="num">fail</th>
     <th class="num">del</th>
-    <th>laatst gebroken door</th>
+    <th>last broken by</th>
   </tr>
 </thead>
 <tbody>
@@ -308,14 +308,14 @@ ${stabRows}
 </tbody>
 </table>
 
-> ⚠ markeert tests waarbij dit kwartaal een test-deletion of weakening-event is gedetecteerd. In een echte setup linkt klik op een test-naam door naar de commit-historie van die specifieke test.
+> ⚠ marks tests where a test-deletion or weakening event has been detected this period. In a real setup, clicking a test name will link through to that test's commit history.
 
-## hoe lees je dit
+## how to read this
 
-- **Veel pass, weinig fail, 0 del**: gezond. Test doet wat hij moet, niemand sloopt 'm.
-- **Veel fail, 0 del**: test vangt actief regressies. Goed nieuws — discipline werkt.
-- **Fail én del > 0**: test wordt onder druk gezet. Coach de agent die 'm gebroken heeft (klik op het tag-icoon).
-- **Snapshot rood + stabiliteit hoog**: bekende, langlopende kapotte test. Apart onderwerp, niet per se een agent-probleem.
+- **Lots of pass, few fail, 0 del**: healthy. The test does what it should, nobody is sabotaging it.
+- **Lots of fail, 0 del**: the test is actively catching regressions. Good news — discipline is working.
+- **Fail and del > 0**: the test is under pressure. Coach the agent that broke it (click the tag icon).
+- **Snapshot red + stability high**: a known, long-running broken test. Separate concern, not necessarily an agent problem.
 
 ---