feat: consolidate tests

refactor: decompose tick_once into helper methods
Extract drain_spawns() (phases 1+4, dedup) and cleanup_dead_actors() (phase 7, 48 lines of complex death notification + GC logic). tick_once: 196 → 133 lines (−32%), max nesting depth: 5 → 3. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski
2026-02-13 14:48:35 +00:00 · 2026-02-13 14:48:35 +00:00 · 2026-02-13 14:48:35 +00:00 · 2026-02-13 14:48:35 +00:00 · 2026-02-13 14:48:35 +00:00 · 2026-02-13 14:48:35 +00:00
34 changed files with 4997 additions and 4831 deletions
--- a/crates/runtime-dashboard/src/actor_detail_html.rs
+++ b/crates/runtime-dashboard/src/actor_detail_html.rs
@ -0,0 +1,403 @@
+pub const ACTOR_DETAIL_HTML: &str = r##"<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>Actor Detail — Swactor Dashboard</title>
+<style>
+  * { margin: 0; padding: 0; box-sizing: border-box; }
+  body { font-family: 'Menlo', 'Consolas', 'Monaco', monospace; background: #0f1117; color: #e0e0e0; font-size: 13px; }
+
+  .header {
+    display: flex; align-items: center; justify-content: space-between;
+    padding: 12px 20px; background: #161822; border-bottom: 1px solid #2a2d3e;
+  }
+  .header h1 { font-size: 16px; font-weight: 600; color: #fff; }
+  .status-dot {
+    width: 10px; height: 10px; border-radius: 50%; background: #4caf50;
+    display: inline-block; margin-left: 8px; vertical-align: middle;
+  }
+  .status-dot.disconnected { background: #f44336; }
+
+  .header-left { display: flex; align-items: center; }
+  .nav-links { display: flex; gap: 4px; margin-left: 20px; }
+  .nav-link {
+    color: #888; text-decoration: none; font-size: 12px;
+    padding: 4px 10px; border-radius: 3px; transition: color 0.2s;
+  }
+  .nav-link:hover { color: #e0e0e0; }
+  .nav-link.active { color: #fff; background: #2a2d3e; }
+
+  .content { padding: 16px 20px; max-width: 900px; }
+
+  .breadcrumb { color: #555; font-size: 12px; margin-bottom: 12px; }
+  .breadcrumb a { color: #888; text-decoration: none; }
+  .breadcrumb a:hover { color: #e0e0e0; }
+
+  .info-card {
+    background: #161822; border: 1px solid #2a2d3e; border-radius: 6px;
+    padding: 16px; margin-bottom: 12px;
+  }
+  .info-row { display: flex; gap: 24px; margin-bottom: 6px; flex-wrap: wrap; }
+  .info-label { color: #888; font-size: 11px; text-transform: uppercase; }
+  .info-value { color: #fff; font-weight: 700; font-size: 15px; }
+  .info-value.healthy { color: #4caf50; }
+  .info-value.poisoned { color: #f44336; }
+
+  .stats-cards {
+    display: grid; grid-template-columns: repeat(4, 1fr); gap: 10px;
+    margin-bottom: 12px;
+  }
+  .stat-card {
+    background: #161822; border: 1px solid #2a2d3e; border-radius: 4px;
+    padding: 12px; text-align: center;
+  }
+  .stat-card .value { font-size: 22px; font-weight: 700; color: #fff; }
+  .stat-card .label { font-size: 10px; color: #888; text-transform: uppercase; margin-top: 2px; }
+
+  .sparkline-panel {
+    background: #161822; border: 1px solid #2a2d3e; border-radius: 6px;
+    padding: 14px; margin-bottom: 12px;
+  }
+  .sparkline-panel h3 { font-size: 11px; color: #888; text-transform: uppercase; letter-spacing: 1px; margin-bottom: 8px; }
+  .sparkline-panel svg { width: 100%; height: 50px; }
+
+  .type-breakdown {
+    background: #161822; border: 1px solid #2a2d3e; border-radius: 6px;
+    padding: 14px; margin-bottom: 12px;
+  }
+  .type-breakdown h3 { font-size: 11px; color: #888; text-transform: uppercase; letter-spacing: 1px; margin-bottom: 8px; }
+  .type-row { display: flex; align-items: center; gap: 8px; margin-bottom: 4px; font-size: 11px; }
+  .type-name { color: #e0e0e0; min-width: 180px; text-overflow: ellipsis; overflow: hidden; white-space: nowrap; }
+  .type-bar-bg { flex: 1; background: #1e2030; height: 14px; border-radius: 2px; overflow: hidden; }
+  .type-bar-fill { height: 100%; border-radius: 2px; }
+  .type-count { color: #888; min-width: 60px; text-align: right; }
+  .type-pct { color: #555; min-width: 40px; text-align: right; }
+
+  .logs-panel {
+    background: #161822; border: 1px solid #2a2d3e; border-radius: 6px;
+    padding: 14px; margin-bottom: 12px;
+  }
+  .logs-panel h3 {
+    font-size: 11px; color: #888; text-transform: uppercase; letter-spacing: 1px;
+    margin-bottom: 8px; display: flex; align-items: center; gap: 12px;
+  }
+  .level-filter { display: flex; gap: 4px; }
+  .level-btn {
+    background: #1e2030; border: 1px solid #2a2d3e; border-radius: 3px;
+    color: #888; font-size: 10px; padding: 1px 6px; cursor: pointer;
+    font-family: inherit;
+  }
+  .level-btn.active { border-color: #555; color: #fff; }
+  .level-btn.error { color: #f44336; }
+  .level-btn.warn { color: #ff9800; }
+  .level-btn.info { color: #2196f3; }
+  .level-btn.debug { color: #888; }
+  .level-btn.trace { color: #555; }
+
+  .log-list {
+    max-height: 400px; overflow-y: auto; font-size: 11px; line-height: 1.6;
+  }
+  .log-entry { display: flex; gap: 8px; padding: 1px 0; border-bottom: 1px solid #1a1c2e; }
+  .log-time { color: #555; white-space: nowrap; min-width: 80px; }
+  .log-level { font-weight: 700; min-width: 50px; }
+  .log-level.ERROR { color: #f44336; }
+  .log-level.WARN { color: #ff9800; }
+  .log-level.INFO { color: #2196f3; }
+  .log-level.DEBUG { color: #888; }
+  .log-level.TRACE { color: #555; }
+  .log-msg { color: #e0e0e0; word-break: break-all; }
+
+  ::-webkit-scrollbar { width: 6px; }
+  ::-webkit-scrollbar-track { background: #0f1117; }
+  ::-webkit-scrollbar-thumb { background: #2a2d3e; border-radius: 3px; }
+</style>
+</head>
+<body>
+<div class="header">
+  <div class="header-left">
+    <h1>
+      Swactor Runtime Dashboard
+      <span id="statusDot" class="status-dot"></span>
+    </h1>
+    <nav class="nav-links">
+      <a href="/" class="nav-link">Overview</a>
+      <a href="/actors" class="nav-link">Actors</a>
+    </nav>
+  </div>
+</div>
+
+<div class="content">
+  <div class="breadcrumb">
+    <a href="/">Overview</a> / <a href="/actors">Actors</a> / <span id="addrBreadcrumb">—</span>
+  </div>
+
+  <div class="info-card">
+    <div class="info-row">
+      <div><div class="info-label">Address</div><div class="info-value" id="addrFull">—</div></div>
+      <div><div class="info-label">Worker</div><div class="info-value" id="addrWorker">—</div></div>
+      <div><div class="info-label">Status</div><div class="info-value" id="addrStatus">—</div></div>
+    </div>
+    <div class="info-row">
+      <div><div class="info-label">Last Message Type</div><div class="info-value" id="addrLastMsg" style="color:#4caf50;font-size:13px;">—</div></div>
+    </div>
+  </div>
+
+  <div class="stats-cards">
+    <div class="stat-card"><div class="value" id="addrMsgs">0</div><div class="label">Messages</div></div>
+    <div class="stat-card"><div class="value" id="addrMailbox">0</div><div class="label">Mailbox</div></div>
+    <div class="stat-card"><div class="value" id="addrRate">0</div><div class="label">Msg/s</div></div>
+    <div class="stat-card"><div class="value" id="addrWorkerLoad">—</div><div class="label">Worker Load</div></div>
+  </div>
+
+  <div class="sparkline-panel">
+    <h3>Message Rate</h3>
+    <svg id="rateSpark" viewBox="0 0 400 50" preserveAspectRatio="none"></svg>
+  </div>
+
+  <div class="sparkline-panel">
+    <h3>Mailbox Depth</h3>
+    <svg id="mboxSpark" viewBox="0 0 400 50" preserveAspectRatio="none"></svg>
+  </div>
+
+  <div class="type-breakdown" id="typeBreakdown" style="display:none;">
+    <h3>Message Types</h3>
+    <div id="typeRows"></div>
+  </div>
+
+  <div class="logs-panel">
+    <h3>
+      <span>Logs</span>
+      <span id="logCount" style="color:#555;">(0)</span>
+      <div class="level-filter">
+        <button class="level-btn error active" data-level="ERROR" onclick="toggleLevel(this)">ERR</button>
+        <button class="level-btn warn active" data-level="WARN" onclick="toggleLevel(this)">WARN</button>
+        <button class="level-btn info active" data-level="INFO" onclick="toggleLevel(this)">INFO</button>
+        <button class="level-btn debug active" data-level="DEBUG" onclick="toggleLevel(this)">DBG</button>
+        <button class="level-btn trace active" data-level="TRACE" onclick="toggleLevel(this)">TRC</button>
+      </div>
+    </h3>
+    <div class="log-list" id="logList"></div>
+  </div>
+</div>
+
+<script>
+(function() {
+  var targetAddr = '__ACTOR_ADDR__';
+  var dot = document.getElementById('statusDot');
+
+  var history = { rates: [], mailbox: [], prev_msgs: 0 };
+
+  function formatAddr(addr) {
+    if (!addr) return '';
+    var bytes = Array.isArray(addr) ? addr : Object.values(addr);
+    var hex = '';
+    for (var i = 0; i < bytes.length; i++) {
+      hex += ('0' + bytes[i].toString(16)).slice(-2);
+    }
+    return hex;
+  }
+
+  function shortAddr(hex) {
+    return hex.length > 16 ? hex.substring(0, 16) + '\u2026' : hex;
+  }
+
+  function shortTypeName(full) {
+    if (!full) return '\u2014';
+    var parts = full.split('::');
+    return parts[parts.length - 1];
+  }
+
+  function updateSparklineSvg(svgEl, data, color) {
+    if (!data || data.length < 2) { svgEl.innerHTML = ''; return; }
+    var max = Math.max.apply(null, data);
+    if (max === 0) max = 1;
+    var w = 400, h = 50;
+    var step = w / (data.length - 1);
+    var points = data.map(function(v, i) {
+      return (i * step).toFixed(1) + ',' + (h - (v / max) * (h - 4) - 2).toFixed(1);
+    }).join(' ');
+    svgEl.innerHTML = '<polyline fill="none" stroke="' + color + '" stroke-width="2" points="' + points + '"/>';
+  }
+
+  function findActor(stats) {
+    if (!stats.actor_details) return null;
+    for (var i = 0; i < stats.actor_details.length; i++) {
+      var a = stats.actor_details[i];
+      var hex = formatAddr(a.address);
+      if (hex === targetAddr || hex.indexOf(targetAddr) === 0) return a;
+    }
+    return null;
+  }
+
+  function updateDetail(stats) {
+    var actor = findActor(stats);
+    if (!actor) return;
+
+    var hex = formatAddr(actor.address);
+    document.getElementById('addrBreadcrumb').textContent = shortAddr(hex);
+    document.getElementById('addrFull').textContent = hex;
+    document.getElementById('addrWorker').textContent = 'W' + actor.worker_id;
+
+    var statusEl = document.getElementById('addrStatus');
+    if (actor.poisoned) {
+      statusEl.textContent = 'POISONED';
+      statusEl.className = 'info-value poisoned';
+    } else {
+      statusEl.textContent = 'Healthy';
+      statusEl.className = 'info-value healthy';
+    }
+
+    document.getElementById('addrLastMsg').textContent = shortTypeName(actor.last_msg_type);
+    document.getElementById('addrMsgs').textContent = actor.messages_processed.toLocaleString();
+    document.getElementById('addrMailbox').textContent = actor.mailbox_depth;
+
+    // Compute rate
+    var rate = actor.messages_processed - history.prev_msgs;
+    if (rate < 0) rate = 0;
+    history.prev_msgs = actor.messages_processed;
+    history.rates.push(rate);
+    history.mailbox.push(actor.mailbox_depth);
+    if (history.rates.length > 300) { history.rates.shift(); history.mailbox.shift(); }
+
+    // Rate per second (SSE interval is ~200ms, so multiply by 5)
+    document.getElementById('addrRate').textContent = (rate * 5).toLocaleString();
+
+    // Worker load
+    if (stats.workers) {
+      var w = stats.workers.find(function(w) { return w.id === actor.worker_id; });
+      if (w) {
+        document.getElementById('addrWorkerLoad').textContent =
+          w.num_actors + ' actors, mbox ' + w.mailbox_depth;
+      }
+    }
+
+    updateSparklineSvg(document.getElementById('rateSpark'), history.rates, '#4caf50');
+    updateSparklineSvg(document.getElementById('mboxSpark'), history.mailbox, '#2196f3');
+
+    // Update message type breakdown
+    updateTypeBreakdown(actor.message_type_counts);
+  }
+
+  var typeColors = ['#4caf50','#2196f3','#ff9800','#9c27b0','#00bcd4','#f44336','#ffeb3b','#e91e63'];
+
+  function updateTypeBreakdown(types) {
+    var panel = document.getElementById('typeBreakdown');
+    var container = document.getElementById('typeRows');
+    if (!types || types.length === 0) { panel.style.display = 'none'; return; }
+    panel.style.display = '';
+    var total = 0;
+    for (var i = 0; i < types.length; i++) total += types[i][1];
+    if (total === 0) { panel.style.display = 'none'; return; }
+
+    var html = '';
+    for (var i = 0; i < types.length; i++) {
+      var name = types[i][0];
+      var count = types[i][1];
+      var pct = (count / total * 100).toFixed(1);
+      var barPct = (count / types[0][1] * 100).toFixed(1);
+      var color = typeColors[i % typeColors.length];
+      var shortName = name.split('::').pop();
+      html += '<div class="type-row">' +
+        '<span class="type-name" title="' + escapeHtml(name) + '">' + escapeHtml(shortName) + '</span>' +
+        '<div class="type-bar-bg"><div class="type-bar-fill" style="width:' + barPct + '%;background:' + color + ';"></div></div>' +
+        '<span class="type-count">' + count.toLocaleString() + '</span>' +
+        '<span class="type-pct">' + pct + '%</span>' +
+        '</div>';
+    }
+    container.innerHTML = html;
+  }
+
+  // ─── Logging ─────────────────────────────────────────────────
+  var activeLevels = { ERROR: true, WARN: true, INFO: true, DEBUG: true, TRACE: true };
+  var allLogs = [];
+  var logList = document.getElementById('logList');
+  var logCount = document.getElementById('logCount');
+  var autoScroll = true;
+
+  window.toggleLevel = function(btn) {
+    var lvl = btn.getAttribute('data-level');
+    activeLevels[lvl] = !activeLevels[lvl];
+    btn.classList.toggle('active');
+    renderLogs();
+  };
+
+  function formatLogTime(ms) {
+    var d = new Date(ms);
+    return ('0' + d.getHours()).slice(-2) + ':' +
+           ('0' + d.getMinutes()).slice(-2) + ':' +
+           ('0' + d.getSeconds()).slice(-2) + '.' +
+           ('00' + d.getMilliseconds()).slice(-3);
+  }
+
+  function renderLogs() {
+    var visible = allLogs.filter(function(e) { return activeLevels[e.level]; });
+    logCount.textContent = '(' + visible.length + ')';
+    var html = '';
+    for (var i = 0; i < visible.length; i++) {
+      var e = visible[i];
+      html += '<div class="log-entry">' +
+        '<span class="log-time">' + formatLogTime(e.timestamp_ms) + '</span>' +
+        '<span class="log-level ' + e.level + '">' + e.level + '</span>' +
+        '<span class="log-msg">' + escapeHtml(e.message) + '</span>' +
+        '</div>';
+    }
+    logList.innerHTML = html;
+    if (autoScroll) {
+      logList.scrollTop = logList.scrollHeight;
+    }
+  }
+
+  function escapeHtml(s) {
+    if (!s) return '';
+    return s.replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;');
+  }
+
+  function addLogEntries(events) {
+    for (var i = 0; i < events.length; i++) {
+      var e = events[i];
+      if (!e.actor_addr) continue;
+      // Match if event's actor_addr starts with or contains target
+      var a = e.actor_addr.toLowerCase();
+      if (a.indexOf(targetAddr.toLowerCase()) === 0 || a === targetAddr.toLowerCase()) {
+        allLogs.push(e);
+      }
+    }
+    // Keep bounded
+    while (allLogs.length > 500) allLogs.shift();
+    renderLogs();
+  }
+
+  // Fetch initial logs
+  fetch('/api/logs?actor=' + targetAddr + '&limit=200')
+    .then(function(r) { return r.json(); })
+    .then(function(data) { if (Array.isArray(data)) addLogEntries(data); })
+    .catch(function() {});
+
+  logList.addEventListener('scroll', function() {
+    autoScroll = (logList.scrollTop + logList.clientHeight >= logList.scrollHeight - 20);
+  });
+
+  // ─── SSE ────────────────────────────────────────────────────
+  var es = new EventSource('/events');
+
+  es.addEventListener('stats', function(e) {
+    try { updateDetail(JSON.parse(e.data)); } catch(err) { console.error(err); }
+  });
+
+  es.addEventListener('activity', function(e) {
+    try { addLogEntries(JSON.parse(e.data)); } catch(err) { console.error(err); }
+  });
+
+  es.addEventListener('done', function() {
+    dot.className = 'status-dot disconnected';
+    es.close();
+  });
+
+  es.onerror = function() { dot.className = 'status-dot disconnected'; };
+  es.onopen = function() { dot.className = 'status-dot'; };
+})();
+</script>
+</body>
+</html>
+"##;
--- a/crates/runtime-dashboard/src/actors.html
+++ b/crates/runtime-dashboard/src/actors.html
@ -60,7 +60,7 @@

  .panel {
    background: #161822; border: 1px solid #2a2d3e; border-radius: 6px;
-    padding: 14px; overflow: hidden;
+    padding: 14px; overflow: visible;
  }
  .panel h2 { font-size: 12px; color: #888; text-transform: uppercase; letter-spacing: 1px; margin-bottom: 10px; }

@ -75,7 +75,7 @@
  .stat-card .value { font-size: 22px; font-weight: 700; color: #fff; }
  .stat-card .label { font-size: 10px; color: #888; text-transform: uppercase; margin-top: 2px; }

-  canvas { width: 100%; height: 200px; }
+  canvas { width: 100%; height: 220px; }

  .search-wrap { margin-bottom: 10px; display: flex; align-items: center; gap: 12px; }
  .search-input {
@ -109,7 +109,12 @@
  tr.clickable { cursor: pointer; }
  tr.clickable:hover { background: #1a1d2c; }

-  .detail-panel { display: none; }
+  .detail-panel {
+    display: none; position: fixed; bottom: 12px; right: 12px;
+    width: 420px; max-height: 320px; overflow-y: auto;
+    background: #161822; border: 1px solid #2a2d3e; border-radius: 6px;
+    padding: 14px; z-index: 100; box-shadow: 0 4px 24px rgba(0,0,0,0.5);
+  }
  .detail-panel.visible { display: block; }
  .detail-header { display: flex; justify-content: space-between; align-items: center; margin-bottom: 10px; }
  .detail-close {
@ -118,17 +123,17 @@
  }
  .detail-close:hover { color: #e0e0e0; border-color: #555; }
  .detail-grid {
-    display: grid; grid-template-columns: repeat(3, 1fr); gap: 10px; margin-bottom: 12px;
+    display: grid; grid-template-columns: repeat(2, 1fr); gap: 8px; margin-bottom: 10px;
  }
-  .detail-item { background: #1c1f2e; border-radius: 4px; padding: 8px 10px; }
+  .detail-item { background: #1c1f2e; border-radius: 4px; padding: 6px 8px; }
  .detail-item .d-label { font-size: 10px; color: #888; text-transform: uppercase; }
-  .detail-item .d-value { font-size: 14px; font-weight: 700; color: #fff; margin-top: 2px; word-break: break-all; }
+  .detail-item .d-value { font-size: 13px; font-weight: 700; color: #fff; margin-top: 2px; word-break: break-all; }
  .poisoned-badge {
    background: #f44336; color: #fff; font-size: 10px; font-weight: 700;
    padding: 2px 6px; border-radius: 3px; letter-spacing: 0.5px;
  }
  .healthy-badge { color: #4caf50; font-size: 12px; }
-  canvas.sparkline { width: 100%; height: 60px; }
+  canvas.sparkline { width: 100%; height: 50px; }

  ::-webkit-scrollbar { width: 6px; }
  ::-webkit-scrollbar-track { background: #0f1117; }
@ -187,7 +192,16 @@
  <div class="panel full-width">
    <h2>All Actors <span id="actorCount" style="color:#555;font-weight:400;"></span></h2>
    <div class="search-wrap">
-      <input type="text" id="actorSearch" class="search-input" placeholder="Filter by address or worker..." />
+      <input type="text" id="actorSearch" class="search-input" placeholder="Filter by address, type, or worker..." />
+      <select id="workerFilter" class="search-input" style="width:120px;">
+        <option value="">All Workers</option>
+      </select>
+      <select id="statusFilter" class="search-input" style="width:120px;">
+        <option value="">All Status</option>
+        <option value="healthy">Healthy</option>
+        <option value="poisoned">Poisoned</option>
+      </select>
+      <input type="number" id="minDepth" class="search-input" style="width:100px;" placeholder="Min depth" min="0" />
      <span id="searchInfo" class="search-info"></span>
    </div>
    <div class="actor-list-wrap">
@ -207,23 +221,25 @@
    </div>
  </div>

-  <!-- Actor detail panel -->
-  <div id="detailPanel" class="panel full-width detail-panel">
-    <div class="detail-header">
-      <h2>Actor Detail <span id="detailAddr" style="color:#aaa;font-weight:400;"></span></h2>
-      <button class="detail-close" id="detailClose">Close</button>
-    </div>
-    <div class="detail-grid">
-      <div class="detail-item"><div class="d-label">Full Address</div><div class="d-value" id="detailFullAddr" style="font-size:11px;"></div></div>
-      <div class="detail-item"><div class="d-label">Worker</div><div class="d-value" id="detailWorker"></div></div>
-      <div class="detail-item"><div class="d-label">Mailbox Depth</div><div class="d-value" id="detailMailbox"></div></div>
-      <div class="detail-item"><div class="d-label">Messages Processed</div><div class="d-value" id="detailMsgCount"></div></div>
-      <div class="detail-item"><div class="d-label">Last Message Type</div><div class="d-value" id="detailLastMsg"></div></div>
-      <div class="detail-item"><div class="d-label">Status</div><div class="d-value" id="detailStatus"></div></div>
-    </div>
-    <h2>Mailbox Depth History</h2>
-    <canvas id="sparkline" class="sparkline"></canvas>
+
+</div>
+
+<!-- Actor detail overlay (fixed position, doesn't affect grid layout) -->
+<div id="detailPanel" class="detail-panel">
+  <div class="detail-header">
+    <h2 style="font-size:12px;color:#888;text-transform:uppercase;letter-spacing:1px;">Actor <span id="detailAddr" style="color:#aaa;font-weight:400;"></span></h2>
+    <button class="detail-close" id="detailClose">&times;</button>
  </div>
+  <div class="detail-grid">
+    <div class="detail-item"><div class="d-label">Address</div><div class="d-value" id="detailFullAddr" style="font-size:10px;"></div></div>
+    <div class="detail-item"><div class="d-label">Worker</div><div class="d-value" id="detailWorker"></div></div>
+    <div class="detail-item"><div class="d-label">Mailbox</div><div class="d-value" id="detailMailbox"></div></div>
+    <div class="detail-item"><div class="d-label">Messages</div><div class="d-value" id="detailMsgCount"></div></div>
+    <div class="detail-item"><div class="d-label">Last Type</div><div class="d-value" id="detailLastMsg"></div></div>
+    <div class="detail-item"><div class="d-label">Status</div><div class="d-value" id="detailStatus"></div></div>
+  </div>
+  <div style="font-size:10px;color:#888;text-transform:uppercase;letter-spacing:1px;margin-bottom:4px;">Mailbox History</div>
+  <canvas id="sparkline" class="sparkline"></canvas>
 </div>

 <script>
@ -374,11 +390,12 @@

    var maxCount = Math.max(1, Math.max.apply(null, counts));
    var barW = Math.max(12, Math.floor((W - 40) / buckets.length) - 8);
+    var topPad = 20;
    var chartH = H - 35;

    for (var i = 0; i < buckets.length; i++) {
      var x = 20 + i * (barW + 8);
-      var h = (counts[i] / maxCount) * (chartH - 10);
+      var h = (counts[i] / maxCount) * (chartH - topPad);
      depthCtx.fillStyle = bucketColors[i];
      depthCtx.globalAlpha = 0.85;
      depthCtx.fillRect(x, chartH - h, barW, h);
@ -424,11 +441,12 @@

    var maxCount = Math.max(1, Math.max.apply(null, entries.map(function(e) { return e.count; })));
    var barW = Math.max(12, Math.floor((W - 40) / entries.length) - 8);
+    var topPad = 20;
    var chartH = H - 35;

    for (var i = 0; i < entries.length; i++) {
      var x = 20 + i * (barW + 8);
-      var h = (entries[i].count / maxCount) * (chartH - 10);
+      var h = (entries[i].count / maxCount) * (chartH - topPad);
      workerCtx.fillStyle = colors[entries[i].id % colors.length];
      workerCtx.globalAlpha = 0.85;
      workerCtx.fillRect(x, chartH - h, barW, h);
@ -449,13 +467,29 @@

  function renderActorTable() {
    var filter = document.getElementById('actorSearch').value.toLowerCase();
-    var filtered = currentActors;
-    if (filter) {
-      filtered = currentActors.filter(function(a) {
+    var workerFilter = document.getElementById('workerFilter').value;
+    var statusFilter = document.getElementById('statusFilter').value;
+    var minDepthVal = document.getElementById('minDepth').value;
+    var minDepth = minDepthVal ? parseInt(minDepthVal, 10) : 0;
+
+    var filtered = currentActors.filter(function(a) {
+      // Text search
+      if (filter) {
        var hex = addrToHex(a.address).toLowerCase();
-        return hex.indexOf(filter) >= 0 || ('w' + a.worker_id).indexOf(filter) >= 0;
-      });
-    }
+        var msgType = (a.last_msg_type || '').toLowerCase();
+        if (hex.indexOf(filter) < 0 && ('w' + a.worker_id).indexOf(filter) < 0 && msgType.indexOf(filter) < 0) {
+          return false;
+        }
+      }
+      // Worker filter
+      if (workerFilter && a.worker_id !== parseInt(workerFilter, 10)) return false;
+      // Status filter
+      if (statusFilter === 'healthy' && a.poisoned) return false;
+      if (statusFilter === 'poisoned' && !a.poisoned) return false;
+      // Min depth
+      if (minDepth > 0 && a.mailbox_depth < minDepth) return false;
+      return true;
+    });

    // Sort
    filtered.sort(function(a, b) {
@ -531,14 +565,14 @@
  // ── Focus / detail panel ───────────────────────────────
  function focusActor(fullHex) {
    focusedAddrHex = fullHex;
-    document.getElementById('detailPanel').className = 'panel full-width detail-panel visible';
+    document.getElementById('detailPanel').className = 'detail-panel visible';
    updateDetailPanel();
    renderActorTable();
  }

  function clearFocus() {
    focusedAddrHex = null;
-    document.getElementById('detailPanel').className = 'panel full-width detail-panel';
+    document.getElementById('detailPanel').className = 'detail-panel';
    renderActorTable();
  }

@ -558,8 +592,8 @@
      return;
    }

-    document.getElementById('detailAddr').textContent = addrToHex(actor.address);
-    document.getElementById('detailFullAddr').textContent = focusedAddrHex;
+    document.getElementById('detailAddr').innerHTML = '<a href="/actor/' + focusedAddrHex + '" style="color:#aaa;text-decoration:none;">' + escapeHtml(addrToHex(actor.address)) + '</a>';
+    document.getElementById('detailFullAddr').innerHTML = '<a href="/actor/' + focusedAddrHex + '" style="color:#fff;text-decoration:none;">' + escapeHtml(focusedAddrHex) + '</a>';
    document.getElementById('detailWorker').textContent = 'W' + actor.worker_id;
    document.getElementById('detailMailbox').textContent = actor.mailbox_depth;
    document.getElementById('detailMsgCount').textContent = (actor.messages_processed || 0).toLocaleString();
@ -647,11 +681,17 @@
    });
  }

-  // ── Search handler ─────────────────────────────────────
+  // ── Search/filter handlers ─────────────────────────────
  document.getElementById('actorSearch').addEventListener('keyup', function() {
    clearTimeout(searchTimer);
    searchTimer = setTimeout(renderActorTable, 150);
  });
+  document.getElementById('workerFilter').addEventListener('change', renderActorTable);
+  document.getElementById('statusFilter').addEventListener('change', renderActorTable);
+  document.getElementById('minDepth').addEventListener('input', function() {
+    clearTimeout(searchTimer);
+    searchTimer = setTimeout(renderActorTable, 150);
+  });

  // ── Status helpers ─────────────────────────────────────
  function setStatus(s) {
@ -697,6 +737,21 @@
        if (!liveAddrs[key]) delete depthHistory[key];
      }

+      // Update worker filter dropdown
+      var wSelect = document.getElementById('workerFilter');
+      var curVal = wSelect.value;
+      var workerIds = {};
+      for (var i = 0; i < currentActors.length; i++) workerIds[currentActors[i].worker_id] = true;
+      var wids = Object.keys(workerIds).sort(function(a,b) { return +a - +b; });
+      wSelect.innerHTML = '<option value="">All Workers</option>';
+      wids.forEach(function(wid) {
+        var opt = document.createElement('option');
+        opt.value = wid;
+        opt.textContent = 'W' + wid;
+        wSelect.appendChild(opt);
+      });
+      wSelect.value = curVal;
+
      renderActorTable();
      if (focusedAddrHex) updateDetailPanel();
    } catch(err) { console.error('stats parse error', err); }
--- a/crates/runtime-dashboard/src/actors_html.rs
+++ b/crates/runtime-dashboard/src/actors_html.rs
@ -1 +1 @@
-pub const ACTORS_HTML: &str = include_str!("actors.html");
+pub const ACTORS_HTML: &str = include_str\!("actors.html");
--- a/crates/runtime-dashboard/src/collector.rs
+++ b/crates/runtime-dashboard/src/collector.rs
@ -60,6 +60,7 @@ impl StatsHook for StatsCollector {
                last_msg_type: s.last_msg_type.map(|t| t.to_string()),
                messages_processed: s.messages_processed,
                poisoned: s.poisoned,
+                message_type_counts: s.message_type_counts.iter().map(|(k, v)| (k.to_string(), *v)).collect(),
            }));
        }
    }
--- a/crates/runtime-dashboard/src/dashboard.html
+++ b/crates/runtime-dashboard/src/dashboard.html
@ -77,6 +77,17 @@
  .chart-panel { grid-row: span 2; }
  canvas#workerChart { width: 100%; height: 200px; }

+  .worker-cards { display: flex; flex-direction: column; gap: 6px; }
+  .worker-card {
+    display: flex; align-items: center; gap: 10px;
+    background: #1c1f2e; border-radius: 4px; padding: 6px 10px;
+  }
+  .worker-card .wc-id { font-weight: 700; min-width: 32px; }
+  .worker-card .wc-bar-wrap { flex: 1; height: 14px; background: #0f1117; border-radius: 2px; overflow: hidden; display: flex; }
+  .worker-card .wc-bar-seg { height: 100%; }
+  .worker-card .wc-stats { font-size: 11px; color: #888; min-width: 200px; text-align: right; }
+  .worker-card .wc-spark { display: inline-flex; gap: 4px; margin-left: 6px; }
+
  .actor-table-wrap { max-height: 200px; overflow-y: auto; }
  .actor-table-wrap table { width: 100%; border-collapse: collapse; }
  .actor-table-wrap th, .actor-table-wrap td {
@ -113,6 +124,8 @@
  .worker-group-header .wid { font-weight: 700; }
  .worker-group-header .summary { color: #888; font-size: 11px; }
  .worker-group-header .toggle { color: #555; font-size: 14px; }
+  .worker-group-header .sparkline-wrap { display: inline-flex; gap: 8px; margin-left: 12px; }
+  .worker-group-header .sparkline-wrap svg { vertical-align: middle; }
  .worker-group-body { display: none; }
  .worker-group.open .worker-group-body { display: block; }
  .worker-group-body table { width: 100%; border-collapse: collapse; }
@ -152,11 +165,17 @@
  <div id="progressFill" class="progress-fill"></div>
 </div>

+<div id="warningBanner" style="display:none;padding:8px 20px;background:#1c1f2e;border-bottom:1px solid #2a2d3e;font-size:12px;"></div>
 <div class="grid">
  <div class="panel chart-panel">
-    <h2>Worker Distribution</h2>
-    <canvas id="workerChart"></canvas>
-    <div id="workerLegend" style="margin-top:8px;font-size:11px;color:#888;"></div>
+    <h2>Worker Utilization</h2>
+    <div id="workerCards" class="worker-cards"></div>
+    <div style="margin-top:6px;font-size:10px;color:#555;">
+      <span style="color:#4caf50;">\u25A0</span> processing
+      <span style="color:#2196f3;">\u25A0</span> delivery
+      <span style="color:#00bcd4;">\u25A0</span> spawns
+      <span style="color:#f44336;">\u25A0</span> overhead
+    </div>
  </div>

  <div class="panel">
@ -203,6 +222,7 @@
  var isReplay = (DASHBOARD_MODE === 'replay');
  var lastUptimeMs = null;
  var lastStatsTime = null;
+  var workerHistory = {}; // { id: { message_rates: [], mailbox_depths: [] } }

  var dot = document.getElementById('statusDot');
  var uptimeLabel = document.getElementById('uptimeLabel');
@ -240,48 +260,83 @@
  }
  setInterval(updateUptime, 1000);

-  var canvas = document.getElementById('workerChart');
-  var ctx = canvas.getContext('2d');
  var colors = ['#4caf50','#2196f3','#ff9800','#f44336','#9c27b0','#00bcd4','#ffeb3b','#e91e63'];

-  function drawWorkerChart(workers) {
-    var dpr = window.devicePixelRatio || 1;
-    var rect = canvas.getBoundingClientRect();
-    canvas.width = rect.width * dpr;
-    canvas.height = rect.height * dpr;
-    ctx.scale(dpr, dpr);
-    var W = rect.width, H = rect.height;
-    ctx.clearRect(0, 0, W, H);
+  var phaseColors = ['#4caf50', '#2196f3', '#00bcd4', '#f44336'];
+  // Group tick phases: processing=2, delivery=1+4, spawns=0+3, overhead=5

-    if (!workers || workers.length === 0) return;
+  function computePhases(timings) {
+    if (!timings || timings.length === 0) return [0.25, 0.25, 0.25, 0.25];
+    var sums = [0,0,0,0,0,0];
+    var active = 0;
+    for (var i = 0; i < timings.length; i++) {
+      var t = timings[i];
+      if (t.did_work) active++;
+      for (var p = 0; p < 6 && p < t.phase_us.length; p++) sums[p] += t.phase_us[p];
+    }
+    var total = sums.reduce(function(a,b) { return a+b; }, 0);
+    if (total === 0) return [0.25, 0.25, 0.25, 0.25];
+    var processing = sums[2] / total;
+    var delivery = (sums[1] + sums[4]) / total;
+    var spawns = (sums[0] + sums[3]) / total;
+    var overhead = sums[5] / total;
+    var load = timings.length > 0 ? active / timings.length : 0;
+    return { fracs: [processing, delivery, spawns, overhead], load: load };
+  }

-    var maxActors = Math.max(1, Math.max.apply(null, workers.map(function(w) { return w.num_actors; })));
-    var barW = Math.max(8, Math.floor((W - 40) / workers.length) - 6);
-    var chartH = H - 30;
+  function renderWorkerCards(data) {
+    var container = document.getElementById('workerCards');
+    if (!data.workers) return;
+    container.innerHTML = '';

-    workers.forEach(function(w, i) {
-      var x = 20 + i * (barW + 6);
-      var h = (w.num_actors / maxActors) * (chartH * 0.45);
-      ctx.fillStyle = colors[i % colors.length];
-      ctx.globalAlpha = 0.8;
-      ctx.fillRect(x, chartH * 0.5 - h, barW, h);
+    data.workers.forEach(function(w, idx) {
+      var timings = data.tick_timings ? data.tick_timings[idx] : null;
+      var phases = computePhases(timings);
+      var load = phases.load || 0;
+      var fracs = phases.fracs || [0.25, 0.25, 0.25, 0.25];

-      var mh = Math.min(w.mailbox_depth * 2, chartH * 0.4);
-      ctx.globalAlpha = 0.4;
-      ctx.fillRect(x, chartH * 0.55, barW, mh);
+      var card = document.createElement('div');
+      card.className = 'worker-card';

-      ctx.globalAlpha = 1;
-      ctx.fillStyle = '#888';
-      ctx.font = '10px monospace';
-      ctx.textAlign = 'center';
-      ctx.fillText('W' + w.id, x + barW / 2, H - 2);
+      // ID
+      var idSpan = document.createElement('span');
+      idSpan.className = 'wc-id';
+      idSpan.style.color = colors[w.id % colors.length];
+      idSpan.textContent = 'W' + w.id;
+      card.appendChild(idSpan);
+
+      // Phase bar
+      var barWrap = document.createElement('span');
+      barWrap.className = 'wc-bar-wrap';
+      var filledPct = Math.round(load * 100);
+      for (var p = 0; p < 4; p++) {
+        var seg = document.createElement('span');
+        seg.className = 'wc-bar-seg';
+        seg.style.width = (fracs[p] * filledPct) + '%';
+        seg.style.background = phaseColors[p];
+        barWrap.appendChild(seg);
+      }
+      card.appendChild(barWrap);
+
+      // Sparklines
+      var sparkWrap = document.createElement('span');
+      sparkWrap.className = 'wc-spark';
+      var wh = workerHistory[w.id];
+      if (wh) {
+        sparkWrap.innerHTML = renderSparklineSvg(wh.message_rates, 60, 14, '#4caf50');
+      }
+      card.appendChild(sparkWrap);
+
+      // Stats
+      var statsSpan = document.createElement('span');
+      statsSpan.className = 'wc-stats';
+      statsSpan.textContent = w.num_actors + ' actors  ' +
+        w.messages_processed.toLocaleString() + ' msgs  mbox ' + w.mailbox_depth +
+        '  ' + Math.round(load * 100) + '%';
+      card.appendChild(statsSpan);
+
+      container.appendChild(card);
    });
-
-    var legend = document.getElementById('workerLegend');
-    legend.innerHTML = workers.map(function(w, i) {
-      return '<span style="color:' + colors[i % colors.length] + '">W' + w.id +
-        ': ' + w.num_actors + ' actors, ' + w.messages_processed + ' msgs, mbox ' + w.mailbox_depth + '</span>';
-    }).join(' &nbsp;|&nbsp; ');
  }

  function updateStats(data) {
@ -298,7 +353,7 @@
    document.getElementById('statWorkers').textContent = data.num_workers || 0;
    document.getElementById('statMailbox').textContent = totalMailbox;

-    drawWorkerChart(data.workers);
+    renderWorkerCards(data);

    var tbody = document.getElementById('actorTableBody');
    tbody.innerHTML = '';
@ -316,7 +371,7 @@
          hex += '\u2026';
        }
        var tr = document.createElement('tr');
-        tr.innerHTML = '<td style="color:#aaa;font-size:11px;">' + hex + '</td><td>W' + wid + '</td>';
+        tr.innerHTML = '<td style="color:#aaa;font-size:11px;"><a href="/actor/' + hex + '" style="color:#aaa;text-decoration:none;">' + hex + '</a></td><td>W' + wid + '</td>';
        tbody.appendChild(tr);
      });
      if (data.actors.length > 200) {
@ -347,6 +402,34 @@
    return parts[parts.length - 1];
  }

+  function renderSparklineSvg(data, w, h, color) {
+    if (!data || data.length < 2) return '';
+    var max = Math.max.apply(null, data);
+    if (max === 0) max = 1;
+    var step = w / (data.length - 1);
+    var points = data.map(function(v, i) {
+      return (i * step).toFixed(1) + ',' + (h - (v / max) * (h - 2) - 1).toFixed(1);
+    }).join(' ');
+    return '<svg width="' + w + '" height="' + h + '" style="vertical-align:middle">' +
+      '<polyline fill="none" stroke="' + color + '" stroke-width="1.5" points="' + points + '"/></svg>';
+  }
+
+  function pushHistorySample(stats) {
+    if (!stats.workers) return;
+    stats.workers.forEach(function(w) {
+      if (!workerHistory[w.id]) {
+        workerHistory[w.id] = { message_rates: [], mailbox_depths: [], prev_msgs: w.messages_processed };
+      }
+      var wh = workerHistory[w.id];
+      var rate = w.messages_processed - wh.prev_msgs;
+      if (rate < 0) rate = 0;
+      wh.prev_msgs = w.messages_processed;
+      wh.message_rates.push(rate);
+      wh.mailbox_depths.push(w.mailbox_depth);
+      if (wh.message_rates.length > 300) { wh.message_rates.shift(); wh.mailbox_depths.shift(); }
+    });
+  }
+
  function updateWorkerDetails(data) {
    var container = document.getElementById('workerDetailContainer');
    if (!data.workers) return;
@ -377,8 +460,17 @@
      var hdr = document.createElement('div');
      hdr.className = 'worker-group-header';
      var panicHtml = g.info.panics > 0 ? ', <span style="color:#f44336">' + g.info.panics + ' panics</span>' : '';
+      var wh = workerHistory[wid];
+      var sparkHtml = '';
+      if (wh) {
+        sparkHtml = '<span class="sparkline-wrap">' +
+          renderSparklineSvg(wh.message_rates, 80, 16, '#4caf50') +
+          renderSparklineSvg(wh.mailbox_depths, 80, 16, '#2196f3') +
+          '</span>';
+      }
      hdr.innerHTML =
        '<span class="wid" style="color:' + colors[wid % colors.length] + '">W' + wid + '</span>' +
+        sparkHtml +
        '<span class="summary">' + g.actors.length + ' actors, ' +
          g.info.messages_processed.toLocaleString() + ' msgs, mbox ' + g.info.mailbox_depth + panicHtml + '</span>' +
        '<span class="toggle">' + (isOpen ? '\u25BC' : '\u25B6') + '</span>';
@ -400,7 +492,7 @@
          var msgShort = hasMsg ? shortTypeName(a.last_msg_type) : 'none';
          var msgClass = hasMsg ? 'msg-type' : 'msg-type none';
          var title = hasMsg ? ' title="' + escapeHtml(a.last_msg_type) + '"' : '';
-          rows += '<tr><td style="color:#aaa;">' + hex + '</td>' +
+          rows += '<tr><td style="color:#aaa;"><a href="/actor/' + hex + '" style="color:#aaa;text-decoration:none;">' + hex + '</a></td>' +
            '<td>' + a.mailbox_depth + '</td>' +
            '<td class="' + msgClass + '"' + title + '>' + escapeHtml(msgShort) + '</td></tr>';
        });
@ -467,7 +559,44 @@
  var es = new EventSource('/events');

  es.addEventListener('stats', function(e) {
-    try { updateStats(JSON.parse(e.data)); } catch(err) { console.error('stats parse error', err); }
+    try {
+      var data = JSON.parse(e.data);
+      pushHistorySample(data);
+      updateStats(data);
+    } catch(err) { console.error('stats parse error', err); }
+  });
+
+  es.addEventListener('history', function(e) {
+    try {
+      var data = JSON.parse(e.data);
+      if (data.workers) {
+        data.workers.forEach(function(w) {
+          workerHistory[w.id] = {
+            message_rates: w.message_rates || [],
+            mailbox_depths: w.mailbox_depths || [],
+            prev_msgs: 0
+          };
+        });
+      }
+    } catch(err) { console.error('history parse error', err); }
+  });
+
+  es.addEventListener('warnings', function(e) {
+    try {
+      var warnings = JSON.parse(e.data);
+      var banner = document.getElementById('warningBanner');
+      if (warnings.length === 0) {
+        banner.style.display = 'none';
+        return;
+      }
+      banner.style.display = 'block';
+      var sevColors = {critical:'#f44336',high:'#ff5722',medium:'#ff9800',low:'#888'};
+      var html = warnings.map(function(w) {
+        var c = sevColors[w.severity] || '#888';
+        return '<span style="color:' + c + ';">\u26A0 ' + w.description + '</span>';
+      }).join(' &nbsp; ');
+      banner.innerHTML = '<span style="color:#ff9800;font-weight:700;">WARNINGS (' + warnings.length + ')</span> &nbsp; ' + html;
+    } catch(err) { console.error('warnings parse error', err); }
  });

  es.addEventListener('activity', function(e) {
--- a/crates/runtime-dashboard/src/dashboard_html.rs
+++ b/crates/runtime-dashboard/src/dashboard_html.rs
@ -1 +1 @@
-pub const DASHBOARD_HTML: &str = include_str!("dashboard.html");
+pub const DASHBOARD_HTML: &str = include_str\!("dashboard.html");
--- a/crates/runtime-dashboard/src/distribution.html
+++ b/crates/runtime-dashboard/src/distribution.html
@ -40,9 +40,9 @@

  .graph-panel {
    grid-row: 1 / 3; border-right: 1px solid #2a2d3e; position: relative;
-    min-height: 0;
+    min-height: 0; overflow: hidden;
  }
-  .graph-panel canvas { width: 100%; height: 100%; display: block; }
+  .graph-panel canvas { position: absolute; top: 0; left: 0; width: 100%; height: 100%; display: block; }

  .side-panel { display: flex; flex-direction: column; overflow: hidden; min-height: 0; }

@ -176,8 +176,17 @@
      </div>
    </div>
    <div class="bottom-section">
-      <h2>Recent Probes</h2>
-      <div class="scroll-wrap" id="probesWrap"></div>
+      <h2>Recent Probes <span id="probeCount" style="color:#555;font-weight:400;"></span></h2>
+      <div class="scroll-wrap">
+        <table style="width:100%;border-collapse:collapse;">
+          <thead><tr>
+            <th style="color:#888;font-weight:500;font-size:11px;">State</th>
+            <th style="color:#888;font-weight:500;font-size:11px;">Node</th>
+            <th style="color:#888;font-weight:500;font-size:11px;">Address</th>
+          </tr></thead>
+          <tbody id="probesBody"></tbody>
+        </table>
+      </div>
    </div>
    <div class="bottom-section">
      <h2>Routing Buckets</h2>
@ -653,14 +662,26 @@
      cacheBody.appendChild(tr);
    }

-    // Recent probes
-    var probesWrap = document.getElementById('probesWrap');
-    probesWrap.innerHTML = '';
+    // Recent probes — cross-reference with members for state + address
+    var memberMap = {};
+    for (var i = 0; i < d.members.length; i++) {
+      memberMap[d.members[i].node_id] = d.members[i];
+    }
+    var probesBody = document.getElementById('probesBody');
+    probesBody.innerHTML = '';
+    document.getElementById('probeCount').textContent = '(' + d.recent_probe_targets.length + ')';
    for (var i = d.recent_probe_targets.length - 1; i >= 0; i--) {
-      var div = document.createElement('div');
-      div.style.cssText = 'padding:2px 0;color:#aaa;font-size:11px;border-bottom:1px solid #1c1f2e;';
-      div.textContent = d.recent_probe_targets[i].substring(0, 16) + '\u2026';
-      probesWrap.appendChild(div);
+      var pid = d.recent_probe_targets[i];
+      var mem = memberMap[pid];
+      var state = mem ? mem.state : 'unknown';
+      var addr = mem ? mem.addr : '\u2014';
+      var cls = 'state-' + state;
+      var tr = document.createElement('tr');
+      tr.innerHTML =
+        '<td class="' + cls + '" style="font-size:10px;">' + state + '</td>' +
+        '<td style="color:#aaa;font-size:10px;">' + pid.substring(0, 12) + '\u2026</td>' +
+        '<td style="font-size:10px;">' + addr + '</td>';
+      probesBody.appendChild(tr);
    }

    // Routing bucket histogram
--- a/crates/runtime-dashboard/src/distribution_html.rs
+++ b/crates/runtime-dashboard/src/distribution_html.rs
@ -1 +1 @@
-pub const DISTRIBUTION_HTML: &str = include_str!("distribution.html");
+pub const DISTRIBUTION_HTML: &str = include_str\!("distribution.html");
--- a/crates/runtime-dashboard/src/history.rs
+++ b/crates/runtime-dashboard/src/history.rs
@ -0,0 +1,344 @@
+//! In-process time-series history for dashboard sparklines and trend detection.
+//!
+//! Stores bounded ring buffers of per-worker and per-actor stats, sampled at
+//! a configurable interval. All data is kept in memory with automatic eviction
+//! of the oldest samples when capacity is reached.
+
+use std::collections::{HashMap, VecDeque};
+use std::sync::RwLock;
+
+use swactor::actor::ActorAddress;
+use swactor::stats::{ActorInfo, RuntimeStats};
+
+/// Configuration for history collection.
+#[derive(Debug, Clone)]
+pub struct HistoryConfig {
+    /// Maximum samples per worker (default: 300 = 5 min at 1/sec).
+    pub max_worker_samples: usize,
+    /// Maximum samples per actor (default: 300).
+    pub max_actor_samples: usize,
+    /// Maximum number of actors tracked (LRU eviction). Default: 1000.
+    pub max_tracked_actors: usize,
+}
+
+impl Default for HistoryConfig {
+    fn default() -> Self {
+        Self {
+            max_worker_samples: 300,
+            max_actor_samples: 300,
+            max_tracked_actors: 1000,
+        }
+    }
+}
+
+/// Time-series data for a single worker.
+#[derive(Debug, Clone)]
+pub struct WorkerHistory {
+    pub message_rates: VecDeque<f64>,
+    pub mailbox_depths: VecDeque<u64>,
+    pub actor_counts: VecDeque<u32>,
+    prev_messages: u64,
+}
+
+impl WorkerHistory {
+    fn new() -> Self {
+        Self {
+            message_rates: VecDeque::new(),
+            mailbox_depths: VecDeque::new(),
+            actor_counts: VecDeque::new(),
+            prev_messages: 0,
+        }
+    }
+
+    fn push(&mut self, messages_processed: u64, mailbox_depth: usize, num_actors: usize, cap: usize) {
+        let rate = messages_processed.saturating_sub(self.prev_messages) as f64;
+        self.prev_messages = messages_processed;
+
+        push_bounded(&mut self.message_rates, rate, cap);
+        push_bounded(&mut self.mailbox_depths, mailbox_depth as u64, cap);
+        push_bounded(&mut self.actor_counts, num_actors as u32, cap);
+    }
+}
+
+/// Time-series data for a single actor.
+#[derive(Debug, Clone)]
+pub struct ActorHistory {
+    pub mailbox_depths: VecDeque<u64>,
+    pub message_rates: VecDeque<f64>,
+    prev_messages: u64,
+    last_seen_sample: u64,
+}
+
+impl ActorHistory {
+    fn new(sample_counter: u64) -> Self {
+        Self {
+            mailbox_depths: VecDeque::new(),
+            message_rates: VecDeque::new(),
+            prev_messages: 0,
+            last_seen_sample: sample_counter,
+        }
+    }
+
+    fn push(&mut self, info: &ActorInfo, cap: usize, sample_counter: u64) {
+        let rate = info.messages_processed.saturating_sub(self.prev_messages) as f64;
+        self.prev_messages = info.messages_processed;
+        self.last_seen_sample = sample_counter;
+
+        push_bounded(&mut self.mailbox_depths, info.mailbox_depth as u64, cap);
+        push_bounded(&mut self.message_rates, rate, cap);
+    }
+}
+
+fn push_bounded<T>(buf: &mut VecDeque<T>, val: T, cap: usize) {
+    if buf.len() >= cap {
+        buf.pop_front();
+    }
+    buf.push_back(val);
+}
+
+/// Thread-safe history store. Written by the sampler, read by SSE/TUI.
+pub struct DashboardHistory {
+    inner: RwLock<HistoryInner>,
+    config: HistoryConfig,
+}
+
+struct HistoryInner {
+    workers: Vec<WorkerHistory>,
+    actors: HashMap<ActorAddress, ActorHistory>,
+    sample_counter: u64,
+}
+
+impl DashboardHistory {
+    pub fn new(config: HistoryConfig) -> Self {
+        Self {
+            inner: RwLock::new(HistoryInner {
+                workers: Vec::new(),
+                actors: HashMap::new(),
+                sample_counter: 0,
+            }),
+            config,
+        }
+    }
+
+    /// Record a stats snapshot. Called by the sampler thread.
+    pub fn record(&self, stats: &RuntimeStats) {
+        let mut inner = self.inner.write().unwrap();
+        inner.sample_counter += 1;
+        let counter = inner.sample_counter;
+
+        // Resize workers vec if needed
+        while inner.workers.len() < stats.workers.len() {
+            inner.workers.push(WorkerHistory::new());
+        }
+
+        // Record per-worker data
+        for w in &stats.workers {
+            if let Some(wh) = inner.workers.get_mut(w.id) {
+                wh.push(
+                    w.messages_processed,
+                    w.mailbox_depth,
+                    w.num_actors,
+                    self.config.max_worker_samples,
+                );
+            }
+        }
+
+        // Record per-actor data
+        for a in &stats.actor_details {
+            let ah = inner.actors.entry(a.address).or_insert_with(|| ActorHistory::new(counter));
+            ah.push(a, self.config.max_actor_samples, counter);
+        }
+
+        // LRU eviction: remove actors not seen recently if over capacity
+        if inner.actors.len() > self.config.max_tracked_actors {
+            let mut entries: Vec<(ActorAddress, u64)> = inner
+                .actors
+                .iter()
+                .map(|(addr, ah)| (*addr, ah.last_seen_sample))
+                .collect();
+            entries.sort_by_key(|&(_, seen)| seen);
+            let to_remove = inner.actors.len() - self.config.max_tracked_actors;
+            for (addr, _) in entries.into_iter().take(to_remove) {
+                inner.actors.remove(&addr);
+            }
+        }
+    }
+
+    /// Get a snapshot of worker history for rendering sparklines.
+    /// Returns Vec indexed by worker_id, each containing recent message rates.
+    pub fn worker_sparklines(&self) -> Vec<Vec<u64>> {
+        let inner = self.inner.read().unwrap();
+        inner
+            .workers
+            .iter()
+            .map(|wh| wh.message_rates.iter().map(|r| *r as u64).collect())
+            .collect()
+    }
+
+    /// Get worker mailbox depth history.
+    pub fn worker_mailbox_sparklines(&self) -> Vec<Vec<u64>> {
+        let inner = self.inner.read().unwrap();
+        inner
+            .workers
+            .iter()
+            .map(|wh| wh.mailbox_depths.iter().copied().collect())
+            .collect()
+    }
+
+    /// Get sparkline data for a specific actor.
+    pub fn actor_sparkline(&self, addr: &ActorAddress) -> Option<(Vec<u64>, Vec<u64>)> {
+        let inner = self.inner.read().unwrap();
+        inner.actors.get(addr).map(|ah| {
+            let mailbox: Vec<u64> = ah.mailbox_depths.iter().copied().collect();
+            let rates: Vec<u64> = ah.message_rates.iter().map(|r| *r as u64).collect();
+            (mailbox, rates)
+        })
+    }
+
+    /// Get total sample count (useful for knowing if history is available).
+    pub fn sample_count(&self) -> u64 {
+        self.inner.read().unwrap().sample_counter
+    }
+
+    /// Serialize worker history as JSON for the SSE initial payload.
+    pub fn worker_history_json(&self) -> String {
+        let sparklines = self.worker_sparklines();
+        let mailbox = self.worker_mailbox_sparklines();
+        serde_json::json!({
+            "workers": sparklines.iter().enumerate().map(|(i, rates)| {
+                serde_json::json!({
+                    "id": i,
+                    "message_rates": rates,
+                    "mailbox_depths": mailbox.get(i).unwrap_or(&Vec::new()),
+                })
+            }).collect::<Vec<_>>(),
+        })
+        .to_string()
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use swactor::stats::{ActorInfo, WorkerInfo};
+
+    fn make_stats(workers: Vec<(u64, usize, usize)>, actors: Vec<ActorInfo>) -> RuntimeStats {
+        RuntimeStats {
+            num_workers: workers.len(),
+            uptime_ms: 0,
+            actors: actors.iter().map(|a| (a.address, a.worker_id)).collect(),
+            workers: workers
+                .into_iter()
+                .enumerate()
+                .map(|(id, (msgs, depth, n_actors))| WorkerInfo {
+                    id,
+                    num_actors: n_actors,
+                    mailbox_depth: depth,
+                    messages_processed: msgs,
+                    local_sends: 0,
+                    cross_sends: 0,
+                    inbox_sends: 0,
+                    type_mismatches: 0,
+                    panics: 0,
+                    messages_dropped: 0,
+                    restarts: 0,
+                    stops: 0,
+                })
+                .collect(),
+            actor_details: actors,
+            tick_timings: Vec::new(),
+        }
+    }
+
+    fn make_actor(id: u8, worker: usize, depth: usize, msgs: u64) -> ActorInfo {
+        ActorInfo {
+            address: ActorAddress([id; 32]),
+            worker_id: worker,
+            mailbox_depth: depth,
+            last_msg_type: None,
+            messages_processed: msgs,
+            poisoned: false,
+            message_type_counts: Vec::new(),
+        }
+    }
+
+    #[test]
+    fn worker_rates_accumulate_over_samples() {
+        let history = DashboardHistory::new(HistoryConfig::default());
+
+        // First sample: establishes baseline (rate will be the raw value since prev=0)
+        let stats1 = make_stats(vec![(100, 5, 2)], vec![]);
+        history.record(&stats1);
+
+        // Second sample: delta = 150 - 100 = 50
+        let stats2 = make_stats(vec![(150, 3, 2)], vec![]);
+        history.record(&stats2);
+
+        let sparklines = history.worker_sparklines();
+        assert_eq!(sparklines.len(), 1);
+        assert_eq!(sparklines[0].len(), 2);
+        assert_eq!(sparklines[0][0], 100); // first sample: 100 - 0
+        assert_eq!(sparklines[0][1], 50); // second sample: 150 - 100
+    }
+
+    #[test]
+    fn bounded_eviction_drops_oldest() {
+        let config = HistoryConfig {
+            max_worker_samples: 3,
+            ..Default::default()
+        };
+        let history = DashboardHistory::new(config);
+
+        for i in 0..5u64 {
+            let stats = make_stats(vec![(i * 10, 0, 0)], vec![]);
+            history.record(&stats);
+        }
+
+        let sparklines = history.worker_sparklines();
+        assert_eq!(sparklines[0].len(), 3); // capped at 3
+    }
+
+    #[test]
+    fn actor_lru_eviction_keeps_most_recent() {
+        let config = HistoryConfig {
+            max_tracked_actors: 2,
+            ..Default::default()
+        };
+        let history = DashboardHistory::new(config);
+
+        // Sample 1: actors A and B
+        let stats1 = make_stats(
+            vec![(0, 0, 2)],
+            vec![make_actor(1, 0, 0, 0), make_actor(2, 0, 0, 0)],
+        );
+        history.record(&stats1);
+
+        // Sample 2: actors B and C (A not seen)
+        let stats2 = make_stats(
+            vec![(0, 0, 2)],
+            vec![make_actor(2, 0, 0, 0), make_actor(3, 0, 0, 0)],
+        );
+        history.record(&stats2);
+
+        // A should be evicted (LRU), B and C kept
+        assert!(history.actor_sparkline(&ActorAddress([1; 32])).is_none());
+        assert!(history.actor_sparkline(&ActorAddress([2; 32])).is_some());
+        assert!(history.actor_sparkline(&ActorAddress([3; 32])).is_some());
+    }
+
+    #[test]
+    fn actor_rates_track_deltas() {
+        let history = DashboardHistory::new(HistoryConfig::default());
+
+        let stats1 = make_stats(vec![(0, 0, 1)], vec![make_actor(1, 0, 5, 100)]);
+        history.record(&stats1);
+
+        let stats2 = make_stats(vec![(0, 0, 1)], vec![make_actor(1, 0, 3, 175)]);
+        history.record(&stats2);
+
+        let (mailbox, rates) = history.actor_sparkline(&ActorAddress([1; 32])).unwrap();
+        assert_eq!(mailbox, vec![5, 3]);
+        assert_eq!(rates[0], 100); // first: 100 - 0
+        assert_eq!(rates[1], 75); // second: 175 - 100
+    }
+}
--- a/crates/runtime-dashboard/src/layer.rs
+++ b/crates/runtime-dashboard/src/layer.rs
@ -20,6 +20,7 @@ pub struct DashboardEvent {
    pub level: String,
    pub message: String,
    pub worker_id: Option<usize>,
+    pub actor_addr: Option<String>,
    pub fields: serde_json::Map<String, serde_json::Value>,
 }

@ -83,6 +84,28 @@ impl EventStore {
        (batch, new_cursor)
    }

+    /// Read recent events for a specific actor address (hex prefix match).
+    /// Returns up to `limit` most recent matching events.
+    pub fn read_for_actor(&self, actor_hex: &str, limit: usize) -> Vec<DashboardEvent> {
+        let events = self.events.lock().unwrap();
+        let lower = actor_hex.to_lowercase();
+        events
+            .iter()
+            .rev()
+            .filter(|e| {
+                e.actor_addr
+                    .as_ref()
+                    .map(|a| a.to_lowercase().starts_with(&lower) || a.to_lowercase().contains(&lower))
+                    .unwrap_or(false)
+            })
+            .take(limit)
+            .cloned()
+            .collect::<Vec<_>>()
+            .into_iter()
+            .rev()
+            .collect()
+    }
+
    /// Drains the full recording log. Only available when recording is enabled.
    /// This is destructive — events are consumed. Intended for `save_trace()`.
    pub fn all_events(&self) -> Option<Vec<DashboardEvent>> {
@ -183,24 +206,39 @@ where
        let mut visitor = FieldVisitor::new();
        event.record(&mut visitor);

-        // Walk span context to find worker_id
+        // Walk span context to find worker_id and actor_addr
        let mut worker_id = None;
+        let mut actor_addr = None;
        if let Some(scope) = ctx.event_scope(event) {
            for span in scope {
                let exts = span.extensions();
-                if let Some(wid) = exts.get::<WorkerIdField>() {
-                    worker_id = Some(wid.0);
+                if worker_id.is_none() {
+                    if let Some(wid) = exts.get::<WorkerIdField>() {
+                        worker_id = Some(wid.0);
+                    }
+                }
+                if actor_addr.is_none() {
+                    if let Some(aa) = exts.get::<ActorAddrField>() {
+                        actor_addr = Some(aa.0.clone());
+                    }
+                }
+                if worker_id.is_some() && actor_addr.is_some() {
                    break;
                }
            }
        }

-        // Also check if worker_id was a field on the event itself
+        // Also check if worker_id or actor_addr was a field on the event itself
        if worker_id.is_none() {
            if let Some(serde_json::Value::Number(n)) = visitor.fields.get("worker_id") {
                worker_id = n.as_u64().map(|v| v as usize);
            }
        }
+        if actor_addr.is_none() {
+            if let Some(serde_json::Value::String(s)) = visitor.fields.get("actor_addr") {
+                actor_addr = Some(s.clone());
+            }
+        }

        let dashboard_event = DashboardEvent {
            seq: 0, // filled by push()
@ -208,6 +246,7 @@ where
            level: event.metadata().level().to_string(),
            message: visitor.message,
            worker_id,
+            actor_addr,
            fields: visitor.fields,
        };

@ -215,19 +254,25 @@ where
    }

    fn on_new_span(&self, attrs: &span::Attributes<'_>, id: &span::Id, ctx: Context<'_, S>) {
-        // Extract worker_id from span fields and store in extensions
+        // Extract worker_id and actor_addr from span fields and store in extensions
        let mut visitor = FieldVisitor::new();
        attrs.record(&mut visitor);

-        if let Some(serde_json::Value::Number(n)) = visitor.fields.get("worker_id") {
-            if let Some(wid) = n.as_u64() {
-                if let Some(span) = ctx.span(id) {
+        if let Some(span) = ctx.span(id) {
+            if let Some(serde_json::Value::Number(n)) = visitor.fields.get("worker_id") {
+                if let Some(wid) = n.as_u64() {
                    span.extensions_mut().insert(WorkerIdField(wid as usize));
                }
            }
+            if let Some(serde_json::Value::String(s)) = visitor.fields.get("actor_addr") {
+                span.extensions_mut().insert(ActorAddrField(s.clone()));
+            }
        }
    }
 }

 /// Stored in span extensions to propagate worker_id to child events.
 struct WorkerIdField(usize);
+
+/// Stored in span extensions to propagate actor_addr to child events.
+struct ActorAddrField(String);
--- a/crates/runtime-dashboard/src/lib.rs
+++ b/crates/runtime-dashboard/src/lib.rs
@ -1,10 +1,15 @@
 pub mod collector;
+pub mod history;
 pub mod investigate;
 pub mod layer;
 pub mod trace;
+pub mod warnings;
+mod actor_detail_html;
 mod actors_html;
 mod dashboard_html;
 mod server;
+pub mod topology;
+mod topology_html;

 #[cfg(feature = "tui")]
 pub mod tui;
@ -27,6 +32,7 @@ use tracing_subscriber::layer::SubscriberExt;
 use tracing_subscriber::util::SubscriberInitExt;

 use crate::collector::StatsCollector;
+use crate::history::{DashboardHistory, HistoryConfig};
 use crate::layer::{now_ms, DashboardLayer, EventStore};
 use crate::trace::{RuntimeTrace, TimestampedStats};

@ -80,6 +86,7 @@ pub struct DashboardHandle {
    collector: Arc<Mutex<Option<Arc<StatsCollector>>>>,
    shutdown: Arc<AtomicBool>,
    stats_timeline: Arc<ArrayQueue<TimestampedStats>>,
+    history: Arc<DashboardHistory>,
    recording: bool,
    #[cfg(feature = "distribution")]
    distribution: Arc<Mutex<Option<Arc<dyn distribution_collector::DistributionStatsProvider>>>>,
@ -115,6 +122,11 @@ impl DashboardHandle {
        *self.distribution.lock().unwrap() = Some(provider);
    }

+    /// Access the time-series history store (for TUI sparklines, etc.).
+    pub fn history(&self) -> &Arc<DashboardHistory> {
+        &self.history
+    }
+
    /// Signal the dashboard to shut down (SSE clients receive "done").
    pub fn shutdown(&self) {
        self.shutdown.store(true, Ordering::Release);
@ -159,6 +171,7 @@ pub fn start_dashboard(config: DashboardConfig) -> DashboardHandle {
    let collector: Arc<Mutex<Option<Arc<StatsCollector>>>> = Arc::new(Mutex::new(None));
    let shutdown = Arc::new(AtomicBool::new(false));
    let stats_timeline = Arc::new(ArrayQueue::new(config.record_stats_capacity.max(1)));
+    let history = Arc::new(DashboardHistory::new(HistoryConfig::default()));

    #[cfg(feature = "distribution")]
    let distribution: Arc<Mutex<Option<Arc<dyn distribution_collector::DistributionStatsProvider>>>> =
@ -169,6 +182,7 @@ pub fn start_dashboard(config: DashboardConfig) -> DashboardHandle {
        Arc::clone(&runtime),
        Arc::clone(&collector),
        Arc::clone(&shutdown),
+        Arc::clone(&history),
        config.port,
        #[cfg(feature = "distribution")]
        Arc::clone(&distribution),
@ -210,6 +224,7 @@ pub fn start_dashboard(config: DashboardConfig) -> DashboardHandle {
        collector,
        shutdown,
        stats_timeline,
+        history,
        recording: config.record,
        #[cfg(feature = "distribution")]
        distribution,
--- a/crates/runtime-dashboard/src/server.rs
+++ b/crates/runtime-dashboard/src/server.rs
@ -8,11 +8,16 @@ use std::collections::HashMap;

 use swactor::runtime::Runtime;

+use crate::actor_detail_html::ACTOR_DETAIL_HTML;
 use crate::actors_html::ACTORS_HTML;
 use crate::collector::StatsCollector;
 use crate::dashboard_html::DASHBOARD_HTML;
+use crate::history::DashboardHistory;
 use crate::layer::EventStore;
+use crate::topology;
+use crate::topology_html::TOPOLOGY_HTML;
 use crate::trace::RuntimeTrace;
+use crate::warnings::{WarningConfig, WarningDetector};

 #[cfg(feature = "distribution")]
 use crate::distribution_collector::DistributionStatsProvider;
@ -119,6 +124,7 @@ pub(crate) fn spawn_http_server(
    runtime: Arc<Mutex<Option<Arc<Runtime>>>>,
    collector: Arc<Mutex<Option<Arc<StatsCollector>>>>,
    shutdown: Arc<AtomicBool>,
+    history: Arc<DashboardHistory>,
    port: u16,
    #[cfg(feature = "distribution")]
    distribution: Arc<Mutex<Option<Arc<dyn DistributionStatsProvider>>>>,
@ -134,6 +140,7 @@ pub(crate) fn spawn_http_server(
        let runtime = Arc::clone(&runtime);
        let collector = Arc::clone(&collector);
        let shutdown = Arc::clone(&shutdown);
+        let history = Arc::clone(&history);
        let cmd_router = Arc::clone(&cmd_router);
        #[cfg(feature = "distribution")]
        let distribution = Arc::clone(&distribution);
@ -149,6 +156,7 @@ pub(crate) fn spawn_http_server(
                match path {
                    "/" => respond_html(request, DASHBOARD_HTML, "live"),
                    "/actors" => respond_html(request, ACTORS_HTML, "live"),
+                    "/topology" => respond_html(request, TOPOLOGY_HTML, "live"),
                    #[cfg(feature = "distribution")]
                    "/distribution" => respond_html(request, DISTRIBUTION_HTML, "live"),
                    "/events" => {
@ -158,6 +166,7 @@ pub(crate) fn spawn_http_server(
                            Arc::clone(&runtime),
                            Arc::clone(&collector),
                            Arc::clone(&shutdown),
+                            Arc::clone(&history),
                            #[cfg(feature = "distribution")]
                            Arc::clone(&distribution),
                        );
@ -169,6 +178,16 @@ pub(crate) fn spawn_http_server(
                            Arc::clone(&collector),
                        );
                    }
+                    "/api/history" => {
+                        handle_history_api(request, Arc::clone(&history));
+                    }
+                    "/api/topology" => {
+                        handle_topology_api(
+                            request,
+                            Arc::clone(&runtime),
+                            Arc::clone(&collector),
+                        );
+                    }
                    "/api/investigate" => {
                        handle_investigate_api(
                            request,
@ -185,6 +204,13 @@ pub(crate) fn spawn_http_server(
                            Arc::clone(&distribution),
                        );
                    }
+                    "/api/logs" => {
+                        handle_logs_api(request, &url, Arc::clone(&store));
+                    }
+                    _ if path.starts_with("/actor/") => {
+                        let hex = &path[7..]; // strip "/actor/"
+                        respond_actor_detail(request, hex);
+                    }
                    _ => respond_404(request),
                }
            }
@ -192,12 +218,25 @@ pub(crate) fn spawn_http_server(
    }
 }

+fn respond_actor_detail(request: tiny_http::Request, hex_addr: &str) {
+    let html = ACTOR_DETAIL_HTML
+        .replace("__DASHBOARD_MODE__", "live")
+        .replace("__ACTOR_ADDR__", hex_addr);
+    let response = tiny_http::Response::from_string(html).with_header(
+        "Content-Type: text/html; charset=utf-8"
+            .parse::<tiny_http::Header>()
+            .unwrap(),
+    );
+    let _ = request.respond(response);
+}
+
 fn handle_live_sse(
    request: tiny_http::Request,
    store: Arc<EventStore>,
    runtime: Arc<Mutex<Option<Arc<Runtime>>>>,
    collector: Arc<Mutex<Option<Arc<StatsCollector>>>>,
    shutdown: Arc<AtomicBool>,
+    history: Arc<DashboardHistory>,
    #[cfg(feature = "distribution")]
    distribution: Arc<Mutex<Option<Arc<dyn DistributionStatsProvider>>>>,
 ) {
@ -207,6 +246,14 @@ fn handle_live_sse(
    // Spawn producer thread
    thread::spawn(move || {
        let mut cursor: u64 = 0;
+        let mut warning_detector = WarningDetector::new(WarningConfig::default());
+        let mut tick_count: u64 = 0;
+
+        // Send initial history snapshot so sparklines render immediately
+        if history.sample_count() > 0 {
+            let json = history.worker_history_json();
+            let _ = tx.send(format_sse("history", &json));
+        }

        loop {
            // Send stats if runtime is available
@ -217,10 +264,33 @@ fn handle_live_sse(
                    if let Some(col) = collector.lock().unwrap().as_ref() {
                        col.enrich(&mut stats);
                    }
+                    history.record(&stats);
+
+                    // Run warning detection
+                    let warnings = warning_detector.check(&stats);
+                    if !warnings.is_empty() {
+                        if let Ok(wjson) = serde_json::to_string(&warnings) {
+                            if tx.send(format_sse("warnings", &wjson)).is_err() {
+                                return;
+                            }
+                        }
+                    }
+
                    let json = serde_json::to_string(&stats).unwrap();
                    if tx.send(format_sse("stats", &json)).is_err() {
                        return;
                    }
+
+                    // Send topology every 5th tick (~1/sec)
+                    tick_count += 1;
+                    if tick_count % 5 == 0 {
+                        let topo = topology::worker_topology(&stats);
+                        if let Ok(tjson) = serde_json::to_string(&topo) {
+                            if tx.send(format_sse("topology", &tjson)).is_err() {
+                                return;
+                            }
+                        }
+                    }
                }
            }

@ -353,6 +423,67 @@ fn handle_distribution_api(
    let _ = request.respond(response);
 }

+fn handle_topology_api(
+    request: tiny_http::Request,
+    runtime: Arc<Mutex<Option<Arc<Runtime>>>>,
+    collector: Arc<Mutex<Option<Arc<StatsCollector>>>>,
+) {
+    let maybe_rt = runtime.lock().unwrap().clone();
+    let json = match maybe_rt {
+        Some(rt) => {
+            let mut stats = rt.stats();
+            if let Some(col) = collector.lock().unwrap().as_ref() {
+                col.enrich(&mut stats);
+            }
+            let topo = topology::worker_topology(&stats);
+            serde_json::to_string(&topo).unwrap_or_else(|_| "{}".into())
+        }
+        None => "{}".to_string(),
+    };
+    let response = tiny_http::Response::from_string(json).with_header(
+        "Content-Type: application/json"
+            .parse::<tiny_http::Header>()
+            .unwrap(),
+    );
+    let _ = request.respond(response);
+}
+
+fn handle_logs_api(request: tiny_http::Request, url: &str, store: Arc<EventStore>) {
+    let params = parse_query_string(url);
+    let actor = params.get("actor").cloned().unwrap_or_default();
+    let limit: usize = params
+        .get("limit")
+        .and_then(|s| s.parse().ok())
+        .unwrap_or(200);
+    let level = params.get("level").cloned();
+
+    let mut events = store.read_for_actor(&actor, limit);
+
+    // Filter by level if specified
+    if let Some(ref lvl) = level {
+        let lvl_upper = lvl.to_uppercase();
+        events.retain(|e| e.level == lvl_upper);
+    }
+
+    let json = serde_json::to_string(&events).unwrap_or_else(|_| "[]".into());
+    let response = tiny_http::Response::from_string(json).with_header(
+        "Content-Type: application/json"
+            .parse::<tiny_http::Header>()
+            .unwrap(),
+    );
+    let _ = request.respond(response);
+}
+
+fn handle_history_api(request: tiny_http::Request, history: Arc<DashboardHistory>) {
+    let json = history.worker_history_json();
+    let response = tiny_http::Response::from_string(json).with_header(
+        "Content-Type: application/json"
+            .parse::<tiny_http::Header>()
+            .unwrap(),
+    );
+    let _ = request.respond(response);
+}
+
 fn parse_query_string(url: &str) -> HashMap<String, String> {
    let mut params = HashMap::new();
    if let Some(qs) = url.split('?').nth(1) {
--- a/crates/runtime-dashboard/src/topology.rs
+++ b/crates/runtime-dashboard/src/topology.rs
@ -0,0 +1,82 @@
+//! Actor-to-actor (and worker-to-worker) message flow topology.
+//!
+//! Currently derives topology from per-worker cross_sends/local_sends stats.
+//! Future: sample-based per-actor source→destination tracking with core instrumentation.
+
+use swactor::stats::RuntimeStats;
+
+/// An edge in the topology graph.
+#[derive(Debug, Clone, serde::Serialize)]
+pub struct TopologyEdge {
+    pub source: String,
+    pub target: String,
+    pub weight: u64,
+    pub label: String,
+}
+
+/// A node in the topology graph.
+#[derive(Debug, Clone, serde::Serialize)]
+pub struct TopologyNode {
+    pub id: String,
+    pub label: String,
+    pub actor_count: usize,
+    pub group: usize,
+}
+
+/// A snapshot of the current topology.
+#[derive(Debug, Clone, serde::Serialize)]
+pub struct TopologySnapshot {
+    pub nodes: Vec<TopologyNode>,
+    pub edges: Vec<TopologyEdge>,
+}
+
+/// Build a worker-level topology from RuntimeStats.
+///
+/// Workers are nodes, edges represent message flow:
+/// - Self-loops for local_sends
+/// - Cross-edges distributed proportionally (until per-destination tracking exists)
+pub fn worker_topology(stats: &RuntimeStats) -> TopologySnapshot {
+    let mut nodes = Vec::new();
+    let mut edges = Vec::new();
+
+    for w in &stats.workers {
+        nodes.push(TopologyNode {
+            id: format!("w{}", w.id),
+            label: format!("W{}", w.id),
+            actor_count: w.num_actors,
+            group: w.id,
+        });
+
+        // Local sends = self-loop
+        if w.local_sends > 0 {
+            edges.push(TopologyEdge {
+                source: format!("w{}", w.id),
+                target: format!("w{}", w.id),
+                weight: w.local_sends,
+                label: format!("{} local", w.local_sends),
+            });
+        }
+
+        // Cross sends — without per-destination data, distribute evenly to other workers
+        if w.cross_sends > 0 && stats.workers.len() > 1 {
+            let others: Vec<&swactor::stats::WorkerInfo> =
+                stats.workers.iter().filter(|o| o.id != w.id).collect();
+            let per_worker = w.cross_sends / others.len() as u64;
+            let remainder = w.cross_sends % others.len() as u64;
+
+            for (i, other) in others.iter().enumerate() {
+                let count = per_worker + if (i as u64) < remainder { 1 } else { 0 };
+                if count > 0 {
+                    edges.push(TopologyEdge {
+                        source: format!("w{}", w.id),
+                        target: format!("w{}", other.id),
+                        weight: count,
+                        label: format!("{} cross", count),
+                    });
+                }
+            }
+        }
+    }
+
+    TopologySnapshot { nodes, edges }
+}
--- a/crates/runtime-dashboard/src/topology_html.rs
+++ b/crates/runtime-dashboard/src/topology_html.rs
@ -0,0 +1,280 @@
+pub const TOPOLOGY_HTML: &str = r##"<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>Topology — Swactor Dashboard</title>
+<style>
+  * { margin: 0; padding: 0; box-sizing: border-box; }
+  body { font-family: 'Menlo', 'Consolas', 'Monaco', monospace; background: #0f1117; color: #e0e0e0; font-size: 13px; }
+
+  .header {
+    display: flex; align-items: center; justify-content: space-between;
+    padding: 12px 20px; background: #161822; border-bottom: 1px solid #2a2d3e;
+  }
+  .header h1 { font-size: 16px; font-weight: 600; color: #fff; }
+  .status-dot {
+    width: 10px; height: 10px; border-radius: 50%; background: #4caf50;
+    display: inline-block; margin-left: 8px; vertical-align: middle;
+  }
+  .status-dot.disconnected { background: #f44336; }
+
+  .header-left { display: flex; align-items: center; }
+  .nav-links { display: flex; gap: 4px; margin-left: 20px; }
+  .nav-link {
+    color: #888; text-decoration: none; font-size: 12px;
+    padding: 4px 10px; border-radius: 3px;
+  }
+  .nav-link:hover { color: #e0e0e0; }
+  .nav-link.active { color: #fff; background: #2a2d3e; }
+
+  .content { padding: 0; display: flex; flex-direction: column; height: calc(100vh - 49px); }
+  canvas#topoCanvas { flex: 1; width: 100%; cursor: grab; }
+  canvas#topoCanvas:active { cursor: grabbing; }
+
+  .legend {
+    padding: 8px 20px; background: #161822; border-top: 1px solid #2a2d3e;
+    font-size: 11px; color: #888;
+  }
+</style>
+</head>
+<body>
+<div class="header">
+  <div class="header-left">
+    <h1>Swactor Runtime Dashboard <span id="statusDot" class="status-dot"></span></h1>
+    <nav class="nav-links">
+      <a href="/" class="nav-link">Overview</a>
+      <a href="/actors" class="nav-link">Actors</a>
+      <a href="/topology" class="nav-link active">Topology</a>
+      <a href="/distribution" class="nav-link">Distribution</a>
+    </nav>
+  </div>
+</div>
+
+<div class="content">
+  <canvas id="topoCanvas"></canvas>
+  <div class="legend">
+    Node size = actor count. Edge thickness = message volume. Green = local sends. Blue = cross-worker sends.
+  </div>
+</div>
+
+<script>
+(function() {
+  var canvas = document.getElementById('topoCanvas');
+  var ctx = canvas.getContext('2d');
+  var dot = document.getElementById('statusDot');
+
+  var colors = ['#4caf50','#2196f3','#ff9800','#f44336','#9c27b0','#00bcd4','#ffeb3b','#e91e63'];
+  var nodes = [];
+  var edges = [];
+  var positions = {};
+
+  function resize() {
+    var dpr = window.devicePixelRatio || 1;
+    var rect = canvas.getBoundingClientRect();
+    canvas.width = rect.width * dpr;
+    canvas.height = rect.height * dpr;
+    ctx.setTransform(dpr, 0, 0, dpr, 0, 0);
+  }
+  window.addEventListener('resize', resize);
+  resize();
+
+  function initPositions() {
+    var W = canvas.getBoundingClientRect().width;
+    var H = canvas.getBoundingClientRect().height;
+    var cx = W / 2, cy = H / 2;
+    var r = Math.min(W, H) * 0.3;
+
+    nodes.forEach(function(n, i) {
+      if (!positions[n.id]) {
+        var angle = (2 * Math.PI * i) / Math.max(1, nodes.length);
+        positions[n.id] = {
+          x: cx + r * Math.cos(angle),
+          y: cy + r * Math.sin(angle),
+          vx: 0, vy: 0
+        };
+      }
+    });
+  }
+
+  function simulate() {
+    var W = canvas.getBoundingClientRect().width;
+    var H = canvas.getBoundingClientRect().height;
+    var cx = W / 2, cy = H / 2;
+
+    // Repulsion between nodes
+    for (var i = 0; i < nodes.length; i++) {
+      var pi = positions[nodes[i].id];
+      if (!pi) continue;
+      for (var j = i + 1; j < nodes.length; j++) {
+        var pj = positions[nodes[j].id];
+        if (!pj) continue;
+        var dx = pi.x - pj.x;
+        var dy = pi.y - pj.y;
+        var dist = Math.sqrt(dx * dx + dy * dy) || 1;
+        var force = 8000 / (dist * dist);
+        pi.vx += dx / dist * force;
+        pi.vy += dy / dist * force;
+        pj.vx -= dx / dist * force;
+        pj.vy -= dy / dist * force;
+      }
+    }
+
+    // Attraction along edges
+    edges.forEach(function(e) {
+      if (e.source === e.target) return;
+      var ps = positions[e.source];
+      var pt = positions[e.target];
+      if (!ps || !pt) return;
+      var dx = pt.x - ps.x;
+      var dy = pt.y - ps.y;
+      var dist = Math.sqrt(dx * dx + dy * dy) || 1;
+      var force = (dist - 150) * 0.01;
+      ps.vx += dx / dist * force;
+      ps.vy += dy / dist * force;
+      pt.vx -= dx / dist * force;
+      pt.vy -= dy / dist * force;
+    });
+
+    // Gravity toward center
+    for (var i = 0; i < nodes.length; i++) {
+      var p = positions[nodes[i].id];
+      if (!p) continue;
+      p.vx += (cx - p.x) * 0.002;
+      p.vy += (cy - p.y) * 0.002;
+    }
+
+    // Apply velocity with damping
+    for (var i = 0; i < nodes.length; i++) {
+      var p = positions[nodes[i].id];
+      if (!p) continue;
+      p.vx *= 0.85;
+      p.vy *= 0.85;
+      p.x += p.vx;
+      p.y += p.vy;
+      p.x = Math.max(30, Math.min(W - 30, p.x));
+      p.y = Math.max(30, Math.min(H - 30, p.y));
+    }
+  }
+
+  function draw() {
+    var W = canvas.getBoundingClientRect().width;
+    var H = canvas.getBoundingClientRect().height;
+    ctx.clearRect(0, 0, W, H);
+
+    if (nodes.length === 0) {
+      ctx.fillStyle = '#555';
+      ctx.font = '14px monospace';
+      ctx.textAlign = 'center';
+      ctx.fillText('Waiting for topology data...', W / 2, H / 2);
+      return;
+    }
+
+    // Draw edges
+    var maxWeight = Math.max(1, Math.max.apply(null, edges.map(function(e) { return e.weight; })));
+
+    edges.forEach(function(e) {
+      var ps = positions[e.source];
+      var pt = positions[e.target];
+      if (!ps || !pt) return;
+
+      var isSelf = e.source === e.target;
+      var thickness = Math.max(1, (e.weight / maxWeight) * 6);
+      var color = isSelf ? 'rgba(76, 175, 80, 0.5)' : 'rgba(33, 150, 243, 0.5)';
+
+      if (isSelf) {
+        // Self-loop: small arc above the node
+        ctx.beginPath();
+        ctx.arc(ps.x, ps.y - 25, 15, 0.3, Math.PI - 0.3);
+        ctx.strokeStyle = color;
+        ctx.lineWidth = thickness;
+        ctx.stroke();
+      } else {
+        ctx.beginPath();
+        ctx.moveTo(ps.x, ps.y);
+        ctx.lineTo(pt.x, pt.y);
+        ctx.strokeStyle = color;
+        ctx.lineWidth = thickness;
+        ctx.stroke();
+
+        // Arrow
+        var angle = Math.atan2(pt.y - ps.y, pt.x - ps.x);
+        var headLen = 8;
+        var mx = (ps.x + pt.x) / 2;
+        var my = (ps.y + pt.y) / 2;
+        ctx.beginPath();
+        ctx.moveTo(mx, my);
+        ctx.lineTo(mx - headLen * Math.cos(angle - 0.3), my - headLen * Math.sin(angle - 0.3));
+        ctx.moveTo(mx, my);
+        ctx.lineTo(mx - headLen * Math.cos(angle + 0.3), my - headLen * Math.sin(angle + 0.3));
+        ctx.strokeStyle = color;
+        ctx.lineWidth = 1.5;
+        ctx.stroke();
+
+        // Edge label
+        ctx.fillStyle = '#666';
+        ctx.font = '9px monospace';
+        ctx.textAlign = 'center';
+        ctx.fillText(e.label, mx, my - 6);
+      }
+    });
+
+    // Draw nodes
+    nodes.forEach(function(n) {
+      var p = positions[n.id];
+      if (!p) return;
+      var r = Math.max(12, 8 + n.actor_count * 2);
+      var color = colors[n.group % colors.length];
+
+      ctx.beginPath();
+      ctx.arc(p.x, p.y, r, 0, 2 * Math.PI);
+      ctx.fillStyle = color;
+      ctx.globalAlpha = 0.7;
+      ctx.fill();
+      ctx.globalAlpha = 1;
+      ctx.strokeStyle = '#fff';
+      ctx.lineWidth = 1.5;
+      ctx.stroke();
+
+      ctx.fillStyle = '#fff';
+      ctx.font = 'bold 11px monospace';
+      ctx.textAlign = 'center';
+      ctx.textBaseline = 'middle';
+      ctx.fillText(n.label, p.x, p.y);
+
+      ctx.fillStyle = '#888';
+      ctx.font = '9px monospace';
+      ctx.fillText(n.actor_count + ' actors', p.x, p.y + r + 12);
+    });
+  }
+
+  function updateTopology(data) {
+    nodes = data.nodes || [];
+    edges = data.edges || [];
+    initPositions();
+  }
+
+  function tick() {
+    simulate();
+    draw();
+    requestAnimationFrame(tick);
+  }
+  tick();
+
+  var es = new EventSource('/events');
+
+  es.addEventListener('topology', function(e) {
+    try { updateTopology(JSON.parse(e.data)); } catch(err) { console.error(err); }
+  });
+
+  es.addEventListener('done', function() {
+    dot.className = 'status-dot disconnected';
+    es.close();
+  });
+  es.onerror = function() { dot.className = 'status-dot disconnected'; };
+  es.onopen = function() { dot.className = 'status-dot'; };
+})();
+</script>
+</body>
+</html>
+"##;
--- a/crates/runtime-dashboard/src/tui/app.rs
+++ b/crates/runtime-dashboard/src/tui/app.rs
@ -1,9 +1,14 @@
+use std::collections::{HashMap, VecDeque};
+use std::sync::Arc;
 use std::time::Instant;

 use crossterm::event::{KeyCode, KeyEvent, KeyModifiers};
 use swactor::actor::ActorAddress;
 use swactor::stats::{RuntimeStats, TickTiming};

+use crate::layer::{DashboardEvent, EventStore};
+use crate::warnings::{Warning, WarningConfig, WarningDetector};
+
 /// Per-worker data prepared for rendering.
 pub struct WorkerView {
    pub id: usize,
@ -15,6 +20,10 @@ pub struct WorkerView {
    /// Fraction of bar for each phase group: [processing, delivery, spawns, overhead]
    pub phase_fractions: [f64; 4],
    pub panics: u64,
+    /// Recent message rates for sparkline rendering.
+    pub sparkline_rates: Vec<u64>,
+    /// Recent mailbox depths for sparkline rendering.
+    pub sparkline_mailbox: Vec<u64>,
 }

 /// One row in the actor table.
@ -25,6 +34,12 @@ pub struct ActorRow {
    pub last_msg_type: Option<String>,
    pub messages_processed: u64,
    pub poisoned: bool,
+    /// Per-actor mailbox sparkline (from local ring buffer).
+    pub sparkline_mailbox: Vec<u64>,
+    /// Per-actor message rate sparkline.
+    pub sparkline_rates: Vec<u64>,
+    /// Per-message-type counts, sorted descending.
+    pub message_type_counts: Vec<(String, u64)>,
 }

 #[derive(Clone, Copy, PartialEq, Eq)]
@ -62,6 +77,7 @@ impl SortColumn {
 pub enum ViewMode {
    Overview,
    WorkerDetail,
+    ActorDetail,
    #[cfg(feature = "distribution")]
    Distribution,
 }
@ -79,7 +95,13 @@ pub struct App {
    pub total_panics: u64,
    pub num_workers: usize,
    pub view_mode: ViewMode,
+    pub prev_view_mode: ViewMode,
    pub focused_worker: usize,
+    pub focused_actor: Option<ActorAddress>,
+    pub search_active: bool,
+    pub search_query: String,
+    pub search_locked: bool,
+    pub warnings: Vec<Warning>,

    #[cfg(feature = "distribution")]
    pub distribution: Option<distribution::snapshot::DistributionNodeSnapshot>,
@ -90,6 +112,21 @@ pub struct App {
    prev_time: Instant,
    /// Rolling msg rates (smoothed)
    msg_rates: Vec<f64>,
+    /// Per-worker sparkline history (message rate deltas).
+    sparkline_rates: Vec<VecDeque<u64>>,
+    /// Per-worker sparkline history (mailbox depths).
+    sparkline_mailbox: Vec<VecDeque<u64>>,
+    /// Per-actor sparkline history: address → (prev_msgs, rates, mailbox_depths).
+    actor_sparklines: HashMap<ActorAddress, (u64, VecDeque<u64>, VecDeque<u64>)>,
+    warning_detector: WarningDetector,
+    /// Event store for fetching per-actor logs.
+    event_store: Option<Arc<EventStore>>,
+    /// Cached log entries for the focused actor.
+    pub actor_logs: Vec<DashboardEvent>,
+    /// Scroll offset for actor log view.
+    pub log_scroll: usize,
+    /// Active log level filter (all enabled by default).
+    pub log_levels: [bool; 5], // ERROR, WARN, INFO, DEBUG, TRACE
 }

 impl App {
@ -107,7 +144,13 @@ impl App {
            total_panics: 0,
            num_workers: 0,
            view_mode: ViewMode::Overview,
+            prev_view_mode: ViewMode::Overview,
            focused_worker: 0,
+            focused_actor: None,
+            search_active: false,
+            search_query: String::new(),
+            search_locked: false,
+            warnings: Vec::new(),
            #[cfg(feature = "distribution")]
            distribution: None,
            #[cfg(feature = "distribution")]
@ -115,9 +158,22 @@ impl App {
            prev_messages: Vec::new(),
            prev_time: Instant::now(),
            msg_rates: Vec::new(),
+            sparkline_rates: Vec::new(),
+            sparkline_mailbox: Vec::new(),
+            actor_sparklines: HashMap::new(),
+            warning_detector: WarningDetector::new(WarningConfig::default()),
+            event_store: None,
+            actor_logs: Vec::new(),
+            log_scroll: 0,
+            log_levels: [true; 5],
        }
    }

+    /// Set the event store for per-actor log retrieval.
+    pub fn set_event_store(&mut self, store: Arc<EventStore>) {
+        self.event_store = Some(store);
+    }
+
    #[cfg(feature = "distribution")]
    pub fn update_distribution(&mut self, snapshot: distribution::snapshot::DistributionNodeSnapshot) {
        let max = if snapshot.members.is_empty() { 0 } else { snapshot.members.len() - 1 };
@ -125,6 +181,62 @@ impl App {
        self.distribution = Some(snapshot);
    }

+    /// Get visible actor rows (filtered by search query if active).
+    pub fn visible_actor_rows(&self) -> Vec<&ActorRow> {
+        if self.search_query.is_empty() {
+            self.actor_rows.iter().collect()
+        } else {
+            let q = self.search_query.to_lowercase();
+            self.actor_rows.iter().filter(|r| {
+                let addr = format!("{}", r.address).to_lowercase();
+                let msg = r.last_msg_type.as_deref().unwrap_or("").to_lowercase();
+                let worker = format!("w{}", r.worker_id);
+                addr.contains(&q) || msg.contains(&q) || worker.contains(&q)
+            }).collect()
+        }
+    }
+
+    /// Get the focused actor's data (for actor detail view).
+    pub fn focused_actor_row(&self) -> Option<&ActorRow> {
+        self.focused_actor.as_ref().and_then(|addr| {
+            self.actor_rows.iter().find(|r| r.address == *addr)
+        })
+    }
+
+    /// Refresh the actor log cache from the event store.
+    pub fn refresh_actor_logs(&mut self) {
+        if let (Some(addr), Some(store)) = (&self.focused_actor, &self.event_store) {
+            let hex = format!("{}", addr);
+            self.actor_logs = store.read_for_actor(&hex, 200);
+        } else {
+            self.actor_logs.clear();
+        }
+    }
+
+    /// Get visible log entries (filtered by level).
+    pub fn visible_logs(&self) -> Vec<&DashboardEvent> {
+        self.actor_logs
+            .iter()
+            .filter(|e| {
+                match e.level.as_str() {
+                    "ERROR" => self.log_levels[0],
+                    "WARN" => self.log_levels[1],
+                    "INFO" => self.log_levels[2],
+                    "DEBUG" => self.log_levels[3],
+                    "TRACE" => self.log_levels[4],
+                    _ => true,
+                }
+            })
+            .collect()
+    }
+
+    /// Toggle a log level filter (0=ERROR, 1=WARN, 2=INFO, 3=DEBUG, 4=TRACE).
+    pub fn toggle_log_level(&mut self, idx: usize) {
+        if idx < 5 {
+            self.log_levels[idx] = !self.log_levels[idx];
+        }
+    }
+
    /// Actor rows filtered to the focused worker (for worker detail view).
    pub fn focused_actor_rows(&self) -> Vec<&ActorRow> {
        self.actor_rows
@ -144,6 +256,8 @@ impl App {
        if self.prev_messages.len() != stats.workers.len() {
            self.prev_messages = stats.workers.iter().map(|w| w.messages_processed).collect();
            self.msg_rates = vec![0.0; stats.workers.len()];
+            self.sparkline_rates.resize_with(stats.workers.len(), VecDeque::new);
+            self.sparkline_mailbox.resize_with(stats.workers.len(), VecDeque::new);
        }

        // Compute per-worker views
@ -163,8 +277,19 @@ impl App {
                self.msg_rates[i] * 0.6 + rate * 0.4
            };
            self.msg_rates[i] = smoothed;
+
+            // Track sparkline history
+            let delta = w.messages_processed.saturating_sub(self.prev_messages[i]);
            self.prev_messages[i] = w.messages_processed;

+            let spark_rates = &mut self.sparkline_rates[i];
+            if spark_rates.len() >= 60 { spark_rates.pop_front(); }
+            spark_rates.push_back(delta);
+
+            let spark_mbox = &mut self.sparkline_mailbox[i];
+            if spark_mbox.len() >= 60 { spark_mbox.pop_front(); }
+            spark_mbox.push_back(w.mailbox_depth as u64);
+
            // Load % and phase fractions from tick timings
            let timings = stats.tick_timings.get(i).map(|v| v.as_slice()).unwrap_or(&[]);
            let (load_pct, phase_fractions) = compute_load_and_phases(timings);
@ -178,6 +303,8 @@ impl App {
                load_pct,
                phase_fractions,
                panics: w.panics,
+                sparkline_rates: spark_rates.iter().copied().collect(),
+                sparkline_mailbox: spark_mbox.iter().copied().collect(),
            });
        }

@ -187,9 +314,20 @@ impl App {
        self.total_mailbox = stats.workers.iter().map(|w| w.mailbox_depth).sum();
        self.total_panics = stats.workers.iter().map(|w| w.panics).sum();

-        // Build actor table
+        // Build actor table with sparkline history
        self.actor_rows.clear();
        for a in &stats.actor_details {
+            let (prev, rates_buf, mbox_buf) = self.actor_sparklines
+                .entry(a.address)
+                .or_insert_with(|| (0, VecDeque::new(), VecDeque::new()));
+
+            let rate_delta = a.messages_processed.saturating_sub(*prev);
+            *prev = a.messages_processed;
+            if rates_buf.len() >= 60 { rates_buf.pop_front(); }
+            rates_buf.push_back(rate_delta);
+            if mbox_buf.len() >= 60 { mbox_buf.pop_front(); }
+            mbox_buf.push_back(a.mailbox_depth as u64);
+
            self.actor_rows.push(ActorRow {
                address: a.address,
                worker_id: a.worker_id,
@ -197,10 +335,21 @@ impl App {
                last_msg_type: a.last_msg_type.clone(),
                messages_processed: a.messages_processed,
                poisoned: a.poisoned,
+                sparkline_rates: rates_buf.iter().copied().collect(),
+                sparkline_mailbox: mbox_buf.iter().copied().collect(),
+                message_type_counts: a.message_type_counts.clone(),
            });
        }
        self.sort_actors();

+        // Run warning detection
+        self.warnings = self.warning_detector.check(&stats);
+
+        // Refresh actor logs if in detail view
+        if self.view_mode == ViewMode::ActorDetail {
+            self.refresh_actor_logs();
+        }
+
        // Clamp selection
        if !self.actor_rows.is_empty() {
            self.selected = self.selected.min(self.actor_rows.len() - 1);
@ -246,12 +395,45 @@ impl App {
    }

    pub fn handle_key(&mut self, key: KeyEvent) {
+        // Search mode input handling
+        if self.search_active {
+            match key.code {
+                KeyCode::Esc => {
+                    self.search_active = false;
+                    if !self.search_locked {
+                        self.search_query.clear();
+                    }
+                    return;
+                }
+                KeyCode::Enter => {
+                    self.search_active = false;
+                    self.search_locked = !self.search_query.is_empty();
+                    return;
+                }
+                KeyCode::Backspace => {
+                    self.search_query.pop();
+                    return;
+                }
+                KeyCode::Char(c) => {
+                    self.search_query.push(c);
+                    return;
+                }
+                _ => return,
+            }
+        }
+
        // Global keys
        match key.code {
            KeyCode::Char('q') => { self.should_quit = true; return; }
            KeyCode::Char('c') if key.modifiers.contains(KeyModifiers::CONTROL) => {
                self.should_quit = true; return;
            }
+            KeyCode::Char('/') => {
+                self.search_active = true;
+                self.search_locked = false;
+                self.search_query.clear();
+                return;
+            }
            KeyCode::Char('s') => {
                self.sort_column = self.sort_column.next();
                self.sort_actors();
@ -265,6 +447,7 @@ impl App {
            KeyCode::Tab => {
                self.view_mode = match self.view_mode {
                    ViewMode::Overview => ViewMode::WorkerDetail,
+                    ViewMode::ActorDetail => ViewMode::Overview,
                    #[cfg(feature = "distribution")]
                    ViewMode::WorkerDetail => ViewMode::Distribution,
                    #[cfg(not(feature = "distribution"))]
@ -280,13 +463,15 @@ impl App {
        match self.view_mode {
            ViewMode::Overview => self.handle_key_overview(key),
            ViewMode::WorkerDetail => self.handle_key_worker_detail(key),
+            ViewMode::ActorDetail => self.handle_key_actor_detail(key),
            #[cfg(feature = "distribution")]
            ViewMode::Distribution => self.handle_key_distribution(key),
        }
    }

    fn handle_key_overview(&mut self, key: KeyEvent) {
-        let max = if self.actor_rows.is_empty() { 0 } else { self.actor_rows.len() - 1 };
+        let visible = self.visible_actor_rows();
+        let max = if visible.is_empty() { 0 } else { visible.len() - 1 };
        match key.code {
            KeyCode::Up | KeyCode::Char('k') => {
                self.selected = self.selected.saturating_sub(1);
@ -303,11 +488,14 @@ impl App {
            KeyCode::Home => { self.selected = 0; }
            KeyCode::End => { self.selected = max; }
            KeyCode::Enter | KeyCode::Char('l') | KeyCode::Right => {
-                // Enter worker detail for the selected actor's worker
-                if let Some(row) = self.actor_rows.get(self.selected) {
-                    self.focused_worker = row.worker_id;
+                // Enter actor detail for the selected visible actor
+                if let Some(&row) = visible.get(self.selected) {
+                    self.focused_actor = Some(row.address);
+                    self.prev_view_mode = ViewMode::Overview;
+                    self.view_mode = ViewMode::ActorDetail;
+                    self.log_scroll = 0;
+                    self.refresh_actor_logs();
                }
-                self.view_mode = ViewMode::WorkerDetail;
            }
            _ => {}
        }
@ -337,6 +525,38 @@ impl App {
        }
    }

+    fn handle_key_actor_detail(&mut self, key: KeyEvent) {
+        let max_scroll = self.visible_logs().len().saturating_sub(1);
+        match key.code {
+            KeyCode::Esc | KeyCode::Char('h') | KeyCode::Left => {
+                self.view_mode = self.prev_view_mode;
+                self.focused_actor = None;
+                self.actor_logs.clear();
+                self.log_scroll = 0;
+            }
+            KeyCode::Down | KeyCode::Char('j') => {
+                self.log_scroll = (self.log_scroll + 1).min(max_scroll);
+            }
+            KeyCode::Up | KeyCode::Char('k') => {
+                self.log_scroll = self.log_scroll.saturating_sub(1);
+            }
+            KeyCode::PageDown => {
+                self.log_scroll = (self.log_scroll + 20).min(max_scroll);
+            }
+            KeyCode::PageUp => {
+                self.log_scroll = self.log_scroll.saturating_sub(20);
+            }
+            KeyCode::Home => { self.log_scroll = 0; }
+            KeyCode::End => { self.log_scroll = max_scroll; }
+            KeyCode::Char('1') => self.toggle_log_level(0),
+            KeyCode::Char('2') => self.toggle_log_level(1),
+            KeyCode::Char('3') => self.toggle_log_level(2),
+            KeyCode::Char('4') => self.toggle_log_level(3),
+            KeyCode::Char('5') => self.toggle_log_level(4),
+            _ => {}
+        }
+    }
+
    #[cfg(feature = "distribution")]
    fn handle_key_distribution(&mut self, key: KeyEvent) {
        let max = self
--- a/crates/runtime-dashboard/src/tui/mod.rs
+++ b/crates/runtime-dashboard/src/tui/mod.rs
@ -18,6 +18,7 @@ use swactor::runtime::Runtime;
 use crate::collector::StatsCollector;
 #[cfg(feature = "distribution")]
 use crate::distribution_collector::DistributionStatsProvider;
+use crate::layer::EventStore;
 use self::app::App;
 use self::event::{AppEvent, EventLoop};
 use self::types::RuntimeEndpoint;
@ -47,6 +48,7 @@ pub fn start_tui(
    #[cfg(feature = "distribution")]
    distribution: Option<Arc<dyn DistributionStatsProvider>>,
    config: TuiConfig,
+    event_store: Option<Arc<EventStore>>,
 ) -> io::Result<()> {
    // Set up terminal
    crossterm::terminal::enable_raw_mode()?;
@ -72,6 +74,7 @@ pub fn start_tui(
        #[cfg(feature = "distribution")]
        distribution,
        config,
+        event_store,
    );

    // Restore terminal
@ -92,8 +95,12 @@ fn run_loop(
    #[cfg(feature = "distribution")]
    distribution: Option<Arc<dyn DistributionStatsProvider>>,
    config: TuiConfig,
+    event_store: Option<Arc<EventStore>>,
 ) -> io::Result<()> {
    let mut app = App::new();
+    if let Some(store) = event_store {
+        app.set_event_store(store);
+    }
    let mut table_state = TableState::default();
    let events = EventLoop::new(config.poll_interval_ms);

--- a/crates/runtime-dashboard/src/tui/ui.rs
+++ b/crates/runtime-dashboard/src/tui/ui.rs
@ -2,7 +2,7 @@ use ratatui::Frame;
 use ratatui::layout::{Constraint, Layout, Rect};
 use ratatui::style::{Color, Modifier, Style};
 use ratatui::text::{Line, Span};
-use ratatui::widgets::{Block, Borders, Cell, Paragraph, Row, Table, TableState};
+use ratatui::widgets::{Block, Borders, Cell, Paragraph, Row, Sparkline, Table, TableState};

 use super::app::{App, SortColumn, ViewMode};

@ -11,6 +11,7 @@ pub fn draw(f: &mut Frame, app: &App, table_state: &mut TableState) {
    match app.view_mode {
        ViewMode::Overview => draw_overview(f, app, table_state),
        ViewMode::WorkerDetail => draw_worker_detail(f, app, table_state),
+        ViewMode::ActorDetail => draw_actor_detail(f, app),
        #[cfg(feature = "distribution")]
        ViewMode::Distribution => draw_distribution(f, app, table_state),
    }
@ -20,17 +21,23 @@ pub fn draw(f: &mut Frame, app: &App, table_state: &mut TableState) {

 fn draw_overview(f: &mut Frame, app: &App, table_state: &mut TableState) {
    let num_workers = app.workers.len().max(1);
+    let has_sparkline_data = app.workers.iter().any(|w| w.sparkline_rates.len() > 1);
+    let sparkline_height = if has_sparkline_data { 4u16 } else { 0u16 };

    let chunks = Layout::vertical([
        Constraint::Length(num_workers as u16),
+        Constraint::Length(sparkline_height),
        Constraint::Length(1),
        Constraint::Fill(1),
    ])
    .split(f.area());

    draw_worker_bars(f, app, chunks[0]);
-    draw_summary(f, app, chunks[1]);
-    draw_actor_table(f, app, table_state, chunks[2]);
+    if has_sparkline_data {
+        draw_worker_sparklines(f, app, chunks[1]);
+    }
+    draw_summary(f, app, chunks[2]);
+    draw_actor_table(f, app, table_state, chunks[3]);
 }

 /// Render htop-style worker bars.
@ -45,6 +52,36 @@ fn draw_worker_bars(f: &mut Frame, app: &App, area: Rect) {
    f.render_widget(paragraph, area);
 }

+/// Render per-worker sparklines showing message rate trends.
+fn draw_worker_sparklines(f: &mut Frame, app: &App, area: Rect) {
+    if app.workers.is_empty() {
+        return;
+    }
+    // Split area horizontally: one sparkline per worker
+    let constraints: Vec<Constraint> = app
+        .workers
+        .iter()
+        .map(|_| Constraint::Ratio(1, app.workers.len() as u32))
+        .collect();
+    let cols = Layout::horizontal(constraints).split(area);
+
+    for (i, w) in app.workers.iter().enumerate() {
+        if let Some(&col_area) = cols.get(i) {
+            let block = Block::default()
+                .borders(Borders::NONE)
+                .title(Span::styled(
+                    format!(" W{} ", w.id),
+                    Style::default().fg(Color::DarkGray),
+                ));
+            let sparkline = Sparkline::default()
+                .block(block)
+                .data(&w.sparkline_rates)
+                .style(Style::default().fg(Color::Green));
+            f.render_widget(sparkline, col_area);
+        }
+    }
+}
+
 fn build_worker_line(w: &super::app::WorkerView, total_width: usize) -> Line<'static> {
    let id_str = format!("{:>3}", w.id);
    let suffix = format!(
@ -129,7 +166,7 @@ fn draw_summary(f: &mut Frame, app: &App, area: Rect) {
        Style::default().fg(Color::DarkGray)
    };

-    let line = Line::from(vec![
+    let mut spans = vec![
        Span::styled(" Workers: ", Style::default().fg(Color::DarkGray)),
        Span::styled(
            format!("{}", app.num_workers),
@ -152,13 +189,47 @@ fn draw_summary(f: &mut Frame, app: &App, area: Rect) {
        ),
        Span::styled("  Panics: ", Style::default().fg(Color::DarkGray)),
        Span::styled(format!("{}", app.total_panics), panics_style),
-    ]);
+    ];

+    if !app.warnings.is_empty() {
+        spans.push(Span::styled("  Warnings: ", Style::default().fg(Color::DarkGray)));
+        spans.push(Span::styled(
+            format!("{}", app.warnings.len()),
+            Style::default().fg(Color::Yellow).add_modifier(Modifier::BOLD),
+        ));
+    }
+
+    let line = Line::from(spans);
    f.render_widget(Paragraph::new(line), area);
 }

-/// Render the actor table with scrolling and selection.
+/// Render the actor table with scrolling, selection, and search.
 fn draw_actor_table(f: &mut Frame, app: &App, table_state: &mut TableState, area: Rect) {
+    // Split area: optional search bar + table
+    let has_search = app.search_active || !app.search_query.is_empty();
+    let search_height = if has_search { 1u16 } else { 0 };
+    let chunks = Layout::vertical([
+        Constraint::Length(search_height),
+        Constraint::Fill(1),
+    ])
+    .split(area);
+
+    // Draw search bar
+    if has_search {
+        let search_style = if app.search_active {
+            Style::default().fg(Color::Yellow)
+        } else {
+            Style::default().fg(Color::DarkGray)
+        };
+        let cursor = if app.search_active { "\u{2588}" } else { "" };
+        let prefix = if app.search_locked { " [locked] /" } else { " /" };
+        let line = Line::from(vec![
+            Span::styled(prefix, Style::default().fg(Color::DarkGray)),
+            Span::styled(format!("{}{}", app.search_query, cursor), search_style),
+        ]);
+        f.render_widget(Paragraph::new(line), chunks[0]);
+    }
+
    let sort_arrow = if app.sort_desc { " \u{25bc}" } else { " \u{25b2}" };

    let columns = [
@ -181,8 +252,8 @@ fn draw_actor_table(f: &mut Frame, app: &App, table_state: &mut TableState, area
    });
    let header = Row::new(header_cells).height(1);

-    let rows: Vec<Row> = app
-        .actor_rows
+    let visible = app.visible_actor_rows();
+    let rows: Vec<Row> = visible
        .iter()
        .map(|a| actor_row_cells(a))
        .collect();
@ -190,7 +261,22 @@ fn draw_actor_table(f: &mut Frame, app: &App, table_state: &mut TableState, area
    table_state.select(Some(app.selected));

    let help_text =
-        " q: quit  \u{2191}\u{2193}: scroll  s: sort column  r: reverse  Tab: worker view  Enter: drill in";
+        " q: quit  /: search  \u{2191}\u{2193}: scroll  s: sort  r: reverse  Tab: worker view  Enter: detail";
+
+    let title = if !app.search_query.is_empty() {
+        format!(
+            " Actors ({} of {} matching \"{}\") ",
+            visible.len(),
+            app.actor_rows.len(),
+            app.search_query,
+        )
+    } else {
+        format!(
+            " Actors (sorted by {}{}) ",
+            app.sort_column.label(),
+            sort_arrow
+        )
+    };

    let table = Table::new(
        rows,
@ -206,11 +292,7 @@ fn draw_actor_table(f: &mut Frame, app: &App, table_state: &mut TableState, area
    .block(
        Block::default()
            .borders(Borders::ALL)
-            .title(format!(
-                " Actors (sorted by {}{}) ",
-                app.sort_column.label(),
-                sort_arrow
-            ))
+            .title(title)
            .title_bottom(Line::from(help_text).centered()),
    )
    .row_highlight_style(
@ -443,6 +525,231 @@ fn draw_focused_actor_table(f: &mut Frame, app: &App, table_state: &mut TableSta
    f.render_stateful_widget(table, area, table_state);
 }

+// ─── Actor Detail View ──────────────────────────────────────────────────────
+
+fn draw_actor_detail(f: &mut Frame, app: &App) {
+    let actor = match app.focused_actor_row() {
+        Some(a) => a,
+        None => {
+            let msg = Paragraph::new(" No actor selected. Press Esc to go back.")
+                .style(Style::default().fg(Color::DarkGray));
+            f.render_widget(msg, f.area());
+            return;
+        }
+    };
+
+    let has_types = !actor.message_type_counts.is_empty();
+    let type_height = if has_types {
+        (actor.message_type_counts.len() as u16 + 2).min(10)
+    } else {
+        0
+    };
+
+    let chunks = Layout::vertical([
+        Constraint::Length(5),           // Info card
+        Constraint::Length(5),           // Rate sparkline
+        Constraint::Length(5),           // Mailbox sparkline
+        Constraint::Length(type_height), // Type breakdown
+        Constraint::Length(1),           // Help bar
+        Constraint::Fill(1),            // Logs panel
+    ])
+    .split(f.area());
+
+    // Info card
+    let addr_str = format!("{}", actor.address);
+    let status = if actor.poisoned { "POISONED" } else { "Healthy" };
+    let status_color = if actor.poisoned { Color::Red } else { Color::Green };
+    let msg_type = actor.last_msg_type.as_deref()
+        .map(|s| short_type_name(Some(s)))
+        .unwrap_or_else(|| "\u{2014}".to_string());
+
+    let info_lines = vec![
+        Line::from(vec![
+            Span::styled("  Address: ", Style::default().fg(Color::DarkGray)),
+            Span::styled(addr_str, Style::default().fg(Color::White).add_modifier(Modifier::BOLD)),
+        ]),
+        Line::from(vec![
+            Span::styled("  Worker: ", Style::default().fg(Color::DarkGray)),
+            Span::styled(format!("W{}", actor.worker_id), Style::default().fg(Color::Cyan)),
+            Span::styled("   Status: ", Style::default().fg(Color::DarkGray)),
+            Span::styled(status, Style::default().fg(status_color).add_modifier(Modifier::BOLD)),
+        ]),
+        Line::from(vec![
+            Span::styled("  Messages: ", Style::default().fg(Color::DarkGray)),
+            Span::styled(
+                format_num(actor.messages_processed),
+                Style::default().fg(Color::Cyan).add_modifier(Modifier::BOLD),
+            ),
+            Span::styled("   Mailbox: ", Style::default().fg(Color::DarkGray)),
+            Span::styled(
+                format!("{}", actor.mailbox_depth),
+                Style::default().fg(Color::Blue).add_modifier(Modifier::BOLD),
+            ),
+            Span::styled("   Last Msg: ", Style::default().fg(Color::DarkGray)),
+            Span::styled(msg_type, Style::default().fg(Color::Green)),
+        ]),
+    ];
+
+    let info_block = Block::default()
+        .borders(Borders::ALL)
+        .title(" Actor Detail ");
+    let info = Paragraph::new(info_lines).block(info_block);
+    f.render_widget(info, chunks[0]);
+
+    // Rate sparkline
+    let rate_block = Block::default()
+        .borders(Borders::ALL)
+        .title(Span::styled(" Msg Rate ", Style::default().fg(Color::Green)));
+    let rate_sparkline = Sparkline::default()
+        .block(rate_block)
+        .data(&actor.sparkline_rates)
+        .style(Style::default().fg(Color::Green));
+    f.render_widget(rate_sparkline, chunks[1]);
+
+    // Mailbox sparkline
+    let mbox_block = Block::default()
+        .borders(Borders::ALL)
+        .title(Span::styled(" Mailbox Depth ", Style::default().fg(Color::Blue)));
+    let mbox_sparkline = Sparkline::default()
+        .block(mbox_block)
+        .data(&actor.sparkline_mailbox)
+        .style(Style::default().fg(Color::Blue));
+    f.render_widget(mbox_sparkline, chunks[2]);
+
+    // Message type breakdown
+    if has_types {
+        draw_type_breakdown(f, &actor.message_type_counts, chunks[3]);
+    }
+
+    // Help bar
+    let level_names = ["ERR", "WARN", "INFO", "DBG", "TRC"];
+    let level_colors = [Color::Red, Color::Yellow, Color::Blue, Color::DarkGray, Color::DarkGray];
+    let mut help_spans: Vec<Span> = vec![
+        Span::styled(
+            " Esc: back  \u{2191}\u{2193}: scroll logs  ",
+            Style::default().fg(Color::DarkGray),
+        ),
+    ];
+    for (i, &name) in level_names.iter().enumerate() {
+        let active = app.log_levels[i];
+        let color = if active { level_colors[i] } else { Color::DarkGray };
+        let style = if active {
+            Style::default().fg(color).add_modifier(Modifier::BOLD)
+        } else {
+            Style::default().fg(color)
+        };
+        help_spans.push(Span::styled(format!("{}:{} ", i + 1, name), style));
+    }
+    f.render_widget(Paragraph::new(Line::from(help_spans)), chunks[4]);
+
+    // Logs panel
+    draw_actor_logs(f, app, chunks[5]);
+}
+
+fn draw_type_breakdown(f: &mut Frame, types: &[(String, u64)], area: Rect) {
+    let total: u64 = types.iter().map(|(_, c)| *c).sum();
+    let max_count = types.first().map(|(_, c)| *c).unwrap_or(1).max(1);
+    let inner_height = area.height.saturating_sub(2) as usize;
+
+    let bar_colors = [Color::Green, Color::Blue, Color::Yellow, Color::Magenta, Color::Cyan, Color::Red];
+
+    let lines: Vec<Line> = types
+        .iter()
+        .take(inner_height)
+        .enumerate()
+        .map(|(i, (name, count))| {
+            let short = name.rsplit("::").next().unwrap_or(name);
+            let pct = if total > 0 { *count as f64 / total as f64 * 100.0 } else { 0.0 };
+            let bar_width = 20usize;
+            let filled = ((*count as f64 / max_count as f64) * bar_width as f64).round() as usize;
+            let color = bar_colors[i % bar_colors.len()];
+
+            Line::from(vec![
+                Span::styled(
+                    format!(" {:>16} ", short),
+                    Style::default().fg(Color::White),
+                ),
+                Span::styled(
+                    "\u{2588}".repeat(filled),
+                    Style::default().fg(color),
+                ),
+                Span::styled(
+                    " ".repeat(bar_width.saturating_sub(filled)),
+                    Style::default(),
+                ),
+                Span::styled(
+                    format!(" {:>8} ", count),
+                    Style::default().fg(Color::Cyan).add_modifier(Modifier::BOLD),
+                ),
+                Span::styled(
+                    format!("{:>5.1}%", pct),
+                    Style::default().fg(Color::DarkGray),
+                ),
+            ])
+        })
+        .collect();
+
+    let block = Block::default()
+        .borders(Borders::ALL)
+        .title(format!(" Message Types ({}) ", types.len()));
+    let paragraph = Paragraph::new(lines).block(block);
+    f.render_widget(paragraph, area);
+}
+
+fn draw_actor_logs(f: &mut Frame, app: &App, area: Rect) {
+    let visible = app.visible_logs();
+    let log_count = visible.len();
+    let inner_height = area.height.saturating_sub(2) as usize; // borders
+
+    // Compute scroll window
+    let scroll = app.log_scroll.min(log_count.saturating_sub(inner_height));
+
+    let lines: Vec<Line> = visible
+        .iter()
+        .skip(scroll)
+        .take(inner_height)
+        .map(|e| {
+            let level_color = match e.level.as_str() {
+                "ERROR" => Color::Red,
+                "WARN" => Color::Yellow,
+                "INFO" => Color::Blue,
+                "DEBUG" => Color::DarkGray,
+                "TRACE" => Color::DarkGray,
+                _ => Color::White,
+            };
+            let ts = {
+                let secs = e.timestamp_ms / 1000;
+                let ms = e.timestamp_ms % 1000;
+                let h = (secs / 3600) % 24;
+                let m = (secs / 60) % 60;
+                let s = secs % 60;
+                format!("{:02}:{:02}:{:02}.{:03}", h, m, s, ms)
+            };
+            Line::from(vec![
+                Span::styled(
+                    format!(" {} ", ts),
+                    Style::default().fg(Color::DarkGray),
+                ),
+                Span::styled(
+                    format!("{:<5} ", e.level),
+                    Style::default().fg(level_color).add_modifier(Modifier::BOLD),
+                ),
+                Span::styled(
+                    e.message.clone(),
+                    Style::default().fg(Color::White),
+                ),
+            ])
+        })
+        .collect();
+
+    let title = format!(" Logs ({}) ", log_count);
+    let block = Block::default()
+        .borders(Borders::ALL)
+        .title(title);
+    let paragraph = Paragraph::new(lines).block(block);
+    f.render_widget(paragraph, area);
+}
+
 // ─── Distribution View ──────────────────────────────────────────────────────

 #[cfg(feature = "distribution")]
--- a/crates/runtime-dashboard/src/warnings.rs
+++ b/crates/runtime-dashboard/src/warnings.rs
@ -0,0 +1,392 @@
+//! Automated anomaly detection for the runtime dashboard.
+//!
+//! Runs on each stats sample, comparing consecutive snapshots to detect
+//! growing mailboxes, stalled actors, worker imbalance, and other conditions.
+
+use std::collections::HashMap;
+
+use swactor::actor::ActorAddress;
+use swactor::stats::RuntimeStats;
+
+/// Types of warnings the detector can produce.
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, serde::Serialize)]
+#[serde(rename_all = "snake_case")]
+pub enum WarningType {
+    GrowingMailbox,
+    StalledActor,
+    PoisonedActor,
+    WorkerImbalance,
+    EmptyWorker,
+    MailboxOverflow,
+}
+
+/// Severity levels for warnings.
+#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, serde::Serialize)]
+#[serde(rename_all = "lowercase")]
+pub enum Severity {
+    Low,
+    Medium,
+    High,
+    Critical,
+}
+
+/// An active warning.
+#[derive(Debug, Clone, serde::Serialize)]
+pub struct Warning {
+    pub warning_type: WarningType,
+    pub severity: Severity,
+    pub entity: String,
+    pub description: String,
+}
+
+/// Configuration for warning thresholds.
+#[derive(Debug, Clone)]
+pub struct WarningConfig {
+    /// Consecutive samples with increasing mailbox depth before warning.
+    pub growing_mailbox_threshold: usize,
+    /// Consecutive ticks with no message processing while mailbox > 0.
+    pub stalled_actor_threshold: usize,
+    /// A worker is "imbalanced" if it has > this ratio times the average load.
+    pub worker_imbalance_ratio: f64,
+}
+
+impl Default for WarningConfig {
+    fn default() -> Self {
+        Self {
+            growing_mailbox_threshold: 5,
+            stalled_actor_threshold: 10,
+            worker_imbalance_ratio: 2.0,
+        }
+    }
+}
+
+/// Per-actor tracking state.
+struct ActorState {
+    prev_mailbox: usize,
+    prev_messages: u64,
+    growing_streak: usize,
+    stalled_streak: usize,
+}
+
+/// Warning detection engine. Call `check()` on each stats sample.
+pub struct WarningDetector {
+    config: WarningConfig,
+    actors: HashMap<ActorAddress, ActorState>,
+}
+
+impl WarningDetector {
+    pub fn new(config: WarningConfig) -> Self {
+        Self {
+            config,
+            actors: HashMap::new(),
+        }
+    }
+
+    /// Analyze a stats snapshot and return all active warnings.
+    pub fn check(&mut self, stats: &RuntimeStats) -> Vec<Warning> {
+        let mut warnings = Vec::new();
+
+        // Track which actors are still alive
+        let mut live_addrs: std::collections::HashSet<ActorAddress> =
+            std::collections::HashSet::new();
+
+        for actor in &stats.actor_details {
+            live_addrs.insert(actor.address);
+
+            // Poisoned actor — immediate critical warning
+            if actor.poisoned {
+                warnings.push(Warning {
+                    warning_type: WarningType::PoisonedActor,
+                    severity: Severity::Critical,
+                    entity: format!("{}", actor.address),
+                    description: "Actor is poisoned (panicked)".to_string(),
+                });
+            }
+
+            let state = self.actors.entry(actor.address).or_insert(ActorState {
+                prev_mailbox: actor.mailbox_depth,
+                prev_messages: actor.messages_processed,
+                growing_streak: 0,
+                stalled_streak: 0,
+            });
+
+            // Growing mailbox detection
+            if actor.mailbox_depth > state.prev_mailbox && actor.mailbox_depth > 0 {
+                state.growing_streak += 1;
+            } else {
+                state.growing_streak = 0;
+            }
+
+            if state.growing_streak >= self.config.growing_mailbox_threshold {
+                warnings.push(Warning {
+                    warning_type: WarningType::GrowingMailbox,
+                    severity: Severity::Medium,
+                    entity: format!("{}", actor.address),
+                    description: format!(
+                        "Mailbox growing for {} consecutive samples (depth: {})",
+                        state.growing_streak, actor.mailbox_depth,
+                    ),
+                });
+            }
+
+            // Stalled actor detection
+            if actor.messages_processed == state.prev_messages && actor.mailbox_depth > 0 {
+                state.stalled_streak += 1;
+            } else {
+                state.stalled_streak = 0;
+            }
+
+            if state.stalled_streak >= self.config.stalled_actor_threshold {
+                warnings.push(Warning {
+                    warning_type: WarningType::StalledActor,
+                    severity: Severity::High,
+                    entity: format!("{}", actor.address),
+                    description: format!(
+                        "No messages processed for {} ticks with {} pending",
+                        state.stalled_streak, actor.mailbox_depth,
+                    ),
+                });
+            }
+
+            state.prev_mailbox = actor.mailbox_depth;
+            state.prev_messages = actor.messages_processed;
+        }
+
+        // Clean up dead actors
+        self.actors.retain(|addr, _| live_addrs.contains(addr));
+
+        // Mailbox overflow detection
+        for w in &stats.workers {
+            if w.messages_dropped > 0 {
+                warnings.push(Warning {
+                    warning_type: WarningType::MailboxOverflow,
+                    severity: Severity::Medium,
+                    entity: format!("Worker {}", w.id),
+                    description: format!("{} messages dropped", w.messages_dropped),
+                });
+            }
+        }
+
+        // Worker imbalance and empty worker detection
+        if stats.workers.len() > 1 {
+            let total_actors: usize = stats.workers.iter().map(|w| w.num_actors).sum();
+            let avg = total_actors as f64 / stats.workers.len() as f64;
+
+            for w in &stats.workers {
+                if avg > 0.0 && w.num_actors as f64 > avg * self.config.worker_imbalance_ratio {
+                    warnings.push(Warning {
+                        warning_type: WarningType::WorkerImbalance,
+                        severity: Severity::Low,
+                        entity: format!("Worker {}", w.id),
+                        description: format!(
+                            "{} actors vs {:.0} average ({:.1}x)",
+                            w.num_actors, avg, w.num_actors as f64 / avg,
+                        ),
+                    });
+                }
+
+                if w.num_actors == 0 && total_actors > 0 {
+                    warnings.push(Warning {
+                        warning_type: WarningType::EmptyWorker,
+                        severity: Severity::Low,
+                        entity: format!("Worker {}", w.id),
+                        description: "Worker has no actors while others do".to_string(),
+                    });
+                }
+            }
+        }
+
+        // Sort by severity (critical first)
+        warnings.sort_by(|a, b| b.severity.cmp(&a.severity));
+        warnings
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use swactor::stats::{ActorInfo, WorkerInfo};
+
+    fn make_worker(id: usize, actors: usize, dropped: u64) -> WorkerInfo {
+        WorkerInfo {
+            id,
+            num_actors: actors,
+            mailbox_depth: 0,
+            messages_processed: 0,
+            local_sends: 0,
+            cross_sends: 0,
+            inbox_sends: 0,
+            type_mismatches: 0,
+            panics: 0,
+            messages_dropped: dropped,
+            restarts: 0,
+            stops: 0,
+        }
+    }
+
+    fn make_actor(id: u8, depth: usize, msgs: u64, poisoned: bool) -> ActorInfo {
+        ActorInfo {
+            address: ActorAddress([id; 32]),
+            worker_id: 0,
+            mailbox_depth: depth,
+            last_msg_type: None,
+            messages_processed: msgs,
+            poisoned,
+            message_type_counts: Vec::new(),
+        }
+    }
+
+    fn make_stats(workers: Vec<WorkerInfo>, actors: Vec<ActorInfo>) -> RuntimeStats {
+        RuntimeStats {
+            num_workers: workers.len(),
+            uptime_ms: 0,
+            actors: actors.iter().map(|a| (a.address, a.worker_id)).collect(),
+            workers,
+            actor_details: actors,
+            tick_timings: Vec::new(),
+        }
+    }
+
+    #[test]
+    fn poisoned_actor_triggers_critical_warning() {
+        let mut detector = WarningDetector::new(WarningConfig::default());
+        let stats = make_stats(
+            vec![make_worker(0, 1, 0)],
+            vec![make_actor(1, 0, 10, true)],
+        );
+        let warnings = detector.check(&stats);
+        assert!(warnings.iter().any(|w| w.warning_type == WarningType::PoisonedActor));
+        assert!(warnings.iter().any(|w| w.severity == Severity::Critical));
+    }
+
+    #[test]
+    fn growing_mailbox_triggers_after_threshold() {
+        let config = WarningConfig {
+            growing_mailbox_threshold: 3,
+            ..Default::default()
+        };
+        let mut detector = WarningDetector::new(config);
+
+        // 4 samples with increasing mailbox: should trigger at sample 4
+        for depth in 1..=4 {
+            let stats = make_stats(
+                vec![make_worker(0, 1, 0)],
+                vec![make_actor(1, depth, 0, false)],
+            );
+            let warnings = detector.check(&stats);
+            if depth < 4 {
+                assert!(!warnings.iter().any(|w| w.warning_type == WarningType::GrowingMailbox),
+                    "should not trigger at depth {}", depth);
+            } else {
+                assert!(warnings.iter().any(|w| w.warning_type == WarningType::GrowingMailbox),
+                    "should trigger at depth {}", depth);
+            }
+        }
+    }
+
+    #[test]
+    fn growing_mailbox_resets_on_decrease() {
+        let config = WarningConfig {
+            growing_mailbox_threshold: 3,
+            ..Default::default()
+        };
+        let mut detector = WarningDetector::new(config);
+
+        // Grow for 2 samples, then decrease, then grow again
+        for depth in [1, 2, 1, 2, 3, 4] {
+            let stats = make_stats(
+                vec![make_worker(0, 1, 0)],
+                vec![make_actor(1, depth, 0, false)],
+            );
+            detector.check(&stats);
+        }
+        // After 1,2 → streak=2; then 1 → streak=0; then 2,3,4 → streak=3 → triggers
+        let stats = make_stats(
+            vec![make_worker(0, 1, 0)],
+            vec![make_actor(1, 5, 0, false)],
+        );
+        let warnings = detector.check(&stats);
+        assert!(warnings.iter().any(|w| w.warning_type == WarningType::GrowingMailbox));
+    }
+
+    #[test]
+    fn stalled_actor_triggers_when_not_processing() {
+        let config = WarningConfig {
+            stalled_actor_threshold: 3,
+            ..Default::default()
+        };
+        let mut detector = WarningDetector::new(config);
+
+        // Same messages_processed, nonzero mailbox for 4 ticks
+        for _ in 0..4 {
+            let stats = make_stats(
+                vec![make_worker(0, 1, 0)],
+                vec![make_actor(1, 5, 100, false)],
+            );
+            let warnings = detector.check(&stats);
+            // Last one should trigger
+            if warnings.iter().any(|w| w.warning_type == WarningType::StalledActor) {
+                return; // test passed
+            }
+        }
+        panic!("expected StalledActor warning");
+    }
+
+    #[test]
+    fn worker_imbalance_detected() {
+        let mut detector = WarningDetector::new(WarningConfig::default());
+        // Worker 0: 10 actors, Worker 1: 1 actor. Avg=5.5, ratio=10/5.5=1.8
+        // With ratio threshold 2.0, this should NOT trigger
+        let stats = make_stats(
+            vec![make_worker(0, 10, 0), make_worker(1, 1, 0)],
+            vec![],
+        );
+        let warnings = detector.check(&stats);
+        assert!(!warnings.iter().any(|w| w.warning_type == WarningType::WorkerImbalance));
+
+        // Worker 0: 20 actors, Worker 1: 1 actor. Avg=10.5, ratio=20/10.5=1.9 — still no
+        // Worker 0: 30 actors, Worker 1: 1 actor. Avg=15.5, ratio=30/15.5=1.9 — still no
+        // Worker 0: 100 actors, Worker 1: 1 actor. Avg=50.5, ratio=100/50.5=1.98 — almost
+        // Worker 0: 100 actors, Worker 1: 0 actor. Avg=50, ratio=100/50=2.0 — at threshold
+
+        let stats2 = make_stats(
+            vec![make_worker(0, 100, 0), make_worker(1, 1, 0)],
+            vec![],
+        );
+        let warnings2 = detector.check(&stats2);
+        // 100 / 50.5 = 1.98 — not > 2.0
+        assert!(!warnings2.iter().any(|w| w.warning_type == WarningType::WorkerImbalance));
+
+        // Now 200 vs 1: 200/100.5 = ~1.99 — still not. Let's do 300 vs 1: 300/150.5 = ~2.0
+        // Actually need > 2x. Let's do 50 vs 1: avg=25.5, ratio=50/25.5=1.96. Nope.
+        // 10 vs 1 vs 1: avg=4, ratio=10/4=2.5 — triggers!
+        let stats3 = make_stats(
+            vec![make_worker(0, 10, 0), make_worker(1, 1, 0), make_worker(2, 1, 0)],
+            vec![],
+        );
+        let warnings3 = detector.check(&stats3);
+        assert!(warnings3.iter().any(|w| w.warning_type == WarningType::WorkerImbalance));
+    }
+
+    #[test]
+    fn empty_worker_detected() {
+        let mut detector = WarningDetector::new(WarningConfig::default());
+        let stats = make_stats(
+            vec![make_worker(0, 5, 0), make_worker(1, 0, 0)],
+            vec![],
+        );
+        let warnings = detector.check(&stats);
+        assert!(warnings.iter().any(|w| w.warning_type == WarningType::EmptyWorker));
+    }
+
+    #[test]
+    fn mailbox_overflow_detected() {
+        let mut detector = WarningDetector::new(WarningConfig::default());
+        let stats = make_stats(
+            vec![make_worker(0, 1, 42)],
+            vec![],
+        );
+        let warnings = detector.check(&stats);
+        assert!(warnings.iter().any(|w| w.warning_type == WarningType::MailboxOverflow));
+    }
+}
--- a/docs/development_history/dashboard-improvements-research.md
+++ b/docs/development_history/dashboard-improvements-research.md
@ -0,0 +1,28 @@
+# Dashboard Improvements — Research Phase
+
+## Summary
+Researched 10 comparable monitoring/dashboard systems to inform swactor's dashboard improvement plan.
+
+## Systems Analyzed
+- **Actor runtimes**: Erlang Observer (GUI/CLI/Web), Akka Insights, Ray Dashboard, Orleans Dashboard
+- **Async/runtime tools**: tokio-console, Lunatic
+- **Message/infrastructure**: RabbitMQ Management, Consul UI, Nomad UI
+- **Web frameworks**: Phoenix LiveDashboard
+
+## Key Findings
+1. **Time-series history** is table-stakes — every system provides it
+2. **Actor detail drill-down** is universal (Observer has 6-tab process info, Orleans has grain state inspection)
+3. **Search/filter** exists in every system
+4. **Warning/anomaly detection** (tokio-console's lint system) is a high-value differentiator
+5. **Topology visualization** (Consul golden metrics, Observer supervision tree) is rare but powerful
+
+## Implementation Plan
+8 feature stages defined (see `CLAUDE/notes/feature-stages/`):
+1. Time-Series History Infrastructure
+2. Actor Detail Drill-Down
+3. Search and Filter
+4. Per-Worker Utilization Visualization
+5. Warning/Anomaly Detection
+6. Actor-to-Actor Message Flow Topology
+7. Per-Actor Logging
+8. Per-Message-Type Breakdown
--- a/src/stats.rs
+++ b/src/stats.rs
@ -120,6 +120,8 @@ pub struct ActorSnapshot {
    pub last_msg_type: Option<&'static str>,
    pub messages_processed: u64,
    pub poisoned: bool,
+    /// Per-message-type counts, sorted descending by count.
+    pub message_type_counts: Vec<(&'static str, u64)>,
 }

 /// Observer hook called by workers after productive ticks.
@ -150,6 +152,9 @@ pub struct ActorInfo {
    /// Whether the actor has panicked and is no longer processing messages.
    #[cfg_attr(feature = "serde", serde(default))]
    pub poisoned: bool,
+    /// Per-message-type counts, sorted descending by count. Top 32 types.
+    #[cfg_attr(feature = "serde", serde(default))]
+    pub message_type_counts: Vec<(String, u64)>,
 }

 /// Snapshot of overall runtime state.
--- a/src/worker.rs
+++ b/src/worker.rs
@ -372,6 +372,8 @@ struct ActorSlot {
    started: bool,
    last_msg_type: Option<&'static str>,
    messages_processed: u64,
+    /// Per-message-type counters (bounded to 32 entries).
+    msg_type_counts: HashMap<&'static str, u64>,
    /// Per-actor mailbox capacity. 0 = unbounded.
    mailbox_capacity: usize,
    overflow_policy: MailboxOverflow,
@ -407,6 +409,7 @@ impl ActorPool {
            started: false,
            last_msg_type: None,
            messages_processed: 0,
+            msg_type_counts: HashMap::new(),
            mailbox_capacity: self.default_mailbox_capacity,
            overflow_policy: self.default_overflow_policy,
        });
@ -459,6 +462,9 @@ impl ActorPool {
                continue;
            }

+            #[cfg(feature = "tracing")]
+            let _actor_span = tracing::trace_span!("actor.tick", actor_addr = %addr).entered();
+
            let ctx = Ctx::new(inner, addr);

            // Call on_start once, before first message
@ -521,6 +527,10 @@ impl ActorPool {
                    Ok(Some(type_name)) => {
                        slot.last_msg_type = Some(type_name);
                        slot.messages_processed += 1;
+                        // Track per-type counts (bounded to 32 distinct types)
+                        if slot.msg_type_counts.len() < 32 || slot.msg_type_counts.contains_key(type_name) {
+                            *slot.msg_type_counts.entry(type_name).or_insert(0) += 1;
+                        }
                    }
                }
                count += 1;
@ -592,12 +602,16 @@ impl ActorPool {
    pub fn mailbox_depths_into(&self, out: &mut Vec<ActorSnapshot>) {
        out.clear();
        out.extend(self.actors.iter().map(|(&addr, slot)| {
+            let mut type_counts: Vec<(&'static str, u64)> =
+                slot.msg_type_counts.iter().map(|(&k, &v)| (k, v)).collect();
+            type_counts.sort_by(|a, b| b.1.cmp(&a.1));
            ActorSnapshot {
                address: addr,
                mailbox_depth: slot.mailbox.len(),
                last_msg_type: slot.last_msg_type,
                messages_processed: slot.messages_processed,
                poisoned: slot.poisoned,
+                message_type_counts: type_counts,
            }
        }));
    }
--- a/tests/actor_lifecycle.rs
+++ b/tests/actor_lifecycle.rs
@ -0,0 +1,790 @@
+//! Actor Lifecycle Tests — birth, life, death of individual actors.
+//!
+//! Covers: spawning, on_start, parent-child delegation, graceful stop,
+//! panic isolation, dead actor cleanup, watching (ActorExited), and
+//! monitoring (Down notifications).
+
+mod common;
+use common::*;
+
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::Arc;
+
+// ── Local actors ────────────────────────────────────────────────────────────
+
+/// Records lifecycle events to shared counters.
+struct LifecycleActor {
+    started: Arc<AtomicUsize>,
+    stopped: Arc<AtomicUsize>,
+    handled: Arc<AtomicUsize>,
+}
+
+impl ActorInterface for LifecycleActor {
+    type Incoming = Ping;
+    type Response = Pong;
+    fn on_start(&mut self, _ctx: &Ctx) {
+        self.started.fetch_add(1, Ordering::Relaxed);
+    }
+    fn on_stop(&mut self, _ctx: &Ctx) {
+        self.stopped.fetch_add(1, Ordering::Relaxed);
+    }
+    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
+        self.handled.fetch_add(1, Ordering::Relaxed);
+        let _ = ctx.send(msg.reply_to, Pong);
+    }
+}
+
+/// Stops itself after processing `stop_after` messages.
+struct SelfStopActor {
+    count: usize,
+    stop_after: usize,
+    stopped: Arc<AtomicUsize>,
+}
+
+impl ActorInterface for SelfStopActor {
+    type Incoming = Forward;
+    type Response = Done;
+    fn on_stop(&mut self, _ctx: &Ctx) {
+        self.stopped.fetch_add(1, Ordering::Relaxed);
+    }
+    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
+        self.count += 1;
+        let _ = ctx.send(msg.reply_to, Done(msg.value));
+        if self.count >= self.stop_after {
+            ctx.stop_self();
+        }
+    }
+}
+
+/// Sends a farewell Pong in on_stop.
+struct FarewellActor {
+    farewell_to: ActorAddress,
+}
+
+impl ActorInterface for FarewellActor {
+    type Incoming = Ping;
+    type Response = Pong;
+    fn on_stop(&mut self, ctx: &Ctx) {
+        let _ = ctx.send(self.farewell_to, Pong);
+    }
+    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
+        let _ = ctx.send(msg.reply_to, Pong);
+    }
+}
+
+/// Panics in on_start.
+struct PanicOnStartActor {
+    handled: Arc<AtomicUsize>,
+}
+
+impl ActorInterface for PanicOnStartActor {
+    type Incoming = Ping;
+    type Response = Pong;
+    fn on_start(&mut self, _ctx: &Ctx) {
+        panic!("on_start panic");
+    }
+    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {
+        self.handled.fetch_add(1, Ordering::Relaxed);
+    }
+}
+
+/// Spawns a DoubleActor child, sends it work, then panics.
+struct SpawnThenPanicActor;
+
+impl ActorInterface for SpawnThenPanicActor {
+    type Incoming = Forward;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
+        let child = ctx.spawn(DoubleActor).unwrap();
+        let _ = ctx.send(child, Forward { value: msg.value, reply_to: msg.reply_to });
+        panic!("intentional panic after spawn+send");
+    }
+}
+
+/// Sends a Pong reply, then panics.
+struct SendThenPanicActor;
+
+impl ActorInterface for SendThenPanicActor {
+    type Incoming = Ping;
+    type Response = Pong;
+    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
+        let _ = ctx.send(msg.reply_to, Pong);
+        panic!("intentional panic after send");
+    }
+}
+
+/// Processes `remaining_good` messages then panics.
+struct PanicAfterNActor {
+    remaining_good: usize,
+    counter: Arc<AtomicUsize>,
+}
+
+impl ActorInterface for PanicAfterNActor {
+    type Incoming = Ping;
+    type Response = ();
+    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {
+        if self.remaining_good == 0 {
+            panic!("intentional delayed panic");
+        }
+        self.remaining_good -= 1;
+        self.counter.fetch_add(1, Ordering::SeqCst);
+    }
+}
+
+/// Stops on a trigger message.
+struct StopOnTrigger(Arc<AtomicUsize>);
+
+#[derive(Clone)]
+struct Trigger(bool);
+
+impl ActorInterface for StopOnTrigger {
+    type Incoming = Trigger;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, msg: Trigger) {
+        self.0.fetch_add(1, Ordering::Relaxed);
+        if msg.0 {
+            ctx.stop_self();
+        }
+    }
+}
+
+/// Watches targets and counts exit notifications via on_actor_exit.
+struct ExitWatcher {
+    exit_count: Arc<AtomicUsize>,
+    last_reason: Arc<std::sync::Mutex<Option<ExitReason>>>,
+    last_addr: Arc<std::sync::Mutex<Option<ActorAddress>>>,
+}
+
+#[derive(Clone)]
+enum WatcherCmd {
+    WatchThis(ActorAddress),
+    UnwatchThis(ActorAddress),
+}
+
+impl ActorInterface for ExitWatcher {
+    type Incoming = WatcherCmd;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, msg: WatcherCmd) {
+        match msg {
+            WatcherCmd::WatchThis(target) => ctx.watch(target),
+            WatcherCmd::UnwatchThis(target) => ctx.unwatch(target),
+        }
+    }
+    fn on_actor_exit(&mut self, _ctx: &Ctx, exited: ActorExited) {
+        self.exit_count.fetch_add(1, Ordering::SeqCst);
+        *self.last_reason.lock().unwrap() = Some(exited.reason);
+        *self.last_addr.lock().unwrap() = Some(exited.addr);
+    }
+}
+
+struct WatcherState {
+    exit_count: Arc<AtomicUsize>,
+    last_reason: Arc<std::sync::Mutex<Option<ExitReason>>>,
+}
+
+impl WatcherState {
+    fn count(&self) -> usize {
+        self.exit_count.load(Ordering::SeqCst)
+    }
+    fn last_reason(&self) -> Option<ExitReason> {
+        self.last_reason.lock().unwrap().clone()
+    }
+}
+
+fn new_exit_watcher() -> (ExitWatcher, WatcherState) {
+    let exit_count = Arc::new(AtomicUsize::new(0));
+    let last_reason = Arc::new(std::sync::Mutex::new(None));
+    let last_addr = Arc::new(std::sync::Mutex::new(None));
+    let state = WatcherState {
+        exit_count: exit_count.clone(),
+        last_reason: last_reason.clone(),
+    };
+    (
+        ExitWatcher { exit_count, last_reason, last_addr },
+        state,
+    )
+}
+
+/// Monitors a target and forwards Down to a reply address.
+struct MonitorWatcherActor {
+    watch_target: ActorAddress,
+    reply_to: ActorAddress,
+    mref: Option<MonitorRef>,
+}
+
+impl ActorInterface for MonitorWatcherActor {
+    type Incoming = Down;
+    type Response = ();
+    fn on_start(&mut self, ctx: &Ctx) {
+        self.mref = Some(ctx.monitor(self.watch_target));
+    }
+    fn handle(&mut self, ctx: &Ctx, msg: Down) {
+        ctx.send(self.reply_to, msg).unwrap();
+    }
+}
+
+/// Demonitors on Ping.
+struct DemonitorActor {
+    watch_target: ActorAddress,
+    mref: Option<MonitorRef>,
+}
+
+impl ActorInterface for DemonitorActor {
+    type Incoming = Ping;
+    type Response = ();
+    fn on_start(&mut self, ctx: &Ctx) {
+        self.mref = Some(ctx.monitor(self.watch_target));
+    }
+    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
+        if let Some(mref) = self.mref.take() {
+            ctx.demonitor(mref);
+        }
+    }
+}
+
+/// A silent actor that does nothing (target for watching tests).
+struct Sleeper;
+#[derive(Clone)]
+struct Noop;
+
+impl ActorInterface for Sleeper {
+    type Incoming = Noop;
+    type Response = ();
+    fn handle(&mut self, _ctx: &Ctx, _msg: Noop) {}
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Tests
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Actors are spawned, on_start fires exactly once per instance before any
+/// message, then state accumulates across messages.
+#[test]
+fn actor_from_birth_to_first_message() {
+    let started = Arc::new(AtomicUsize::new(0));
+    let stopped = Arc::new(AtomicUsize::new(0));
+    let handled = Arc::new(AtomicUsize::new(0));
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+
+    // Spawn one tracked actor + 4 more sharing the same counters
+    let addr = rt.spawn(LifecycleActor {
+        started: started.clone(),
+        stopped: stopped.clone(),
+        handled: handled.clone(),
+    }).unwrap();
+    for _ in 0..4 {
+        rt.spawn(LifecycleActor {
+            started: started.clone(),
+            stopped: stopped.clone(),
+            handled: handled.clone(),
+        }).unwrap();
+    }
+
+    // First tick: all 5 on_start fire, no messages processed yet
+    rt.tick();
+    assert_eq!(started.load(Ordering::Relaxed), 5, "on_start per instance");
+    assert_eq!(handled.load(Ordering::Relaxed), 0, "no messages before first send");
+
+    // Send 3 Increments to a CounterActor to verify state accumulation
+    let counter_addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    let count_inbox = rt.new_inbox::<Count>().unwrap();
+    for _ in 0..3 {
+        rt.send_to(counter_addr, Increment { reply_to: *count_inbox.addr() }).unwrap();
+    }
+    let replies = tick_and_drain(&rt, &count_inbox, 10);
+    assert_eq!(replies, vec![Count(1), Count(2), Count(3)], "state accumulates");
+
+    // on_start must not fire again on subsequent ticks
+    rt.tick();
+    rt.tick();
+    assert_eq!(started.load(Ordering::Relaxed), 5, "on_start not repeated");
+
+    // Verify the first actor still responds normally
+    rt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 10);
+    assert!(reply.is_some(), "actor handles messages after on_start");
+}
+
+/// Delegation chains: parent spawns child, child spawns grandchild, fan-out
+/// distributes work. Spawn+send interleaving in a single handler works.
+#[test]
+fn parent_child_delegation_and_spawn_chains() {
+    let rt = std_runtime(RuntimeConfig {
+        max_actors: 2000,
+        ..Default::default()
+    });
+
+    // Act 1: DelegatorActor spawns child, forwards value 7 → Done(14)
+    let delegator = rt.spawn(DelegatorActor).unwrap();
+    let inbox = rt.new_inbox::<Done>().unwrap();
+    rt.send_to(delegator, Forward { value: 7, reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 20);
+    assert_eq!(reply, Some(Done(14)), "delegator child doubles value");
+
+    // Act 2: Chain of depth 20
+    let chain = rt.spawn(ChainActor).unwrap();
+    rt.send_to(chain, ChainMsg { remaining: 20, depth: 0, reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 200);
+    assert_eq!(reply, Some(Done(20)), "chain reaches depth 20");
+
+    // Act 3: Fan-out to 20 children
+    let fan = rt.spawn(FanOutActor).unwrap();
+    rt.send_to(fan, FanOut { count: 20, reply_to: *inbox.addr() }).unwrap();
+    let replies = tick_and_drain(&rt, &inbox, 50);
+    assert_eq!(replies.len(), 20, "all 20 fan-out children reply");
+}
+
+/// The full graceful-stop story: self-stop with on_stop, farewell messages,
+/// external stop ordering vs pending messages, mid-mailbox stop trigger.
+#[test]
+fn graceful_stop_lifecycle() {
+    // --- Part A: SelfStopActor ---
+    let stopped = Arc::new(AtomicUsize::new(0));
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Done>().unwrap();
+
+    let addr = rt.spawn(SelfStopActor {
+        count: 0,
+        stop_after: 3,
+        stopped: stopped.clone(),
+    }).unwrap();
+
+    for i in 0..5 {
+        let _ = rt.send_to(addr, Forward { value: i, reply_to: *inbox.addr() });
+    }
+    tick_n(&rt, 10);
+
+    let mut replies = Vec::new();
+    while let Some(Done(v)) = inbox.try_recv() {
+        replies.push(v);
+    }
+    assert_eq!(replies.len(), 3, "only 3 messages processed before self-stop");
+    assert_eq!(stopped.load(Ordering::Relaxed), 1, "on_stop fired");
+    assert!(rt.send_to(addr, Forward { value: 99, reply_to: *inbox.addr() }).is_err(),
+        "send to stopped actor fails");
+
+    // --- Part B: FarewellActor sends farewell in on_stop ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(FarewellActor { farewell_to: *inbox.addr() }).unwrap();
+    rt.tick();
+    rt.stop_actor(addr).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(inbox.try_recv(), Some(Pong), "farewell message delivered from on_stop");
+
+    // --- Part C: External stop after pending messages ---
+    let stopped = Arc::new(AtomicUsize::new(0));
+    let handled = Arc::new(AtomicUsize::new(0));
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(LifecycleActor {
+        started: Arc::new(AtomicUsize::new(0)),
+        stopped: stopped.clone(),
+        handled: handled.clone(),
+    }).unwrap();
+    for _ in 0..10 {
+        let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
+    }
+    rt.stop_actor(addr).unwrap();
+    tick_n(&rt, 10);
+    assert_eq!(handled.load(Ordering::Relaxed), 10, "all pending messages processed before stop");
+    assert_eq!(stopped.load(Ordering::Relaxed), 1, "on_stop fires after messages");
+
+    // --- Part D: External stop before messages → 0 processed ---
+    let handled = Arc::new(AtomicUsize::new(0));
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(LifecycleActor {
+        started: Arc::new(AtomicUsize::new(0)),
+        stopped: Arc::new(AtomicUsize::new(0)),
+        handled: handled.clone(),
+    }).unwrap();
+    rt.tick(); // on_start
+    rt.stop_actor(addr).unwrap();
+    for _ in 0..5 {
+        let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
+    }
+    tick_n(&rt, 10);
+    assert_eq!(handled.load(Ordering::Relaxed), 0, "stop before messages prevents processing");
+
+    // --- Part E: Mid-mailbox stop trigger ---
+    let processed = Arc::new(AtomicUsize::new(0));
+    let rt = std_runtime(RuntimeConfig::default());
+    let addr = rt.spawn(StopOnTrigger(processed.clone())).unwrap();
+    rt.tick();
+    rt.send_to(addr, Trigger(false)).unwrap();
+    rt.send_to(addr, Trigger(false)).unwrap();
+    rt.send_to(addr, Trigger(true)).unwrap(); // stop trigger
+    rt.send_to(addr, Trigger(false)).unwrap();
+    rt.send_to(addr, Trigger(false)).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(processed.load(Ordering::Relaxed), 3,
+        "only messages up to and including stop trigger processed");
+    assert!(rt.send_to(addr, Trigger(false)).is_err());
+}
+
+/// Panics are caught: healthy siblings survive, panicked actors are poisoned
+/// and cleaned from stats/address map, mid-batch panic discards remaining,
+/// child spawned before parent panic survives, message sent before panic is
+/// delivered, bulk cleanup, on_start panic also poisons.
+#[test]
+fn panic_isolation_and_cleanup() {
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let count_inbox = rt.new_inbox::<Count>().unwrap();
+
+    // Spawn a healthy counter, a PanicActor, and a PanicOnStartActor
+    let good = rt.spawn(CounterActor { count: 0 }).unwrap();
+    let bad = rt.spawn(PanicActor).unwrap();
+    let bad_start_handled = Arc::new(AtomicUsize::new(0));
+    let bad_start = rt.spawn(PanicOnStartActor { handled: bad_start_handled.clone() }).unwrap();
+
+    // Trigger panics
+    rt.send_to(bad, PanicMsg).unwrap();
+    let _ = rt.send_to(bad_start, Ping { reply_to: *inbox.addr() });
+    tick_n(&rt, 10);
+
+    // Healthy actor still works
+    rt.send_to(good, Increment { reply_to: *count_inbox.addr() }).unwrap();
+    rt.send_to(good, Increment { reply_to: *count_inbox.addr() }).unwrap();
+    let replies = tick_and_drain(&rt, &count_inbox, 10);
+    assert_eq!(replies, vec![Count(1), Count(2)], "healthy actor unaffected by peer panics");
+
+    // Poisoned actors are cleaned from address map
+    assert!(rt.send_to(bad, PanicMsg).is_err(), "send to cleaned-up actor fails");
+    assert_eq!(bad_start_handled.load(Ordering::Relaxed), 0, "on_start panic prevents messages");
+
+    // Stats track panics vs stops separately
+    let stats = rt.stats();
+    let panics: u64 = stats.workers.iter().map(|w| w.panics).sum();
+    assert!(panics >= 2, "at least 2 panics recorded (PanicActor + PanicOnStartActor)");
+
+    // --- Mid-batch panic discards remaining ---
+    let counter = Arc::new(AtomicUsize::new(0));
+    let rt = std_runtime(RuntimeConfig::default());
+    let dummy = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(PanicAfterNActor { remaining_good: 2, counter: counter.clone() }).unwrap();
+    for _ in 0..5 {
+        rt.send_to(addr, Ping { reply_to: *dummy.addr() }).unwrap();
+    }
+    tick_n(&rt, 20);
+    assert_eq!(counter.load(Ordering::SeqCst), 2, "only messages before panic processed");
+
+    // --- Child spawned before parent panic survives ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Done>().unwrap();
+    let parent = rt.spawn(SpawnThenPanicActor).unwrap();
+    rt.send_to(parent, Forward { value: 5, reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 30);
+    assert_eq!(reply, Some(Done(10)), "child survives parent panic");
+
+    // --- Message sent before panic is delivered ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(SendThenPanicActor).unwrap();
+    rt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 20);
+    assert!(reply.is_some(), "message sent before panic still delivered");
+
+    // --- Bulk cleanup: 20 panicking actors all cleaned ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let mut addrs = Vec::new();
+    for _ in 0..20 {
+        addrs.push(rt.spawn(PanicActor).unwrap());
+    }
+    for &addr in &addrs {
+        let _ = rt.send_to(addr, PanicMsg);
+    }
+    tick_n(&rt, 10);
+    let stats = rt.stats();
+    assert_eq!(stats.workers[0].num_actors, 0, "all poisoned actors cleaned up");
+}
+
+/// Watch API contract: watchers are notified on death, unwatch cancels,
+/// double-watch is idempotent, multiple watchers all notified,
+/// runtime-level watch works.
+#[test]
+fn watch_notification_contract() {
+    let rt = std_runtime(RuntimeConfig::default());
+
+    // Spawn target + 3 watchers + 1 that unwatches
+    let target = rt.spawn(PanicActor).unwrap();
+    let (w1, s1) = new_exit_watcher();
+    let (w2, s2) = new_exit_watcher();
+    let (w3, s3) = new_exit_watcher();
+    let (w4, s4) = new_exit_watcher(); // will unwatch
+
+    let w1_addr = rt.spawn(w1).unwrap();
+    let w2_addr = rt.spawn(w2).unwrap();
+    let w3_addr = rt.spawn(w3).unwrap();
+    let w4_addr = rt.spawn(w4).unwrap();
+
+    // All watch the target
+    rt.send_to(w1_addr, WatcherCmd::WatchThis(target)).unwrap();
+    rt.send_to(w2_addr, WatcherCmd::WatchThis(target)).unwrap();
+    rt.send_to(w3_addr, WatcherCmd::WatchThis(target)).unwrap();
+    rt.send_to(w4_addr, WatcherCmd::WatchThis(target)).unwrap();
+    tick_n(&rt, 3);
+
+    // w2 double-watches (idempotent test)
+    rt.send_to(w2_addr, WatcherCmd::WatchThis(target)).unwrap();
+    tick_n(&rt, 3);
+
+    // w4 unwatches
+    rt.send_to(w4_addr, WatcherCmd::UnwatchThis(target)).unwrap();
+    tick_n(&rt, 3);
+
+    // Kill target
+    rt.send_to(target, PanicMsg).unwrap();
+    tick_n(&rt, 5);
+
+    assert_eq!(s1.count(), 1, "watcher 1 notified");
+    assert_eq!(s2.count(), 1, "double-watch still only one notification");
+    assert_eq!(s3.count(), 1, "watcher 3 notified");
+    assert_eq!(s4.count(), 0, "unwatched watcher not notified");
+    assert_eq!(s1.last_reason(), Some(ExitReason::Panicked));
+
+    // --- Runtime-level watch ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let target = rt.spawn(PanicActor).unwrap();
+    let (w, s) = new_exit_watcher();
+    let w_addr = rt.spawn(w).unwrap();
+    tick_n(&rt, 2);
+    rt.watch(w_addr, target);
+    rt.send_to(target, PanicMsg).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(s.count(), 1, "runtime-level watch delivers notification");
+}
+
+/// Watch edge cases: watcher dies before target (no crash), self-watch (no
+/// crash), watcher reacts to death by spawning a replacement.
+#[test]
+fn watch_edge_cases() {
+    // Watcher dies before target — no crash
+    let rt = std_runtime(RuntimeConfig::default());
+    let target = rt.spawn(PanicActor).unwrap();
+    let target2 = rt.spawn(PanicActor).unwrap();
+    rt.watch(target2, target);
+    tick_n(&rt, 3);
+    rt.send_to(target2, PanicMsg).unwrap(); // kill watcher first
+    tick_n(&rt, 5);
+    rt.send_to(target, PanicMsg).unwrap(); // kill target — no crash
+    tick_n(&rt, 5);
+
+    // Self-watch — no crash
+    let rt = std_runtime(RuntimeConfig::default());
+    let (w, _s) = new_exit_watcher();
+    let addr = rt.spawn(w).unwrap();
+    rt.send_to(addr, WatcherCmd::WatchThis(addr)).unwrap();
+    tick_n(&rt, 5);
+
+    // Watcher reacts to death by spawning replacement
+    let rt = std_runtime(RuntimeConfig::default());
+    let spawned = Arc::new(AtomicUsize::new(0));
+
+    struct SupervisorWatcher {
+        spawned_count: Arc<AtomicUsize>,
+    }
+
+    #[derive(Clone)]
+    enum SupCmd {
+        WatchThis(ActorAddress),
+    }
+
+    impl ActorInterface for SupervisorWatcher {
+        type Incoming = SupCmd;
+        type Response = ();
+        fn handle(&mut self, ctx: &Ctx, msg: SupCmd) {
+            match msg {
+                SupCmd::WatchThis(target) => ctx.watch(target),
+            }
+        }
+        fn on_actor_exit(&mut self, ctx: &Ctx, _exited: ActorExited) {
+            let _ = ctx.spawn(Sleeper);
+            self.spawned_count.fetch_add(1, Ordering::SeqCst);
+        }
+    }
+
+    let target = rt.spawn(PanicActor).unwrap();
+    let sup = rt.spawn(SupervisorWatcher { spawned_count: spawned.clone() }).unwrap();
+    rt.send_to(sup, SupCmd::WatchThis(target)).unwrap();
+    tick_n(&rt, 3);
+    rt.send_to(target, PanicMsg).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(spawned.load(Ordering::SeqCst), 1, "watcher spawned replacement");
+}
+
+/// Monitor API contract: Down on stop (Normal) and panic (Panicked), multiple
+/// monitors, demonitor cancels, dead watcher cleanup, stacked monitors,
+/// external inbox, handle_down dispatch.
+#[test]
+fn monitor_death_notification_contract() {
+    let rt = std_runtime(RuntimeConfig::default());
+
+    // --- Stop → Down(Normal) ---
+    let inbox = rt.new_inbox::<Down>().unwrap();
+    let target = rt.spawn(PingPongActor).unwrap();
+    rt.spawn(MonitorWatcherActor {
+        watch_target: target,
+        reply_to: *inbox.addr(),
+        mref: None,
+    }).unwrap();
+    rt.tick();
+    rt.stop_actor(target).unwrap();
+    tick_n(&rt, 3);
+    let down = inbox.try_recv().expect("Down on graceful stop");
+    assert_eq!(down.addr, target);
+    assert_eq!(down.reason, StopReason::Normal);
+
+    // --- Panic → Down(Panicked) ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Down>().unwrap();
+    let target = rt.spawn(PanicActor).unwrap();
+    rt.spawn(MonitorWatcherActor {
+        watch_target: target,
+        reply_to: *inbox.addr(),
+        mref: None,
+    }).unwrap();
+    rt.tick();
+    rt.send_to(target, PanicMsg).unwrap();
+    tick_n(&rt, 3);
+    let down = inbox.try_recv().expect("Down on panic");
+    assert_eq!(down.reason, StopReason::Panicked);
+
+    // --- Multiple monitors ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox1 = rt.new_inbox::<Down>().unwrap();
+    let inbox2 = rt.new_inbox::<Down>().unwrap();
+    let target = rt.spawn(PingPongActor).unwrap();
+    rt.spawn(MonitorWatcherActor {
+        watch_target: target, reply_to: *inbox1.addr(), mref: None,
+    }).unwrap();
+    rt.spawn(MonitorWatcherActor {
+        watch_target: target, reply_to: *inbox2.addr(), mref: None,
+    }).unwrap();
+    rt.tick();
+    rt.stop_actor(target).unwrap();
+    tick_n(&rt, 3);
+    assert!(inbox1.try_recv().is_some(), "watcher 1 notified");
+    assert!(inbox2.try_recv().is_some(), "watcher 2 notified");
+
+    // --- Demonitor cancels ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let down_inbox = rt.new_inbox::<Down>().unwrap();
+    let target = rt.spawn(PingPongActor).unwrap();
+    let watcher = rt.spawn(DemonitorActor { watch_target: target, mref: None }).unwrap();
+    rt.tick();
+    rt.send_to(watcher, Ping { reply_to: ActorAddress::default() }).unwrap();
+    rt.tick(); // demonitor
+    rt.stop_actor(target).unwrap();
+    tick_n(&rt, 3);
+    assert!(down_inbox.try_recv().is_none(), "demonitored: no Down delivered");
+
+    // --- Dead watcher cleaned up ---
+    let rt = std_runtime(RuntimeConfig::default());
+    let target = rt.spawn(PingPongActor).unwrap();
+    let watcher = rt.spawn(MonitorWatcherActor {
+        watch_target: target,
+        reply_to: ActorAddress::default(),
+        mref: None,
+    }).unwrap();
+    rt.tick();
+    rt.stop_actor(watcher).unwrap();
+    rt.tick(); // watcher dies
+    rt.stop_actor(target).unwrap();
+    tick_n(&rt, 3); // target dies — no crash trying to deliver to dead watcher
+
+    // --- Stacked monitors produce multiple notifications ---
+    struct DoubleMonitor {
+        target: ActorAddress,
+        reply_to: ActorAddress,
+    }
+    impl ActorInterface for DoubleMonitor {
+        type Incoming = Down;
+        type Response = ();
+        fn on_start(&mut self, ctx: &Ctx) {
+            ctx.monitor(self.target);
+            ctx.monitor(self.target);
+        }
+        fn handle(&mut self, ctx: &Ctx, msg: Down) {
+            ctx.send(self.reply_to, msg).unwrap();
+        }
+    }
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Down>().unwrap();
+    let target = rt.spawn(PingPongActor).unwrap();
+    rt.spawn(DoubleMonitor { target, reply_to: *inbox.addr() }).unwrap();
+    rt.tick();
+    rt.stop_actor(target).unwrap();
+    tick_n(&rt, 3);
+    assert!(inbox.try_recv().is_some(), "first Down from stacked monitor");
+    assert!(inbox.try_recv().is_some(), "second Down from stacked monitor");
+    assert!(inbox.try_recv().is_none(), "no more");
+
+    // --- handle_down dispatch ---
+    struct MonitoringTracker {
+        target: ActorAddress,
+        downs: Vec<Down>,
+        inbox: ActorAddress,
+    }
+    impl ActorInterface for MonitoringTracker {
+        type Incoming = Ping;
+        type Response = ();
+        fn on_start(&mut self, ctx: &Ctx) {
+            ctx.monitor(self.target);
+        }
+        fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
+            let _ = ctx.send(self.inbox, Count(self.downs.len()));
+        }
+        fn handle_down(&mut self, _ctx: &Ctx, down: Down) {
+            self.downs.push(down);
+        }
+    }
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Count>().unwrap();
+    let target = rt.spawn(PanicActor).unwrap();
+    let tracker = rt.spawn(MonitoringTracker {
+        target,
+        downs: vec![],
+        inbox: *inbox.addr(),
+    }).unwrap();
+    rt.tick();
+    rt.send_to(target, PanicMsg).unwrap();
+    tick_n(&rt, 3);
+    rt.send_to(tracker, Ping { reply_to: ActorAddress::default() }).unwrap();
+    rt.tick();
+    assert_eq!(inbox.try_recv(), Some(Count(1)), "handle_down received exactly one Down");
+
+    // --- When Incoming=Down, handle_down is NOT called ---
+    struct DownAsIncoming {
+        target: ActorAddress,
+        inbox: ActorAddress,
+    }
+    impl ActorInterface for DownAsIncoming {
+        type Incoming = Down;
+        type Response = ();
+        fn on_start(&mut self, ctx: &Ctx) {
+            ctx.monitor(self.target);
+        }
+        fn handle(&mut self, ctx: &Ctx, msg: Down) {
+            let _ = ctx.send(self.inbox, msg);
+        }
+        fn handle_down(&mut self, _ctx: &Ctx, _down: Down) {
+            panic!("handle_down must not be called when Incoming=Down");
+        }
+    }
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Down>().unwrap();
+    let target = rt.spawn(PanicActor).unwrap();
+    rt.spawn(DownAsIncoming { target, inbox: *inbox.addr() }).unwrap();
+    rt.tick();
+    rt.send_to(target, PanicMsg).unwrap();
+    tick_n(&rt, 3);
+    let received = inbox.try_recv().expect("Down delivered via handle(), not handle_down");
+    assert_eq!(received.reason, StopReason::Panicked);
+}
--- a/tests/common/mod.rs
+++ b/tests/common/mod.rs
@ -5,11 +5,14 @@
 use std::sync::atomic::{AtomicUsize, Ordering};
 use std::sync::Arc;

-pub use swactor::actor::{ActorAddress, ActorInterface, Down, MonitorRef, StopReason};
+pub use swactor::actor::{
+    ActorAddress, ActorExited, ActorInterface, Down, ExitReason, MonitorRef, StopReason,
+};
 pub use swactor::runtime::{Ctx, Inbox, MailboxOverflow, Runtime, RuntimeConfig};
 pub use swactor_std::{
-    ChildSpec, CtxGroups, CtxMonitoring, CtxNaming, CtxTimers, RestartPolicy, Router,
-    RoutingStrategy, RuntimeGroups, RuntimeNaming, StdExtension, Supervisor, SupervisorStrategy,
+    ChildSpec, CtxGroups, CtxMonitoring, CtxNaming, CtxTimers, CtxWatching, RestartPolicy, Router,
+    RoutingStrategy, RuntimeGroups, RuntimeNaming, RuntimeWatching, StdExtension, Supervisor,
+    SupervisorStrategy,
 };

 // ── Messages ────────────────────────────────────────────────────────────────
@ -232,6 +235,13 @@ pub fn tick_until_recv<M: swactor::actor::Message>(
    None
 }

+/// Tick exactly `n` times (no inbox polling).
+pub fn tick_n(rt: &Runtime, n: usize) {
+    for _ in 0..n {
+        rt.tick();
+    }
+}
+
 /// Tick `n` times, then drain all messages from the inbox.
 pub fn tick_and_drain<M: swactor::actor::Message>(
    rt: &Runtime,
--- a/tests/message_delivery.rs
+++ b/tests/message_delivery.rs
@ -0,0 +1,489 @@
+//! Message Delivery Tests — how data flows through the system.
+//!
+//! Covers: FIFO ordering, routing correctness at scale, delivery from within
+//! handlers, address error handling, fairness/budgets, timers, and mailbox
+//! backpressure policies.
+
+mod common;
+use common::*;
+
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::Arc;
+
+// ── Local actors ────────────────────────────────────────────────────────────
+
+/// Sends a countdown message to itself, then replies Done(0).
+struct SelfSendActor;
+
+#[derive(Clone)]
+struct Countdown {
+    remaining: usize,
+    reply_to: ActorAddress,
+}
+
+impl ActorInterface for SelfSendActor {
+    type Incoming = Countdown;
+    type Response = Done;
+    fn handle(&mut self, ctx: &Ctx, msg: Countdown) {
+        if msg.remaining == 0 {
+            let _ = ctx.send(msg.reply_to, Done(0));
+        } else {
+            let _ = ctx.send(
+                ctx.self_addr(),
+                Countdown { remaining: msg.remaining - 1, reply_to: msg.reply_to },
+            );
+        }
+    }
+}
+
+/// Schedules a one-shot timer in on_start.
+struct TimerStartActor {
+    target: ActorAddress,
+    delay_ticks: u64,
+}
+
+impl ActorInterface for TimerStartActor {
+    type Incoming = Ping;
+    type Response = Pong;
+    fn on_start(&mut self, ctx: &Ctx) {
+        ctx.send_after_ticks(self.target, Ping { reply_to: ctx.self_addr() }, self.delay_ticks);
+    }
+    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {}
+}
+
+/// Schedules a one-shot timer from a handler.
+struct DelayPingPongActor;
+
+impl ActorInterface for DelayPingPongActor {
+    type Incoming = Forward;
+    type Response = Done;
+    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
+        ctx.send_after_ticks(msg.reply_to, Done(msg.value), 3);
+    }
+}
+
+/// Schedules an interval timer on start.
+struct HeartbeatActor {
+    target: ActorAddress,
+    period: u64,
+}
+
+impl ActorInterface for HeartbeatActor {
+    type Incoming = Ping;
+    type Response = Pong;
+    fn on_start(&mut self, ctx: &Ctx) {
+        ctx.send_interval_ticks(self.target, Ping { reply_to: ctx.self_addr() }, self.period);
+    }
+    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {}
+}
+
+/// NumberedMsg/Reply for routing correctness tests.
+#[derive(Clone)]
+struct NumberedMsg {
+    n: usize,
+    reply_to: ActorAddress,
+}
+
+#[derive(Clone, Debug, PartialEq)]
+struct NumberedReply {
+    from: ActorAddress,
+    n: usize,
+}
+
+struct NumberedActor;
+
+impl ActorInterface for NumberedActor {
+    type Incoming = NumberedMsg;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, msg: NumberedMsg) {
+        let _ = ctx.send(msg.reply_to, NumberedReply { from: ctx.self_addr(), n: msg.n });
+    }
+}
+
+/// Ring node for routing chain test.
+#[derive(Clone)]
+struct RingHop {
+    hops_remaining: usize,
+    final_dest: ActorAddress,
+}
+
+#[derive(Clone, Debug, PartialEq)]
+struct RingDone(usize);
+
+struct RingNode {
+    next: ActorAddress,
+}
+
+impl ActorInterface for RingNode {
+    type Incoming = RingHop;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, msg: RingHop) {
+        if msg.hops_remaining == 0 {
+            let _ = ctx.send(msg.final_dest, RingDone(100));
+        } else {
+            let _ = ctx.send(self.next, RingHop {
+                hops_remaining: msg.hops_remaining - 1,
+                final_dest: msg.final_dest,
+            });
+        }
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Tests
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Messages arrive in FIFO order even with small buffers, budget constraints,
+/// and independent mailboxes isolate actors from each other.
+#[test]
+fn fifo_ordering_and_mailbox_isolation() {
+    // FIFO with small buffer and budget
+    let rt = std_runtime(RuntimeConfig {
+        channel_buffer_size: 1,
+        actor_message_budget: 8,
+        ..Default::default()
+    });
+    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    let inbox = rt.new_inbox::<Count>().unwrap();
+    for _ in 0..100 {
+        rt.send_to(addr, Increment { reply_to: *inbox.addr() }).unwrap();
+    }
+    let replies: Vec<_> = tick_and_drain(&rt, &inbox, 50);
+    assert_eq!(replies.len(), 100, "all messages delivered");
+    for (i, reply) in replies.iter().enumerate() {
+        assert_eq!(*reply, Count(i + 1), "FIFO order preserved at position {i}");
+    }
+
+    // Mailbox isolation: 3 actors each get exactly their own messages
+    let rt = std_runtime(RuntimeConfig::default());
+    let mut inboxes = Vec::new();
+    for _ in 0..3 {
+        let addr = rt.spawn(PingPongActor).unwrap();
+        let inbox = rt.new_inbox::<Pong>().unwrap();
+        rt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
+        inboxes.push(inbox);
+    }
+    tick_n(&rt, 10);
+    for (i, inbox) in inboxes.iter().enumerate() {
+        assert!(inbox.try_recv().is_some(), "actor {i} replied");
+        assert!(inbox.try_recv().is_none(), "actor {i} has exactly one reply");
+    }
+}
+
+/// 200 actors each get a unique numbered message and reply correctly.
+/// A 100-hop ring traversal completes.
+#[test]
+fn message_routing_at_scale() {
+    // 200-actor numbered routing
+    let rt = std_runtime(RuntimeConfig {
+        max_actors: 300,
+        channel_buffer_size: 1024,
+        num_threads: 1,
+        ..Default::default()
+    });
+    let inbox = rt.new_inbox::<NumberedReply>().unwrap();
+    let inbox_addr = *inbox.addr();
+    let mut addrs = Vec::new();
+    for _ in 0..200 {
+        addrs.push(rt.spawn(NumberedActor).unwrap());
+    }
+    rt.tick();
+    for (i, addr) in addrs.iter().enumerate() {
+        rt.send_to(*addr, NumberedMsg { n: i, reply_to: inbox_addr }).unwrap();
+    }
+    tick_n(&rt, 3);
+    let replies: Vec<NumberedReply> = std::iter::from_fn(|| inbox.try_recv()).collect();
+    assert_eq!(replies.len(), 200, "all 200 actors replied");
+    for (i, addr) in addrs.iter().enumerate() {
+        let reply = replies.iter().find(|r| r.n == i);
+        assert!(reply.is_some(), "missing reply for actor #{i}");
+        assert_eq!(reply.unwrap().from, *addr, "reply #{i} came from correct actor");
+    }
+
+    // 100-hop ring
+    let rt = std_runtime(RuntimeConfig {
+        max_actors: 200,
+        channel_buffer_size: 1024,
+        num_threads: 1,
+        ..Default::default()
+    });
+    let inbox = rt.new_inbox::<RingDone>().unwrap();
+    let inbox_addr = *inbox.addr();
+    let mut ring_addrs = Vec::new();
+    let mut next = inbox_addr;
+    for _ in (0..100).rev() {
+        let addr = rt.spawn(RingNode { next }).unwrap();
+        ring_addrs.push(addr);
+        next = addr;
+    }
+    ring_addrs.reverse();
+    rt.tick();
+    rt.send_to(ring_addrs[0], RingHop { hops_remaining: 99, final_dest: inbox_addr }).unwrap();
+    let result = tick_until_recv(&rt, &inbox, 110);
+    assert_eq!(result, Some(RingDone(100)), "ring message traverses all 100 hops");
+}
+
+/// Messages sent in handlers are delivered: delegation, self-send chains,
+/// rapid spawn+immediate-send, multiple inbox types coexist.
+#[test]
+fn delivery_from_within_handlers() {
+    let rt = std_runtime(RuntimeConfig::default());
+
+    // Delegation: spawn+send in handler
+    let delegator = rt.spawn(DelegatorActor).unwrap();
+    let inbox = rt.new_inbox::<Done>().unwrap();
+    rt.send_to(delegator, Forward { value: 5, reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 20);
+    assert_eq!(reply, Some(Done(10)), "child spawned during handler receives message");
+
+    // Self-send countdown of 20
+    let self_sender = rt.spawn(SelfSendActor).unwrap();
+    rt.send_to(self_sender, Countdown { remaining: 20, reply_to: *inbox.addr() }).unwrap();
+    let reply = tick_until_recv(&rt, &inbox, 50);
+    assert_eq!(reply, Some(Done(0)), "self-send chain completes");
+
+    // Multiple senders reach same actor
+    let counter = rt.spawn(CounterActor { count: 0 }).unwrap();
+    let inbox_a = rt.new_inbox::<Count>().unwrap();
+    let inbox_b = rt.new_inbox::<Count>().unwrap();
+    rt.send_to(counter, Increment { reply_to: *inbox_a.addr() }).unwrap();
+    rt.send_to(counter, Increment { reply_to: *inbox_b.addr() }).unwrap();
+    tick_n(&rt, 10);
+    assert!(inbox_a.try_recv().is_some());
+    assert_eq!(inbox_b.try_recv(), Some(Count(2)), "both senders reach same actor");
+
+    // 50 rapid spawn+immediate-send pairs
+    let rt = std_runtime(RuntimeConfig::default());
+    let pong_inbox = rt.new_inbox::<Pong>().unwrap();
+    for _ in 0..50 {
+        let addr = rt.spawn(PingPongActor).unwrap();
+        rt.send_to(addr, Ping { reply_to: *pong_inbox.addr() }).unwrap();
+    }
+    let replies = tick_and_drain(&rt, &pong_inbox, 50);
+    assert_eq!(replies.len(), 50, "all spawn+send pairs complete");
+
+    // Multiple inbox types coexist
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter_addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    let pinger_addr = rt.spawn(PingPongActor).unwrap();
+    let count_inbox = rt.new_inbox::<Count>().unwrap();
+    let pong_inbox = rt.new_inbox::<Pong>().unwrap();
+    rt.send_to(counter_addr, Increment { reply_to: *count_inbox.addr() }).unwrap();
+    rt.send_to(pinger_addr, Ping { reply_to: *pong_inbox.addr() }).unwrap();
+    tick_n(&rt, 10);
+    assert_eq!(count_inbox.try_recv(), Some(Count(1)));
+    assert_eq!(pong_inbox.try_recv(), Some(Pong));
+}
+
+/// Sending to nonexistent address returns error, wrong type increments
+/// type_mismatch counter.
+#[test]
+fn address_error_handling() {
+    let rt = std_runtime(RuntimeConfig::default());
+
+    // Nonexistent address
+    let bogus = ActorAddress::new_random();
+    assert!(rt.send_to(bogus, Pong).is_err(), "send to unknown address fails");
+
+    // Wrong type
+    let addr = rt.spawn(PingPongActor).unwrap();
+    rt.send_to(addr, Count(42)).unwrap(); // Count instead of Ping
+    rt.send_to(addr, Count(0)).unwrap();
+    rt.send_to(addr, Count(0)).unwrap();
+    tick_n(&rt, 10);
+    let stats = rt.stats();
+    let mismatches: u64 = stats.workers.iter().map(|w| w.type_mismatches).sum();
+    assert_eq!(mismatches, 3, "3 wrong-type messages counted as mismatches");
+}
+
+/// Budget fairness: hot actor doesn't starve cold actor, budget is respected
+/// with self-sends, unlimited budget drains all.
+#[test]
+fn fairness_budget_prevents_starvation() {
+    // Hot (1000 msgs) vs cold (1 msg), budget=64
+    let rt = std_runtime(RuntimeConfig::default());
+    let hot_counter = Arc::new(AtomicUsize::new(0));
+    let cold_inbox = rt.new_inbox::<Pong>().unwrap();
+    let hot = rt.spawn(CountingPingActor { counter: hot_counter.clone() }).unwrap();
+    let cold = rt.spawn(PingPongActor).unwrap();
+    let dummy = rt.new_inbox::<Pong>().unwrap();
+    for _ in 0..1000 {
+        rt.send_to(hot, Ping { reply_to: *dummy.addr() }).unwrap();
+    }
+    rt.send_to(cold, Ping { reply_to: *cold_inbox.addr() }).unwrap();
+    rt.tick();
+    assert!(cold_inbox.try_recv().is_some(), "cold actor not starved by hot actor");
+    assert!(hot_counter.load(Ordering::SeqCst) <= 64, "hot capped at budget");
+
+    // Budget=4 with self-send chain of 20 → completes across multiple ticks
+    let rt = std_runtime(RuntimeConfig { actor_message_budget: 4, ..Default::default() });
+    let addr = rt.spawn(SelfSendActor).unwrap();
+    let inbox = rt.new_inbox::<Done>().unwrap();
+    rt.send_to(addr, Countdown { remaining: 20, reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 30);
+    assert_eq!(inbox.try_recv(), Some(Done(0)), "self-send chain completes despite budget");
+
+    // Unlimited budget (0) drains all
+    let rt = std_runtime(RuntimeConfig { actor_message_budget: 0, ..Default::default() });
+    let counter = Arc::new(AtomicUsize::new(0));
+    let dummy = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(CountingPingActor { counter: counter.clone() }).unwrap();
+    for _ in 0..500 {
+        rt.send_to(addr, Ping { reply_to: *dummy.addr() }).unwrap();
+    }
+    rt.tick();
+    rt.tick();
+    assert_eq!(counter.load(Ordering::SeqCst), 500, "unlimited budget drains all");
+}
+
+/// One-shot timers fire at the right tick and only once. Interval timers fire
+/// repeatedly at the right period. Timers are cleaned up when actors die.
+#[test]
+fn timer_one_shot_and_interval() {
+    // One-shot: delay=3 from on_start
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Ping>().unwrap();
+    rt.spawn(TimerStartActor { target: *inbox.addr(), delay_ticks: 3 }).unwrap();
+    rt.tick(); // tick 1: on_start schedules
+    assert!(inbox.try_recv().is_none(), "no delivery tick 1");
+    rt.tick(); // tick 2
+    assert!(inbox.try_recv().is_none(), "no delivery tick 2");
+    rt.tick(); // tick 3
+    assert!(inbox.try_recv().is_none(), "no delivery tick 3");
+    rt.tick(); // tick 4: fires
+    assert!(inbox.try_recv().is_some(), "timer fires after 3-tick delay");
+
+    // One-shot from handler
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Done>().unwrap();
+    let addr = rt.spawn(DelayPingPongActor).unwrap();
+    rt.send_to(addr, Forward { value: 42, reply_to: *inbox.addr() }).unwrap();
+    rt.tick(); // process Forward, schedule timer
+    assert!(inbox.try_recv().is_none());
+    rt.tick(); // tick 2
+    rt.tick(); // tick 3
+    assert!(inbox.try_recv().is_none());
+    rt.tick(); // tick 4: fires
+    assert_eq!(inbox.try_recv(), Some(Done(42)), "delayed reply from handler timer");
+
+    // One-shot does NOT repeat
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Ping>().unwrap();
+    rt.spawn(TimerStartActor { target: *inbox.addr(), delay_ticks: 1 }).unwrap();
+    rt.tick(); // schedule
+    rt.tick(); // fires
+    assert!(inbox.try_recv().is_some(), "first fire");
+    tick_n(&rt, 5);
+    assert!(inbox.try_recv().is_none(), "one-shot doesn't repeat");
+
+    // Zero-delay fires next tick
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Ping>().unwrap();
+    rt.spawn(TimerStartActor { target: *inbox.addr(), delay_ticks: 0 }).unwrap();
+    rt.tick(); // schedule
+    assert!(inbox.try_recv().is_none(), "not immediate — fires next tick");
+    rt.tick(); // fires
+    assert!(inbox.try_recv().is_some(), "zero-delay fires next tick");
+
+    // Interval: period=2, fires on ticks 3, 5, 7
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Ping>().unwrap();
+    rt.spawn(HeartbeatActor { target: *inbox.addr(), period: 2 }).unwrap();
+    rt.tick(); // tick 1: schedule
+    assert!(inbox.try_recv().is_none());
+    rt.tick(); // tick 2
+    assert!(inbox.try_recv().is_none());
+    rt.tick(); // tick 3: first fire
+    assert!(inbox.try_recv().is_some(), "fire on tick 3");
+    rt.tick(); // tick 4
+    assert!(inbox.try_recv().is_none());
+    rt.tick(); // tick 5: second fire
+    assert!(inbox.try_recv().is_some(), "fire on tick 5");
+    rt.tick(); // tick 6
+    assert!(inbox.try_recv().is_none());
+    rt.tick(); // tick 7: third fire
+    assert!(inbox.try_recv().is_some(), "fire on tick 7");
+
+    // Timer cleanup when target actor dies
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter_addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    rt.spawn(HeartbeatActor { target: counter_addr, period: 1 }).unwrap();
+    tick_n(&rt, 3);
+    rt.stop_actor(counter_addr).unwrap();
+    tick_n(&rt, 5);
+    let stats = rt.stats();
+    assert_eq!(stats.workers[0].num_actors, 1, "only heartbeat actor remains");
+}
+
+/// Bounded mailboxes: DropNewest caps at capacity, DropOldest keeps newest,
+/// unbounded delivers all, mailbox refills after processing.
+#[test]
+fn mailbox_backpressure_policies() {
+    // DropNewest: capacity=10, send 50 → only 10 delivered
+    let rt = std_runtime(RuntimeConfig {
+        default_mailbox_capacity: 10,
+        mailbox_overflow: MailboxOverflow::DropNewest,
+        ..Default::default()
+    });
+    let inbox = rt.new_inbox::<Count>().unwrap();
+    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    for _ in 0..50 {
+        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
+    }
+    tick_n(&rt, 20);
+    let mut replies = 0;
+    while inbox.try_recv().is_some() { replies += 1; }
+    assert_eq!(replies, 10, "DropNewest caps at mailbox capacity");
+    let drops: u64 = rt.stats().workers.iter().map(|w| w.messages_dropped).sum();
+    assert_eq!(drops, 40, "40 messages dropped");
+
+    // DropOldest: capacity=5, send 10 → newest 5 kept
+    let rt = std_runtime(RuntimeConfig {
+        default_mailbox_capacity: 5,
+        mailbox_overflow: MailboxOverflow::DropOldest,
+        ..Default::default()
+    });
+    let inbox = rt.new_inbox::<Done>().unwrap();
+    let addr = rt.spawn(DoubleActor).unwrap();
+    for i in 0..10 {
+        let _ = rt.send_to(addr, Forward { value: i, reply_to: *inbox.addr() });
+    }
+    tick_n(&rt, 10);
+    let mut replies = Vec::new();
+    while let Some(Done(v)) = inbox.try_recv() { replies.push(v); }
+    assert_eq!(replies.len(), 5, "only 5 kept");
+    assert_eq!(replies, vec![10, 12, 14, 16, 18], "newest values kept (5-9 doubled)");
+
+    // Unbounded: 200 messages all delivered
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Count>().unwrap();
+    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    for _ in 0..200 {
+        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
+    }
+    tick_n(&rt, 50);
+    let mut count = 0;
+    while inbox.try_recv().is_some() { count += 1; }
+    assert_eq!(count, 200, "unbounded delivers all");
+
+    // Refill after processing
+    let rt = std_runtime(RuntimeConfig {
+        default_mailbox_capacity: 5,
+        actor_message_budget: 5,
+        mailbox_overflow: MailboxOverflow::DropNewest,
+        ..Default::default()
+    });
+    let inbox = rt.new_inbox::<Count>().unwrap();
+    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
+    for _ in 0..5 {
+        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
+    }
+    rt.tick(); // process batch 1
+    for _ in 0..5 {
+        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
+    }
+    rt.tick(); // process batch 2
+    let mut count = 0;
+    while inbox.try_recv().is_some() { count += 1; }
+    assert_eq!(count, 10, "mailbox refills after draining");
+}
--- a/tests/runtime_hasher.rs
+++ b/tests/runtime_hasher.rs
@ -1,224 +0,0 @@
-mod common;
-use common::*;
-
-use std::sync::atomic::{AtomicUsize, Ordering};
-use std::sync::Arc;
-
-// ── Identity Hasher Correctness ────────────────────────────────────────────
-
-/// Given: 200 actors each expecting a unique numbered message
-/// When:  Each actor receives its number and replies with (self_addr, number)
-/// Then:  All 200 replies match -- no message was misrouted by the identity hasher
-#[test]
-fn many_actors_all_receive_correct_messages() {
-    #[derive(Clone)]
-    struct NumberedMsg {
-        n: usize,
-        reply_to: ActorAddress,
-    }
-
-    #[derive(Clone, Debug, PartialEq)]
-    struct NumberedReply {
-        from: ActorAddress,
-        n: usize,
-    }
-
-    struct NumberedActor;
-
-    impl ActorInterface for NumberedActor {
-        type Incoming = NumberedMsg;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, msg: NumberedMsg) {
-            let _ = ctx.send(
-                msg.reply_to,
-                NumberedReply {
-                    from: ctx.self_addr(),
-                    n: msg.n,
-                },
-            );
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig {
-        max_actors: 300,
-        channel_buffer_size: 1024,
-        num_threads: 1,
-        ..Default::default()
-    });
-
-    let inbox = rt.new_inbox::<NumberedReply>().unwrap();
-    let inbox_addr = *inbox.addr();
-
-    // Spawn 200 actors
-    let mut addrs = Vec::new();
-    for _ in 0..200 {
-        addrs.push(rt.spawn(NumberedActor).unwrap());
-    }
-    rt.tick(); // on_start
-
-    // Send unique numbered message to each
-    for (i, addr) in addrs.iter().enumerate() {
-        rt.send_to(
-            *addr,
-            NumberedMsg {
-                n: i,
-                reply_to: inbox_addr,
-            },
-        )
-        .unwrap();
-    }
-    rt.tick(); // process + reply
-    rt.tick(); // deliver replies
-
-    // Verify all 200 replies
-    let mut replies: Vec<NumberedReply> = Vec::new();
-    while let Some(reply) = inbox.try_recv() {
-        replies.push(reply);
-    }
-
-    assert_eq!(replies.len(), 200, "should receive exactly 200 replies");
-
-    // Verify each reply came from the correct actor with the correct number
-    for (i, addr) in addrs.iter().enumerate() {
-        let reply = replies.iter().find(|r| r.n == i);
-        assert!(
-            reply.is_some(),
-            "missing reply for actor #{i}"
-        );
-        assert_eq!(
-            reply.unwrap().from, *addr,
-            "reply #{i} came from wrong actor"
-        );
-    }
-}
-
-/// Given: A 100-actor ring where each actor forwards to the next
-/// When:  A message enters the ring and traverses all 100 hops
-/// Then:  The message completes the full circuit (address_map lookups all correct)
-#[test]
-fn ring_routing_unchanged_after_hasher_optimization() {
-    #[derive(Clone)]
-    struct RingHop {
-        hops_remaining: usize,
-        final_dest: ActorAddress,
-    }
-
-    #[derive(Clone, Debug, PartialEq)]
-    struct RingDone(usize); // total hops completed
-
-    struct RingNode {
-        next: ActorAddress,
-    }
-
-    impl ActorInterface for RingNode {
-        type Incoming = RingHop;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, msg: RingHop) {
-            if msg.hops_remaining == 0 {
-                let _ = ctx.send(msg.final_dest, RingDone(100));
-            } else {
-                let _ = ctx.send(
-                    self.next,
-                    RingHop {
-                        hops_remaining: msg.hops_remaining - 1,
-                        final_dest: msg.final_dest,
-                    },
-                );
-            }
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig {
-        max_actors: 200,
-        channel_buffer_size: 1024,
-        num_threads: 1,
-        ..Default::default()
-    });
-
-    let inbox = rt.new_inbox::<RingDone>().unwrap();
-    let inbox_addr = *inbox.addr();
-
-    // Build chain backwards: last node sends to inbox, first node receives
-    let mut addrs = Vec::new();
-    let mut next = inbox_addr;
-    for _ in (0..100).rev() {
-        let node = RingNode { next };
-        let addr = rt.spawn(node).unwrap();
-        addrs.push(addr);
-        next = addr;
-    }
-    addrs.reverse(); // addrs[0] is start of chain
-
-    rt.tick(); // on_start
-
-    // Inject message at the start
-    rt.send_to(
-        addrs[0],
-        RingHop {
-            hops_remaining: 99,
-            final_dest: inbox_addr,
-        },
-    )
-    .unwrap();
-
-    // Tick enough times for the message to traverse all 100 actors
-    for _ in 0..110 {
-        rt.tick();
-    }
-
-    let result = inbox.try_recv();
-    assert!(result.is_some(), "ring message should complete all 100 hops");
-    assert_eq!(result.unwrap(), RingDone(100));
-}
-
-/// Given: An actor that calls ctx.stop_self() upon receiving a trigger message
-/// When:  The trigger is sent, then 5 more messages are sent, then ticked
-/// Then:  The actor is removed, only messages before stop are processed
-#[test]
-fn stop_self_with_pending_messages_still_works() {
-    let processed = Arc::new(AtomicUsize::new(0));
-
-    #[derive(Clone)]
-    struct Msg(bool); // true = trigger stop
-
-    struct StopOnTrigger(Arc<AtomicUsize>);
-
-    impl ActorInterface for StopOnTrigger {
-        type Incoming = Msg;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, msg: Msg) {
-            self.0.fetch_add(1, Ordering::Relaxed);
-            if msg.0 {
-                ctx.stop_self();
-            }
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let p = processed.clone();
-    let addr = rt.spawn(StopOnTrigger(p)).unwrap();
-    rt.tick(); // on_start
-
-    // Send: 2 normal, 1 trigger, 5 more normal
-    rt.send_to(addr, Msg(false)).unwrap();
-    rt.send_to(addr, Msg(false)).unwrap();
-    rt.send_to(addr, Msg(true)).unwrap(); // stop trigger
-    rt.send_to(addr, Msg(false)).unwrap();
-    rt.send_to(addr, Msg(false)).unwrap();
-    rt.send_to(addr, Msg(false)).unwrap();
-    rt.send_to(addr, Msg(false)).unwrap();
-    rt.send_to(addr, Msg(false)).unwrap();
-
-    rt.tick(); // process messages -- stops after trigger
-    rt.tick(); // cleanup
-
-    // Only 3 messages should be processed (2 normal + 1 trigger)
-    assert_eq!(
-        processed.load(Ordering::Relaxed),
-        3,
-        "should process exactly the messages up to and including the stop trigger"
-    );
-
-    // Subsequent sends should fail
-    assert!(rt.send_to(addr, Msg(false)).is_err());
-}
--- a/tests/runtime_hooks.rs
+++ b/tests/runtime_hooks.rs
@ -1,454 +0,0 @@
-mod common;
-use common::*;
-
-use std::sync::atomic::{AtomicUsize, Ordering};
-use std::sync::Arc;
-
-// ── Lifecycle Hook Helpers ────────────────────────────────────────────────
-
-/// An actor that records lifecycle events to shared counters.
-struct LifecycleActor {
-    started: Arc<AtomicUsize>,
-    stopped: Arc<AtomicUsize>,
-    handled: Arc<AtomicUsize>,
-}
-
-impl ActorInterface for LifecycleActor {
-    type Incoming = Ping;
-    type Response = Pong;
-
-    fn on_start(&mut self, _ctx: &Ctx) {
-        self.started.fetch_add(1, Ordering::Relaxed);
-    }
-
-    fn on_stop(&mut self, _ctx: &Ctx) {
-        self.stopped.fetch_add(1, Ordering::Relaxed);
-    }
-
-    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
-        self.handled.fetch_add(1, Ordering::Relaxed);
-        let _ = ctx.send(msg.reply_to, Pong);
-    }
-}
-
-/// An actor that stops itself after processing N messages.
-struct SelfStopActor {
-    count: usize,
-    stop_after: usize,
-    stopped: Arc<AtomicUsize>,
-}
-
-impl ActorInterface for SelfStopActor {
-    type Incoming = Forward;
-    type Response = Done;
-
-    fn on_stop(&mut self, _ctx: &Ctx) {
-        self.stopped.fetch_add(1, Ordering::Relaxed);
-    }
-
-    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
-        self.count += 1;
-        let _ = ctx.send(msg.reply_to, Done(msg.value));
-        if self.count >= self.stop_after {
-            ctx.stop_self();
-        }
-    }
-}
-
-/// An actor that sends a farewell message in on_stop.
-struct FarewellActor {
-    farewell_to: ActorAddress,
-}
-
-impl ActorInterface for FarewellActor {
-    type Incoming = Ping;
-    type Response = Pong;
-
-    fn on_stop(&mut self, ctx: &Ctx) {
-        let _ = ctx.send(self.farewell_to, Pong);
-    }
-
-    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
-        let _ = ctx.send(msg.reply_to, Pong);
-    }
-}
-
-/// An actor whose on_start panics.
-struct PanicOnStartActor {
-    handled: Arc<AtomicUsize>,
-}
-
-impl ActorInterface for PanicOnStartActor {
-    type Incoming = Ping;
-    type Response = Pong;
-
-    fn on_start(&mut self, _ctx: &Ctx) {
-        panic!("on_start panic");
-    }
-
-    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {
-        self.handled.fetch_add(1, Ordering::Relaxed);
-    }
-}
-
-/// Handles Forward messages, replies Done(value * 2), panics on the panic_at-th message.
-/// (Needed for on_start_called_again_after_restart and stop_vs_panic tests.)
-struct RestartTestActor {
-    count: usize,
-    panic_at: usize,
-}
-
-impl ActorInterface for RestartTestActor {
-    type Incoming = Forward;
-    type Response = Done;
-    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
-        self.count += 1;
-        if self.count >= self.panic_at {
-            panic!("intentional panic at message {}", self.count);
-        }
-        let _ = ctx.send(msg.reply_to, Done(msg.value * 2));
-    }
-}
-
-// ── Lifecycle Hook Tests ──────────────────────────────────────────────────
-
-/// Given an actor with on_start implemented,
-/// when it is spawned and the runtime ticks,
-/// then on_start is called exactly once before the first message.
-#[test]
-fn on_start_called_before_first_message() {
-    let started = Arc::new(AtomicUsize::new(0));
-    let stopped = Arc::new(AtomicUsize::new(0));
-    let handled = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let addr = rt.spawn(LifecycleActor {
-        started: started.clone(),
-        stopped: stopped.clone(),
-        handled: handled.clone(),
-    }).unwrap();
-
-    // First tick — should call on_start
-    rt.tick();
-    assert_eq!(started.load(Ordering::Relaxed), 1, "on_start called on first tick");
-    assert_eq!(handled.load(Ordering::Relaxed), 0, "no messages processed yet");
-
-    // Send messages and tick more
-    let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
-    rt.tick();
-    assert_eq!(started.load(Ordering::Relaxed), 1, "on_start not called again");
-    assert_eq!(handled.load(Ordering::Relaxed), 1, "message processed after on_start");
-}
-
-/// Given an actor with on_start,
-/// when multiple actors are spawned,
-/// then each gets its own on_start call exactly once.
-#[test]
-fn on_start_called_per_actor() {
-    let started = Arc::new(AtomicUsize::new(0));
-    let stopped = Arc::new(AtomicUsize::new(0));
-    let handled = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-
-    for _ in 0..5 {
-        let _ = rt.spawn(LifecycleActor {
-            started: started.clone(),
-            stopped: stopped.clone(),
-            handled: handled.clone(),
-        }).unwrap();
-    }
-
-    rt.tick();
-    assert_eq!(started.load(Ordering::Relaxed), 5, "on_start called for each of 5 actors");
-
-    // Subsequent ticks don't repeat on_start
-    rt.tick();
-    rt.tick();
-    assert_eq!(started.load(Ordering::Relaxed), 5, "on_start still 5 after more ticks");
-}
-
-/// Given an actor whose on_start panics,
-/// when it is spawned and the runtime ticks,
-/// then it is immediately poisoned and never processes messages.
-#[test]
-fn on_start_panic_poisons_actor() {
-    let handled = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let addr = rt.spawn(PanicOnStartActor { handled: handled.clone() }).unwrap();
-
-    let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
-    for _ in 0..5 { rt.tick(); }
-
-    assert_eq!(handled.load(Ordering::Relaxed), 0, "actor never processed messages");
-
-    let stats = rt.stats();
-    let total_panics: u64 = stats.workers.iter().map(|w| w.panics).sum();
-    assert_eq!(total_panics, 1, "on_start panic counted");
-}
-
-// ── Graceful Stop Tests ───────────────────────────────────────────────────
-
-/// Given an actor that calls ctx.stop_self() after 3 messages,
-/// when 5 messages are sent,
-/// then only 3 are processed, the actor is removed, and on_stop is called.
-#[test]
-fn actor_can_stop_self() {
-    let stopped = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Done>().unwrap();
-
-    let addr = rt.spawn(SelfStopActor {
-        count: 0,
-        stop_after: 3,
-        stopped: stopped.clone(),
-    }).unwrap();
-
-    for i in 0..5 {
-        let _ = rt.send_to(addr, Forward { value: i, reply_to: *inbox.addr() });
-    }
-    for _ in 0..10 { rt.tick(); }
-
-    // Only 3 messages should be processed (stop_self after 3rd)
-    let mut replies = Vec::new();
-    while let Some(Done(v)) = inbox.try_recv() {
-        replies.push(v);
-    }
-    assert_eq!(replies.len(), 3, "only 3 messages processed before stop");
-    assert!(replies.contains(&0));
-    assert!(replies.contains(&1));
-    assert!(replies.contains(&2));
-
-    assert_eq!(stopped.load(Ordering::Relaxed), 1, "on_stop called exactly once");
-
-    // Actor should be removed from address map
-    let stats = rt.stats();
-    assert_eq!(stats.actors.len(), 0, "stopped actor removed from address map");
-}
-
-/// Given a running actor,
-/// when runtime.stop_actor(addr) is called,
-/// then the actor stops, on_stop is called, and it's removed from the pool.
-#[test]
-fn runtime_can_stop_actor() {
-    let started = Arc::new(AtomicUsize::new(0));
-    let stopped = Arc::new(AtomicUsize::new(0));
-    let handled = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let addr = rt.spawn(LifecycleActor {
-        started: started.clone(),
-        stopped: stopped.clone(),
-        handled: handled.clone(),
-    }).unwrap();
-
-    // Let it start and process a message
-    let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
-    for _ in 0..3 { rt.tick(); }
-    assert_eq!(handled.load(Ordering::Relaxed), 1);
-
-    // Stop it externally
-    rt.stop_actor(addr).unwrap();
-    for _ in 0..3 { rt.tick(); }
-
-    assert_eq!(stopped.load(Ordering::Relaxed), 1, "on_stop called");
-
-    // Actor should be gone
-    let stats = rt.stats();
-    assert_eq!(stats.actors.len(), 0, "stopped actor removed");
-    assert_eq!(stats.workers[0].num_actors, 0);
-}
-
-/// Given a stopped actor,
-/// when new messages are sent to it,
-/// then sends return Err (address not found).
-#[test]
-fn send_to_stopped_actor_returns_error() {
-    let stopped = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Done>().unwrap();
-
-    let addr = rt.spawn(SelfStopActor {
-        count: 0,
-        stop_after: 1,
-        stopped: stopped.clone(),
-    }).unwrap();
-
-    // One message triggers stop
-    let _ = rt.send_to(addr, Forward { value: 1, reply_to: *inbox.addr() });
-    for _ in 0..10 { rt.tick(); }
-
-    // Actor is now removed — send should fail
-    let result = rt.send_to(addr, Forward { value: 2, reply_to: *inbox.addr() });
-    assert!(result.is_err(), "send to stopped actor should return Err");
-}
-
-/// Given a gracefully stopped actor and a panicked actor,
-/// then stats.stops and stats.panics track them separately.
-#[test]
-fn stop_vs_panic_tracked_separately_in_stats() {
-    let stopped = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Done>().unwrap();
-
-    // Actor that stops itself after 1 message
-    let _stop_addr = rt.spawn(SelfStopActor {
-        count: 0,
-        stop_after: 1,
-        stopped: stopped.clone(),
-    }).unwrap();
-
-    // Actor that panics on first message
-    let panic_addr = rt.spawn(RestartTestActor { count: 0, panic_at: 1 }).unwrap();
-
-    let _ = rt.send_to(_stop_addr, Forward { value: 1, reply_to: *inbox.addr() });
-    let _ = rt.send_to(panic_addr, Forward { value: 1, reply_to: *inbox.addr() });
-    for _ in 0..10 { rt.tick(); }
-
-    let stats = rt.stats();
-    let total_stops: u64 = stats.workers.iter().map(|w| w.stops).sum();
-    let total_panics: u64 = stats.workers.iter().map(|w| w.panics).sum();
-
-    assert_eq!(total_stops, 1, "one graceful stop");
-    assert_eq!(total_panics, 1, "one panic");
-}
-
-/// Given an actor with on_stop that sends a farewell message,
-/// when the actor is stopped,
-/// then the farewell message is delivered.
-#[test]
-fn on_stop_can_send_messages() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let addr = rt.spawn(FarewellActor {
-        farewell_to: *inbox.addr(),
-    }).unwrap();
-
-    // Let it start
-    rt.tick();
-
-    // Stop it
-    rt.stop_actor(addr).unwrap();
-    for _ in 0..5 { rt.tick(); }
-
-    // Should receive farewell Pong from on_stop
-    let farewell = inbox.try_recv();
-    assert_eq!(farewell, Some(Pong), "farewell message delivered from on_stop");
-}
-
-/// Given a supervisor with a child that panics and is restarted,
-/// when the child is respawned by the supervisor,
-/// then on_start is called again on the fresh instance.
-#[test]
-fn on_start_called_again_after_restart() {
-    let started = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let started_c = started.clone();
-    let _sup_addr = rt.spawn(Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![ChildSpec::new("child", RestartPolicy::Permanent, move |ctx| {
-            ctx.spawn(LifecycleActor {
-                started: started_c.clone(),
-                stopped: Arc::new(AtomicUsize::new(0)),
-                handled: Arc::new(AtomicUsize::new(0)),
-            })
-        })],
-    )).unwrap();
-
-    // First tick: supervisor starts, spawns child, on_start called
-    for _ in 0..3 { rt.tick(); }
-    assert_eq!(started.load(Ordering::Relaxed), 1, "on_start called once");
-}
-
-/// Given an actor stopped via stop_actor() with messages already queued,
-/// when the stop signal arrives after the queued messages (PoisonPill semantics),
-/// then messages ahead of the signal are processed, then the actor stops.
-#[test]
-fn external_stop_is_queued_after_pending_messages() {
-    let stopped = Arc::new(AtomicUsize::new(0));
-    let handled = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let started = Arc::new(AtomicUsize::new(0));
-    let addr = rt.spawn(LifecycleActor {
-        started: started.clone(),
-        stopped: stopped.clone(),
-        handled: handled.clone(),
-    }).unwrap();
-
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    // Queue 10 messages, then stop — StopSignal is queued AFTER the 10
-    for _ in 0..10 {
-        let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
-    }
-    rt.stop_actor(addr).unwrap();
-    for _ in 0..10 { rt.tick(); }
-
-    // All 10 messages processed (they were ahead of StopSignal in the queue)
-    let total_handled = handled.load(Ordering::Relaxed);
-    assert_eq!(total_handled, 10, "all messages processed before stop signal");
-    assert_eq!(stopped.load(Ordering::Relaxed), 1, "on_stop called");
-
-    // Actor is removed
-    let stats = rt.stats();
-    assert_eq!(stats.actors.len(), 0, "stopped actor removed");
-}
-
-/// Given a running actor with no pending messages,
-/// when stop_actor() is called and then new messages are sent,
-/// then the stop takes priority and new messages are not processed.
-#[test]
-fn external_stop_before_new_messages_prevents_processing() {
-    let stopped = Arc::new(AtomicUsize::new(0));
-    let handled = Arc::new(AtomicUsize::new(0));
-    let started = Arc::new(AtomicUsize::new(0));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let addr = rt.spawn(LifecycleActor {
-        started: started.clone(),
-        stopped: stopped.clone(),
-        handled: handled.clone(),
-    }).unwrap();
-
-    // Let actor start
-    rt.tick();
-
-    // Stop first, then send messages
-    rt.stop_actor(addr).unwrap();
-    for _ in 0..5 {
-        let _ = rt.send_to(addr, Ping { reply_to: *inbox.addr() });
-    }
-    for _ in 0..10 { rt.tick(); }
-
-    // Stop signal was first in queue, so no messages processed
-    assert_eq!(handled.load(Ordering::Relaxed), 0, "no messages processed after stop");
-    assert_eq!(stopped.load(Ordering::Relaxed), 1, "on_stop called");
-}
-
-/// Given stop_actor is called on a nonexistent address,
-/// then it returns Err.
-#[test]
-fn stop_nonexistent_actor_returns_error() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let fake_addr = swactor::actor::ActorAddress::default();
-    let result = rt.stop_actor(fake_addr);
-    assert!(result.is_err(), "stop_actor on nonexistent address should return Err");
-}
--- a/tests/runtime_lifecycle.rs
+++ b/tests/runtime_lifecycle.rs
--- a/tests/runtime_mechanics.rs
+++ b/tests/runtime_mechanics.rs
@ -1,362 +0,0 @@
-mod common;
-use common::*;
-
-// ── Load-Aware Placement Tests ─────────────────────────────────────────────
-
-/// Given a multi-threaded runtime where one worker has many more actors,
-/// when new actors are spawned after a few ticks (so stats propagate),
-/// then they should be placed on the lighter worker.
-#[test]
-fn load_aware_placement_prefers_lighter_worker() {
-    // 2 threads: intentionally imbalance by spawning many actors first
-    let rt = std_runtime(RuntimeConfig {
-        num_threads: 2,
-        ..Default::default()
-    });
-
-    // Phase 1: Spawn 20 actors. With round-robin, they split ~10/10.
-    let mut addrs = Vec::new();
-    for _ in 0..20 {
-        addrs.push(rt.spawn(CounterActor { count: 0 }).unwrap());
-    }
-
-    // Run so stats propagate, then bombard worker 0's actors with messages
-    // to create mailbox depth imbalance.
-    let handle = rt.run().unwrap();
-    std::thread::sleep(std::time::Duration::from_millis(10));
-
-    // Send 500 messages to the first 10 actors (likely on worker 0).
-    for addr in &addrs[..10] {
-        for _ in 0..50 {
-            let _ = handle.runtime.send_to(*addr, Increment {
-                reply_to: *addr, // self-reply to keep mailbox depth up
-            });
-        }
-    }
-
-    std::thread::sleep(std::time::Duration::from_millis(20));
-
-    // Phase 2: Spawn 10 more actors. With load-aware placement,
-    // they should bias toward the lighter worker.
-    let mut late_addrs = Vec::new();
-    for _ in 0..10 {
-        late_addrs.push(handle.runtime.spawn(CounterActor { count: 0 }).unwrap());
-    }
-
-    std::thread::sleep(std::time::Duration::from_millis(20));
-
-    let stats = handle.runtime.stats();
-    handle.shutdown();
-    handle.join();
-
-    // Verify the system is operational — both workers should have actors
-    let total_actors: usize = stats.workers.iter().map(|w| w.num_actors).sum();
-    assert!(total_actors >= 20, "expected at least 20 actors, got {}", total_actors);
-
-    // The lighter worker should have gotten more of the late actors.
-    assert!(
-        stats.workers.iter().all(|w| w.num_actors > 0),
-        "both workers should have actors, got {:?}",
-        stats.workers.iter().map(|w| w.num_actors).collect::<Vec<_>>()
-    );
-}
-
-/// Given a single-threaded runtime (1 worker),
-/// when many actors are spawned,
-/// then all go to worker 0 regardless of load (no panic, no error).
-#[test]
-fn load_aware_placement_single_worker_degrades_gracefully() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    for _ in 0..50 {
-        rt.spawn(CounterActor { count: 0 }).unwrap();
-    }
-
-    // Tick several times to let stats update
-    for _ in 0..10 {
-        rt.tick();
-    }
-
-    let stats = rt.stats();
-    assert_eq!(stats.workers.len(), 1);
-    assert_eq!(stats.workers[0].num_actors, 50);
-}
-
-/// Given a fresh runtime with no prior ticks,
-/// when actors are spawned in a burst,
-/// then they distribute evenly (round-robin fallback when stats are all zero).
-#[test]
-fn load_aware_placement_falls_back_to_round_robin_on_fresh_runtime() {
-    let rt = std_runtime(RuntimeConfig {
-        num_threads: 4,
-        ..Default::default()
-    });
-
-    // Spawn 100 actors before any ticks (all stats are zero)
-    for _ in 0..100 {
-        rt.spawn(CounterActor { count: 0 }).unwrap();
-    }
-
-    let handle = rt.run().unwrap();
-    std::thread::sleep(std::time::Duration::from_millis(20));
-
-    let stats = handle.runtime.stats();
-    handle.shutdown();
-    handle.join();
-
-    // With 4 workers and 100 actors, each should have ~25 (+-5).
-    for w in &stats.workers {
-        assert!(
-            w.num_actors >= 20 && w.num_actors <= 30,
-            "worker {} has {} actors, expected ~25 (round-robin)",
-            w.id, w.num_actors
-        );
-    }
-}
-
-// ── Mailbox Backpressure Tests ─────────────────────────────────────────────
-
-/// Given a runtime with bounded mailboxes (capacity=10, DropNewest),
-/// when 50 messages are sent to an actor before any ticks,
-/// then only the first 10 are delivered and the rest are dropped.
-#[test]
-fn bounded_mailbox_drop_newest_caps_at_capacity() {
-    let rt = std_runtime(RuntimeConfig {
-        default_mailbox_capacity: 10,
-        mailbox_overflow: MailboxOverflow::DropNewest,
-        ..Default::default()
-    });
-
-    let inbox = rt.new_inbox::<Count>().unwrap();
-    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
-
-    // Send 50 messages — only first 10 should be queued
-    for _ in 0..50 {
-        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
-    }
-
-    // Tick enough times to process all queued messages
-    for _ in 0..20 {
-        rt.tick();
-    }
-
-    // Count replies — should be exactly 10 (the mailbox capacity)
-    let mut replies = 0;
-    while inbox.try_recv().is_some() {
-        replies += 1;
-    }
-    assert_eq!(replies, 10, "should deliver exactly mailbox_capacity messages");
-
-    // Stats should show drops
-    let stats = rt.stats();
-    let total_drops: u64 = stats.workers.iter().map(|w| w.messages_dropped).sum();
-    assert_eq!(total_drops, 40, "40 messages should have been dropped");
-}
-
-/// Given a runtime with bounded mailboxes (capacity=5, DropOldest),
-/// when 10 messages are sent before any tick,
-/// then only the 5 most recent messages are delivered.
-#[test]
-fn bounded_mailbox_drop_oldest_keeps_newest() {
-    let rt = std_runtime(RuntimeConfig {
-        default_mailbox_capacity: 5,
-        mailbox_overflow: MailboxOverflow::DropOldest,
-        ..Default::default()
-    });
-
-    let inbox = rt.new_inbox::<Done>().unwrap();
-    let addr = rt.spawn(DoubleActor).unwrap();
-
-    // Send messages with values 0..10. DoubleActor replies Done(value * 2).
-    for i in 0..10 {
-        let _ = rt.send_to(addr, Forward {
-            value: i,
-            reply_to: *inbox.addr(),
-        });
-    }
-
-    for _ in 0..10 {
-        rt.tick();
-    }
-
-    // Collect all replies
-    let mut replies = Vec::new();
-    while let Some(Done(v)) = inbox.try_recv() {
-        replies.push(v);
-    }
-
-    assert_eq!(replies.len(), 5, "should deliver exactly 5 messages");
-    // The 5 most recent: values 5,6,7,8,9 -> doubled: 10,12,14,16,18
-    assert_eq!(replies, vec![10, 12, 14, 16, 18], "should keep the newest messages");
-}
-
-/// Given a runtime with unbounded mailboxes (capacity=0, the default),
-/// when many messages are sent,
-/// then all are delivered (backward compatibility).
-#[test]
-fn unbounded_mailbox_delivers_all_messages() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let inbox = rt.new_inbox::<Count>().unwrap();
-    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
-
-    for _ in 0..200 {
-        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
-    }
-
-    for _ in 0..50 {
-        rt.tick();
-    }
-
-    let mut replies = 0;
-    while inbox.try_recv().is_some() {
-        replies += 1;
-    }
-    assert_eq!(replies, 200, "all 200 messages should be delivered with unbounded mailbox");
-
-    let stats = rt.stats();
-    let total_drops: u64 = stats.workers.iter().map(|w| w.messages_dropped).sum();
-    assert_eq!(total_drops, 0, "no drops with unbounded mailbox");
-}
-
-/// Given bounded mailboxes with budget, when an actor processes messages
-/// and frees mailbox space, then new messages should be accepted on subsequent ticks.
-#[test]
-fn bounded_mailbox_refills_after_processing() {
-    let rt = std_runtime(RuntimeConfig {
-        default_mailbox_capacity: 5,
-        actor_message_budget: 5,
-        mailbox_overflow: MailboxOverflow::DropNewest,
-        ..Default::default()
-    });
-
-    let inbox = rt.new_inbox::<Count>().unwrap();
-    let addr = rt.spawn(CounterActor { count: 0 }).unwrap();
-
-    // Send first batch of 5 — fills mailbox exactly
-    for _ in 0..5 {
-        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
-    }
-
-    // Tick to process all 5 (budget=5, capacity=5)
-    rt.tick();
-
-    // Send second batch of 5 — mailbox is empty, so all 5 should be accepted
-    for _ in 0..5 {
-        let _ = rt.send_to(addr, Increment { reply_to: *inbox.addr() });
-    }
-
-    rt.tick();
-
-    let mut replies = 0;
-    while inbox.try_recv().is_some() {
-        replies += 1;
-    }
-    assert_eq!(replies, 10, "all 10 messages across 2 batches should be processed");
-
-    let stats = rt.stats();
-    let total_drops: u64 = stats.workers.iter().map(|w| w.messages_dropped).sum();
-    assert_eq!(total_drops, 0, "no drops when mailbox drains between batches");
-}
-
-// ── Dead Actor Cleanup Tests ───────────────────────────────────────────────
-
-/// Given an actor that panics and is poisoned,
-/// when ticks continue,
-/// then the actor is removed from stats and sends to its address fail.
-#[test]
-fn dead_actor_cleaned_up_from_stats_and_address_map() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let good = rt.spawn(PingPongActor).unwrap();
-    let bad = rt.spawn(PanicActor).unwrap();
-
-    // Trigger panic
-    let _ = rt.send_to(bad, PanicMsg);
-    for _ in 0..5 { rt.tick(); }
-
-    let stats = rt.stats();
-    // Good actor still present, bad actor cleaned up
-    assert_eq!(stats.workers[0].num_actors, 1, "only the healthy actor should remain");
-    assert!(
-        stats.actors.iter().any(|(a, _)| *a == good),
-        "good actor should be in address map"
-    );
-    assert!(
-        !stats.actors.iter().any(|(a, _)| *a == bad),
-        "poisoned actor should be removed from address map"
-    );
-
-    // Sends to cleaned-up actor fail
-    let result = rt.send_to(bad, PanicMsg);
-    assert!(result.is_err(), "send to cleaned-up actor should fail");
-}
-
-/// Given many actors that all panic,
-/// when ticks proceed,
-/// then all are cleaned up and stats reflect zero actors.
-#[test]
-fn bulk_dead_actor_cleanup() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let mut addrs = Vec::new();
-    for _ in 0..20 {
-        addrs.push(rt.spawn(PanicActor).unwrap());
-    }
-
-    // Trigger all panics
-    for &addr in &addrs {
-        let _ = rt.send_to(addr, PanicMsg);
-    }
-    for _ in 0..10 { rt.tick(); }
-
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 0, "all poisoned actors should be cleaned up");
-    assert_eq!(
-        stats.actors.len(), 0,
-        "address map should be empty after all actors poisoned"
-    );
-}
-
-// ── Actor Recovery Helpers ──────────────────────────────────────────────────
-
-/// Handles Forward messages, replies Done(value * 2), panics on the panic_at-th message.
-struct RestartTestActor {
-    count: usize,
-    panic_at: usize,
-}
-
-impl ActorInterface for RestartTestActor {
-    type Incoming = Forward;
-    type Response = Done;
-    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
-        self.count += 1;
-        if self.count >= self.panic_at {
-            panic!("intentional panic at message {}", self.count);
-        }
-        let _ = ctx.send(msg.reply_to, Done(msg.value * 2));
-    }
-}
-
-// ── Actor Recovery Tests ───────────────────────────────────────────────────
-
-/// Given an actor that panics,
-/// when it panics,
-/// then it is poisoned and future messages are discarded.
-#[test]
-fn non_restartable_actor_still_poisons_on_panic() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Done>().unwrap();
-
-    // Normal spawn — not restartable
-    let addr = rt.spawn(RestartTestActor { count: 0, panic_at: 1 }).unwrap();
-
-    let _ = rt.send_to(addr, Forward { value: 42, reply_to: *inbox.addr() });
-    for _ in 0..5 { rt.tick(); }
-
-    let stats = rt.stats();
-    let total_panics: u64 = stats.workers.iter().map(|w| w.panics).sum();
-    let total_restarts: u64 = stats.workers.iter().map(|w| w.restarts).sum();
-    assert_eq!(total_panics, 1, "should panic");
-    assert_eq!(total_restarts, 0, "should not restart (not restartable)");
-}
--- a/tests/runtime_registry.rs
+++ b/tests/runtime_registry.rs
@ -1,758 +0,0 @@
-mod common;
-use common::*;
-
-// ── Named Actor Registry ────────────────────────────────────────────────────
-
-/// Given a named actor is spawned,
-/// when I look it up by name,
-/// then I get the same address that spawn returned.
-#[test]
-fn named_actor_lookup_returns_spawn_address() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let addr = rt.spawn_named("greeter", PingPongActor).unwrap();
-    assert_eq!(rt.where_is("greeter"), Some(addr));
-}
-
-/// Given a named actor exists,
-/// when I send a message to the looked-up address,
-/// then the actor receives and processes it.
-#[test]
-fn named_actor_receives_messages_via_lookup() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    let addr = rt.spawn_named("ponger", PingPongActor).unwrap();
-    assert_eq!(rt.where_is("ponger"), Some(addr));
-
-    rt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick();
-    assert!(inbox.try_recv().is_some(), "named actor should process message");
-}
-
-/// Given a name is already registered,
-/// when I try to spawn another actor with the same name,
-/// then I get an error and the original binding is preserved.
-#[test]
-fn duplicate_name_returns_error() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let first_addr = rt.spawn_named("singleton", PingPongActor).unwrap();
-    let result = rt.spawn_named("singleton", PingPongActor);
-    assert!(result.is_err(), "duplicate name should fail");
-    assert_eq!(rt.where_is("singleton"), Some(first_addr), "original binding preserved");
-}
-
-/// Given no actors are registered,
-/// when I look up a nonexistent name,
-/// then I get None.
-#[test]
-fn where_is_returns_none_for_unknown_name() {
-    let rt = std_runtime(RuntimeConfig::default());
-    assert_eq!(rt.where_is("ghost"), None);
-}
-
-/// Given a named actor is stopped,
-/// when the next tick runs cleanup,
-/// then the name is automatically unregistered.
-#[test]
-fn name_auto_unregistered_on_actor_death() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let addr = rt.spawn_named("ephemeral", PingPongActor).unwrap();
-    rt.tick(); // on_start
-
-    rt.stop_actor(addr).unwrap();
-    rt.tick(); // process StopSignal + cleanup
-
-    assert_eq!(rt.where_is("ephemeral"), None, "name should be freed after stop");
-}
-
-/// Given a named actor died and its name was freed,
-/// when I spawn a new actor with the same name,
-/// then registration succeeds with a new address.
-#[test]
-fn name_can_be_reused_after_actor_death() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let first = rt.spawn_named("worker", PingPongActor).unwrap();
-    rt.tick();
-    rt.stop_actor(first).unwrap();
-    rt.tick(); // cleanup frees the name
-
-    let second = rt.spawn_named("worker", PingPongActor).unwrap();
-    assert_ne!(first, second, "new actor should have a different address");
-    assert_eq!(rt.where_is("worker"), Some(second));
-}
-
-/// Given a named actor panics (and is not restartable),
-/// when the next tick runs cleanup,
-/// then the name is freed.
-#[test]
-fn name_auto_unregistered_on_panic() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    let _addr = rt.spawn_named("fragile", PanicActor).unwrap();
-    rt.tick(); // on_start
-
-    rt.send_to(_addr, PanicMsg).unwrap();
-    rt.tick(); // panic -> poison -> cleanup
-
-    assert_eq!(rt.where_is("fragile"), None, "name freed after panic");
-    // Can reuse the name
-    let _new = rt.spawn_named("fragile", PingPongActor).unwrap();
-    assert!(rt.where_is("fragile").is_some());
-    drop(inbox);
-}
-
-/// Given multiple named actors are registered,
-/// when I call registered_names(),
-/// then all names are returned.
-#[test]
-fn registered_names_lists_all() {
-    let rt = std_runtime(RuntimeConfig::default());
-    rt.spawn_named("alpha", PingPongActor).unwrap();
-    rt.spawn_named("beta", PingPongActor).unwrap();
-    rt.spawn_named("gamma", PingPongActor).unwrap();
-
-    let mut names = rt.registered_names();
-    names.sort();
-    assert_eq!(names, vec!["alpha", "beta", "gamma"]);
-}
-
-/// Given a named actor exists,
-/// when I manually unregister the name,
-/// then the name is freed but the actor continues running.
-#[test]
-fn manual_unregister_frees_name_but_actor_lives() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    let addr = rt.spawn_named("temp-name", PingPongActor).unwrap();
-    rt.tick(); // on_start
-
-    let removed = rt.unregister("temp-name");
-    assert_eq!(removed, Some(addr));
-    assert_eq!(rt.where_is("temp-name"), None, "name freed");
-
-    // Actor still alive and can receive messages
-    rt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick();
-    assert!(inbox.try_recv().is_some(), "actor still processes messages");
-}
-
-/// An actor that looks up a peer by name using ctx.where_is().
-struct NameLookupActor {
-    target_name: &'static str,
-    reply_to: ActorAddress,
-}
-
-impl ActorInterface for NameLookupActor {
-    type Incoming = Ping;
-    type Response = ();
-    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
-        if let Some(peer) = ctx.where_is(self.target_name) {
-            ctx.send(self.reply_to, MyAddr(peer)).unwrap();
-        }
-    }
-}
-
-/// Given a named actor exists,
-/// when another actor calls ctx.where_is() from inside a handler,
-/// then it resolves the correct address.
-#[test]
-fn ctx_where_is_resolves_inside_handler() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<MyAddr>().unwrap();
-    let target = rt.spawn_named("target", PingPongActor).unwrap();
-
-    let looker = rt.spawn(NameLookupActor {
-        target_name: "target",
-        reply_to: *inbox.addr(),
-    }).unwrap();
-
-    rt.tick(); // on_start
-    rt.send_to(looker, Ping { reply_to: ActorAddress::default() }).unwrap();
-    rt.tick(); // handle -> where_is -> send
-    rt.tick(); // deliver reply
-
-    let result = inbox.try_recv();
-    assert_eq!(result, Some(MyAddr(target)), "ctx.where_is found the named actor");
-}
-
-/// An actor that spawns a named child using ctx.spawn_named().
-struct NamedSpawnerActor {
-    reply_to: ActorAddress,
-}
-
-impl ActorInterface for NamedSpawnerActor {
-    type Incoming = Ping;
-    type Response = ();
-    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
-        match ctx.spawn_named("child", PingPongActor) {
-            Ok(addr) => { ctx.send(self.reply_to, MyAddr(addr)).unwrap(); }
-            Err(_) => {}
-        }
-    }
-}
-
-/// Given an actor calls ctx.spawn_named("child", ...),
-/// when the child is spawned,
-/// then where_is("child") returns the correct address.
-#[test]
-fn ctx_spawn_named_registers_from_handler() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<MyAddr>().unwrap();
-
-    let spawner = rt.spawn(NamedSpawnerActor {
-        reply_to: *inbox.addr(),
-    }).unwrap();
-
-    rt.tick(); // on_start
-    rt.send_to(spawner, Ping { reply_to: ActorAddress::default() }).unwrap();
-    rt.tick(); // handle -> spawn_named
-    rt.tick(); // deliver reply
-
-    let child_addr = inbox.try_recv().expect("should receive child address");
-    assert_eq!(rt.where_is("child"), Some(child_addr.0), "name registered from handler");
-}
-
-// ── Actor Monitoring / Death Watch ──────────────────────────────────────────
-
-/// An actor that monitors a target and forwards Down notifications to a reply address.
-struct WatcherActor {
-    watch_target: ActorAddress,
-    reply_to: ActorAddress,
-    mref: Option<MonitorRef>,
-}
-
-impl ActorInterface for WatcherActor {
-    type Incoming = Down;
-    type Response = ();
-    fn on_start(&mut self, ctx: &Ctx) {
-        self.mref = Some(ctx.monitor(self.watch_target));
-    }
-    fn handle(&mut self, ctx: &Ctx, msg: Down) {
-        // Forward the Down notification to the test inbox
-        ctx.send(self.reply_to, msg).unwrap();
-    }
-}
-
-/// Given actor A monitors actor B,
-/// when B is gracefully stopped,
-/// then A receives a Down { reason: Normal } message.
-#[test]
-fn monitor_notifies_on_graceful_stop() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Down>().unwrap();
-
-    let target = rt.spawn(PingPongActor).unwrap();
-    let _watcher = rt.spawn(WatcherActor {
-        watch_target: target,
-        reply_to: *inbox.addr(),
-        mref: None,
-    }).unwrap();
-
-    rt.tick(); // on_start -> watcher sets up monitor
-    rt.stop_actor(target).unwrap();
-    rt.tick(); // target receives StopSignal -> cleanup_dead emits Down
-    rt.tick(); // watcher receives Down -> forwards to inbox
-
-    let down = inbox.try_recv().expect("should receive Down notification");
-    assert_eq!(down.addr, target);
-    assert_eq!(down.reason, StopReason::Normal);
-}
-
-/// Given actor A monitors actor B,
-/// when B panics,
-/// then A receives a Down { reason: Panicked } message.
-#[test]
-fn monitor_notifies_on_panic() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Down>().unwrap();
-
-    let target = rt.spawn(PanicActor).unwrap();
-    let _watcher = rt.spawn(WatcherActor {
-        watch_target: target,
-        reply_to: *inbox.addr(),
-        mref: None,
-    }).unwrap();
-
-    rt.tick(); // on_start
-    rt.send_to(target, PanicMsg).unwrap();
-    rt.tick(); // target panics -> cleanup_dead emits Down
-    rt.tick(); // watcher receives Down -> forwards to inbox
-
-    let down = inbox.try_recv().expect("should receive Down on panic");
-    assert_eq!(down.addr, target);
-    assert_eq!(down.reason, StopReason::Panicked);
-}
-
-/// Given two actors both monitor the same target,
-/// when the target dies,
-/// then both watchers receive independent Down notifications.
-#[test]
-fn multiple_watchers_all_notified() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox1 = rt.new_inbox::<Down>().unwrap();
-    let inbox2 = rt.new_inbox::<Down>().unwrap();
-
-    let target = rt.spawn(PingPongActor).unwrap();
-    rt.spawn(WatcherActor {
-        watch_target: target,
-        reply_to: *inbox1.addr(),
-        mref: None,
-    }).unwrap();
-    rt.spawn(WatcherActor {
-        watch_target: target,
-        reply_to: *inbox2.addr(),
-        mref: None,
-    }).unwrap();
-
-    rt.tick(); // on_start for all
-    rt.stop_actor(target).unwrap();
-    rt.tick(); // cleanup -> Down emitted to both watchers
-    rt.tick(); // watchers forward Down to inboxes
-
-    assert!(inbox1.try_recv().is_some(), "watcher 1 should receive Down");
-    assert!(inbox2.try_recv().is_some(), "watcher 2 should receive Down");
-}
-
-/// An actor that demonitors in response to a Ping message.
-struct DemonitorActor {
-    watch_target: ActorAddress,
-    mref: Option<MonitorRef>,
-}
-
-impl ActorInterface for DemonitorActor {
-    type Incoming = Ping;
-    type Response = ();
-    fn on_start(&mut self, ctx: &Ctx) {
-        self.mref = Some(ctx.monitor(self.watch_target));
-    }
-    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
-        // Cancel the monitor
-        if let Some(mref) = self.mref.take() {
-            ctx.demonitor(mref);
-        }
-    }
-}
-
-/// Given actor A monitors actor B then demonitors,
-/// when B dies,
-/// then A does NOT receive a Down notification.
-#[test]
-fn demonitor_cancels_notification() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let down_inbox = rt.new_inbox::<Down>().unwrap();
-
-    let target = rt.spawn(PingPongActor).unwrap();
-    let watcher = rt.spawn(DemonitorActor {
-        watch_target: target,
-        mref: None,
-    }).unwrap();
-
-    rt.tick(); // on_start -> monitor set up
-
-    // Trigger demonitor
-    rt.send_to(watcher, Ping { reply_to: ActorAddress::default() }).unwrap();
-    rt.tick(); // handle -> demonitor
-
-    // Now kill the target
-    rt.stop_actor(target).unwrap();
-    rt.tick(); // cleanup -- no Down should be emitted
-    rt.tick(); // extra tick to be sure
-
-    assert!(down_inbox.try_recv().is_none(), "demonitored -- should NOT receive Down");
-}
-
-/// Given actor A monitors B, and A dies before B,
-/// when B dies,
-/// then no Down is delivered (dead watcher cleaned up).
-#[test]
-fn dead_watcher_does_not_receive_down() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let target = rt.spawn(PingPongActor).unwrap();
-    let watcher = rt.spawn(WatcherActor {
-        watch_target: target,
-        reply_to: ActorAddress::default(), // won't matter, watcher dies first
-        mref: None,
-    }).unwrap();
-
-    rt.tick(); // on_start -> monitor set up
-    rt.stop_actor(watcher).unwrap();
-    rt.tick(); // watcher dies -> its monitors are cleaned up
-
-    // Now kill the target -- the dead watcher's subscription should be gone
-    rt.stop_actor(target).unwrap();
-    rt.tick(); // cleanup -- should not panic or try to deliver to dead watcher
-    // If we get here without panic, the test passes
-}
-
-/// Given an external inbox monitors via the runtime,
-/// when the target dies,
-/// then the inbox receives a Down message.
-#[test]
-fn down_delivered_to_external_inbox() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Down>().unwrap();
-
-    let target = rt.spawn(PingPongActor).unwrap();
-
-    let _watcher = rt.spawn(WatcherActor {
-        watch_target: target,
-        reply_to: *inbox.addr(),
-        mref: None,
-    }).unwrap();
-
-    rt.tick(); // on_start
-    rt.stop_actor(target).unwrap();
-    rt.tick(); // cleanup -> Down to watcher
-    rt.tick(); // watcher forwards to inbox
-
-    let down = inbox.try_recv().expect("inbox should receive forwarded Down");
-    assert_eq!(down.addr, target);
-    assert_eq!(down.reason, StopReason::Normal);
-}
-
-/// Given actor A monitors B with two independent monitors,
-/// when B dies,
-/// then A receives two Down messages (one per monitor).
-#[test]
-fn stacked_monitors_produce_multiple_notifications() {
-    /// An actor that creates two monitors on the same target.
-    struct DoubleWatcherActor {
-        target: ActorAddress,
-        reply_to: ActorAddress,
-    }
-
-    impl ActorInterface for DoubleWatcherActor {
-        type Incoming = Down;
-        type Response = ();
-        fn on_start(&mut self, ctx: &Ctx) {
-            ctx.monitor(self.target);
-            ctx.monitor(self.target);
-        }
-        fn handle(&mut self, ctx: &Ctx, msg: Down) {
-            ctx.send(self.reply_to, msg).unwrap();
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Down>().unwrap();
-
-    let target = rt.spawn(PingPongActor).unwrap();
-    rt.spawn(DoubleWatcherActor {
-        target,
-        reply_to: *inbox.addr(),
-    }).unwrap();
-
-    rt.tick(); // on_start -> 2 monitors
-    rt.stop_actor(target).unwrap();
-    rt.tick(); // cleanup -> 2 Down messages to watcher
-    rt.tick(); // watcher forwards both to inbox
-
-    assert!(inbox.try_recv().is_some(), "first Down");
-    assert!(inbox.try_recv().is_some(), "second Down");
-    assert!(inbox.try_recv().is_none(), "no more");
-}
-
-// ── Actor Groups / Pub-Sub ──────────────────────────────────────────────────
-
-/// Given actors join a group,
-/// when I query group_members,
-/// then all joined actors are listed.
-#[test]
-fn group_members_returns_joined_actors() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let a = rt.spawn(PingPongActor).unwrap();
-    let b = rt.spawn(PingPongActor).unwrap();
-
-    rt.join_group(a, "workers");
-    rt.join_group(b, "workers");
-
-    let mut members = rt.group_members("workers");
-    members.sort_by_key(|addr| addr.0);
-    let mut expected = vec![a, b];
-    expected.sort_by_key(|addr| addr.0);
-    assert_eq!(members, expected);
-}
-
-/// Given no actors have joined a group,
-/// when I query group_members,
-/// then the result is empty.
-#[test]
-fn empty_group_returns_no_members() {
-    let rt = std_runtime(RuntimeConfig::default());
-    assert!(rt.group_members("nonexistent").is_empty());
-}
-
-/// Given actors in a group,
-/// when a message is published to the group,
-/// then all members receive the message.
-#[test]
-fn publish_broadcasts_to_all_members() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox1 = rt.new_inbox::<Pong>().unwrap();
-    let inbox2 = rt.new_inbox::<Pong>().unwrap();
-
-    let a = rt.spawn(PingPongActor).unwrap();
-    let b = rt.spawn(PingPongActor).unwrap();
-    rt.join_group(a, "pongers");
-    rt.join_group(b, "pongers");
-
-    rt.tick(); // on_start
-
-    // Publish a Ping with inbox1's addr as reply_to.
-    let count = rt.publish_to("pongers", Ping { reply_to: *inbox1.addr() });
-    assert_eq!(count, 2, "two members, two messages sent");
-
-    rt.tick(); // actors handle Ping -> send Pong to inbox1
-
-    // Both actors send to inbox1
-    assert!(inbox1.try_recv().is_some(), "first Pong");
-    assert!(inbox1.try_recv().is_some(), "second Pong");
-    assert!(inbox1.try_recv().is_none(), "no more");
-    drop(inbox2);
-}
-
-/// Given an actor leaves a group,
-/// when a message is published,
-/// then the leaver does not receive it.
-#[test]
-fn leave_group_stops_receiving_publishes() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let a = rt.spawn(PingPongActor).unwrap();
-    let b = rt.spawn(PingPongActor).unwrap();
-    rt.join_group(a, "pool");
-    rt.join_group(b, "pool");
-    rt.leave_group(b, "pool");
-
-    rt.tick(); // on_start
-    let count = rt.publish_to("pool", Ping { reply_to: *inbox.addr() });
-    assert_eq!(count, 1, "only one member after leave");
-
-    rt.tick();
-    assert!(inbox.try_recv().is_some(), "one Pong from remaining member");
-    assert!(inbox.try_recv().is_none(), "no second Pong");
-}
-
-/// Given a group member dies,
-/// when a message is published,
-/// then the dead member is not included.
-#[test]
-fn dead_actor_auto_removed_from_group() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    let a = rt.spawn(PingPongActor).unwrap();
-    let b = rt.spawn(PingPongActor).unwrap();
-    rt.join_group(a, "team");
-    rt.join_group(b, "team");
-
-    rt.tick(); // on_start
-    rt.stop_actor(b).unwrap();
-    rt.tick(); // b dies, cleaned up from group
-
-    let count = rt.publish_to("team", Ping { reply_to: *inbox.addr() });
-    assert_eq!(count, 1, "dead actor removed from group");
-
-    rt.tick();
-    assert!(inbox.try_recv().is_some());
-    assert!(inbox.try_recv().is_none());
-}
-
-/// Given an actor is in multiple groups,
-/// when the actor dies,
-/// then it is removed from all groups.
-#[test]
-fn actor_removed_from_all_groups_on_death() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(PingPongActor).unwrap();
-    rt.join_group(actor, "alpha");
-    rt.join_group(actor, "beta");
-    rt.join_group(actor, "gamma");
-
-    rt.tick();
-    rt.stop_actor(actor).unwrap();
-    rt.tick(); // cleanup removes from all groups
-
-    assert!(rt.group_members("alpha").is_empty());
-    assert!(rt.group_members("beta").is_empty());
-    assert!(rt.group_members("gamma").is_empty());
-}
-
-/// Given a group becomes empty after its last member leaves,
-/// then the group name disappears from the active groups list.
-#[test]
-fn empty_group_auto_deleted() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(PingPongActor).unwrap();
-    rt.join_group(actor, "temp");
-    assert!(rt.groups().contains(&"temp".to_string()));
-
-    rt.leave_group(actor, "temp");
-    assert!(!rt.groups().contains(&"temp".to_string()), "empty group should be removed");
-}
-
-/// Given actors join groups from handlers using ctx.join_group(),
-/// when group_members is queried,
-/// then the joining actors are listed.
-#[test]
-fn ctx_join_group_from_handler() {
-    struct GroupJoinerActor;
-
-    impl ActorInterface for GroupJoinerActor {
-        type Incoming = Ping;
-        type Response = ();
-        fn on_start(&mut self, ctx: &Ctx) {
-            ctx.join_group("auto-joined");
-        }
-        fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {}
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let a = rt.spawn(GroupJoinerActor).unwrap();
-    let b = rt.spawn(GroupJoinerActor).unwrap();
-
-    rt.tick(); // on_start -> both join "auto-joined"
-
-    let members = rt.group_members("auto-joined");
-    assert_eq!(members.len(), 2);
-    assert!(members.contains(&a));
-    assert!(members.contains(&b));
-}
-
-/// Given an actor uses ctx.publish() from inside a handler,
-/// when the published message is processed,
-/// then all group members receive it.
-#[test]
-fn ctx_publish_broadcasts_from_handler() {
-    #[derive(Clone)]
-    struct BroadcastCmd {
-        reply_to: ActorAddress,
-    }
-
-    struct BroadcasterActor;
-
-    impl ActorInterface for BroadcasterActor {
-        type Incoming = BroadcastCmd;
-        type Response = ();
-        fn on_start(&mut self, ctx: &Ctx) {
-            ctx.join_group("broadcast-test");
-        }
-        fn handle(&mut self, ctx: &Ctx, msg: BroadcastCmd) {
-            ctx.publish("broadcast-test", Ping { reply_to: msg.reply_to });
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-
-    // Spawn 3 PingPongActors and one Broadcaster, all in the same group
-    let _p1 = rt.spawn(PingPongActor).unwrap();
-    let _p2 = rt.spawn(PingPongActor).unwrap();
-    rt.join_group(_p1, "broadcast-test");
-    rt.join_group(_p2, "broadcast-test");
-
-    let broadcaster = rt.spawn(BroadcasterActor).unwrap();
-
-    rt.tick(); // on_start (broadcaster joins group too)
-
-    // Send BroadcastCmd to broadcaster
-    rt.send_to(broadcaster, BroadcastCmd { reply_to: *inbox.addr() }).unwrap();
-    rt.tick(); // broadcaster handles -> publish Ping to all 3 members (including self)
-    rt.tick(); // PingPong actors handle Ping -> send Pong to inbox
-
-    // At least 2 Pongs from the PingPongActors
-    let mut pong_count = 0;
-    while inbox.try_recv().is_some() {
-        pong_count += 1;
-    }
-    assert!(pong_count >= 2, "at least 2 PingPong members should reply, got {pong_count}");
-}
-
-// ── Ask Pattern ─────────────────────────────────────────────────────────────
-
-/// Given a PingPong actor,
-/// when I ask with recv_ticking,
-/// then I get the Pong response.
-#[test]
-fn ask_recv_ticking_returns_response() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(PingPongActor).unwrap();
-    rt.tick(); // on_start
-
-    let pong: Pong = rt.ask(actor, |reply_to| Ping { reply_to })
-        .unwrap()
-        .recv_ticking(&rt, 10)
-        .unwrap();
-    assert_eq!(pong, Pong);
-}
-
-/// Given a CounterActor,
-/// when I ask multiple times,
-/// then each response reflects the updated state.
-#[test]
-fn ask_multiple_times_tracks_state() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(CounterActor { count: 0 }).unwrap();
-    rt.tick(); // on_start
-
-    let c1: Count = rt.ask(actor, |reply_to| Increment { reply_to })
-        .unwrap().recv_ticking(&rt, 10).unwrap();
-    let c2: Count = rt.ask(actor, |reply_to| Increment { reply_to })
-        .unwrap().recv_ticking(&rt, 10).unwrap();
-    let c3: Count = rt.ask(actor, |reply_to| Increment { reply_to })
-        .unwrap().recv_ticking(&rt, 10).unwrap();
-
-    assert_eq!(c1, Count(1));
-    assert_eq!(c2, Count(2));
-    assert_eq!(c3, Count(3));
-}
-
-/// Given a dead actor,
-/// when I ask and tick,
-/// then recv_ticking returns a timeout error.
-#[test]
-fn ask_timeout_when_no_response() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(PingPongActor).unwrap();
-    rt.tick();
-    rt.stop_actor(actor).unwrap();
-    rt.tick(); // actor dies
-
-    // Ask the dead actor -- message is undeliverable, no response
-    let result = rt.ask::<Ping, Pong>(actor, |reply_to| Ping { reply_to });
-    // send_to may succeed or fail
-    if let Ok(ask) = result {
-        let err = ask.recv_ticking(&rt, 5);
-        assert!(err.is_err(), "should timeout with no response");
-    }
-}
-
-/// Given an ask handle,
-/// when I use try_recv before ticking,
-/// then it returns None (response hasn't arrived yet).
-#[test]
-fn ask_try_recv_returns_none_before_tick() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(PingPongActor).unwrap();
-    rt.tick(); // on_start
-
-    let ask = rt.ask::<Ping, Pong>(actor, |reply_to| Ping { reply_to }).unwrap();
-    assert!(ask.try_recv().is_none(), "no response before ticking");
-
-    rt.tick(); // process message
-    assert_eq!(ask.try_recv(), Some(Pong));
-}
-
-/// Given an ask, the reply_addr() returns the inbox address for manual use.
-#[test]
-fn ask_reply_addr_is_accessible() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let actor = rt.spawn(PingPongActor).unwrap();
-    rt.tick();
-
-    let ask = rt.ask::<Ping, Pong>(actor, |reply_to| Ping { reply_to }).unwrap();
-    let addr = *ask.reply_addr();
-    // The address should be valid (non-zero)
-    assert_ne!(addr, ActorAddress::default());
-}
--- a/tests/runtime_stress.rs
+++ b/tests/runtime_stress.rs
@ -0,0 +1,364 @@
+//! Runtime Stress Tests — multi-threaded execution, parking, placement, and scale.
+//!
+//! Covers: single vs multi-threaded processing, high-volume MT delivery,
+//! panic isolation under load, worker parking/shutdown, sustained throughput,
+//! and load-aware actor placement.
+
+mod common;
+use common::*;
+
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::Arc;
+use std::time::{Duration, Instant};
+
+// ── Helpers ──────────────────────────────────────────────────────────────────
+
+/// Poll `inbox` until a message arrives or `timeout` elapses.
+fn poll_inbox<M: swactor::actor::Message>(
+    inbox: &Inbox<M>,
+    timeout: Duration,
+) -> Option<M> {
+    let deadline = Instant::now() + timeout;
+    loop {
+        if let Some(msg) = inbox.try_recv() {
+            return Some(msg);
+        }
+        if Instant::now() > deadline {
+            return None;
+        }
+        std::thread::sleep(Duration::from_millis(1));
+    }
+}
+
+/// Wait until `counter` reaches `target` or `timeout` elapses.
+fn wait_for_count(counter: &AtomicUsize, target: usize, timeout: Duration) -> usize {
+    let deadline = Instant::now() + timeout;
+    loop {
+        let n = counter.load(Ordering::SeqCst);
+        if n >= target {
+            return n;
+        }
+        if Instant::now() > deadline {
+            return n;
+        }
+        std::thread::sleep(Duration::from_millis(5));
+    }
+}
+
+// ── Tests ────────────────────────────────────────────────────────────────────
+
+/// Single-threaded vs multi-threaded runtime basics.
+///
+/// Story: We start with a single-threaded runtime driven by tick(), confirm
+/// nothing happens without ticking, then graduate to a multi-threaded runtime
+/// with run() and verify background processing, cross-worker delegation,
+/// custom thread counts, and clean shutdown.
+#[test]
+fn single_vs_multi_threaded_basics() {
+    // ── Part A: Single-threaded requires tick() ──
+    let rt = std_runtime(RuntimeConfig::default());
+    let addr = rt.spawn(PingPongActor).unwrap();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    rt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
+
+    assert!(inbox.try_recv().is_none(), "no processing before tick");
+    tick_n(&rt, 2);
+    assert!(inbox.try_recv().is_some(), "tick() drives single-threaded processing");
+
+    // ── Part B: Multi-threaded processes without ticking ──
+    let rt_mt = std_runtime(RuntimeConfig { num_threads: 4, ..Default::default() });
+    let addr = rt_mt.spawn(PingPongActor).unwrap();
+    let inbox = rt_mt.new_inbox::<Pong>().unwrap();
+    rt_mt.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
+
+    let handle = rt_mt.run().unwrap();
+    let reply = poll_inbox(&inbox, Duration::from_secs(5));
+    assert!(reply.is_some(), "background workers process without manual ticking");
+
+    // ── Part C: Cross-worker delegation (2 threads, spawn child from handler) ──
+    let addr2 = handle.runtime.spawn(DelegatorActor).unwrap();
+    let done_inbox = handle.runtime.new_inbox::<Done>().unwrap();
+    handle.runtime.send_to(addr2, Forward { value: 3, reply_to: *done_inbox.addr() }).unwrap();
+
+    let reply = poll_inbox(&done_inbox, Duration::from_secs(5));
+    assert_eq!(reply, Some(Done(6)), "cross-worker delegation delivers reply");
+
+    // ── Part D: Custom thread count reflected in stats ──
+    let stats = handle.runtime.stats();
+    assert_eq!(stats.num_workers, 4, "runtime respects requested thread count");
+
+    // ── Part E: Clean shutdown ──
+    handle.shutdown();
+    handle.join();
+    // Test passes by not hanging.
+}
+
+/// High-volume multi-threaded delivery.
+///
+/// Story: We throw large workloads at a 4-thread runtime — 50 senders
+/// each firing 100 messages at one receiver, 200 concurrent spawn+send
+/// pairs, and a 50-level chain that must hop across workers.
+#[test]
+fn mt_high_volume_delivery() {
+    let cfg = || RuntimeConfig {
+        num_threads: 4,
+        max_actors: 5_000,
+        channel_buffer_size: 10_000,
+        ..Default::default()
+    };
+
+    // ── Part A: 50 senders × 100 messages → one receiver ──
+    {
+        let rt = std_runtime(cfg());
+        let counter = Arc::new(AtomicUsize::new(0));
+        let dummy = rt.new_inbox::<Pong>().unwrap();
+        let receiver = rt.spawn(CountingPingActor { counter: counter.clone() }).unwrap();
+
+        let total_expected = 50 * 100;
+        for _ in 0..50 {
+            for _ in 0..100 {
+                rt.send_to(receiver, Ping { reply_to: *dummy.addr() }).unwrap();
+            }
+        }
+
+        let handle = rt.run().unwrap();
+        let processed = wait_for_count(&counter, total_expected, Duration::from_secs(5));
+        handle.shutdown();
+        handle.join();
+        assert_eq!(processed, total_expected, "all 5000 messages delivered to single receiver");
+    }
+
+    // ── Part B: 200 concurrent spawn+send pairs ──
+    {
+        let rt = std_runtime(cfg());
+        let counter = Arc::new(AtomicUsize::new(0));
+        let dummy = rt.new_inbox::<Pong>().unwrap();
+        for _ in 0..200 {
+            let a = rt.spawn(CountingPingActor { counter: counter.clone() }).unwrap();
+            rt.send_to(a, Ping { reply_to: *dummy.addr() }).unwrap();
+        }
+
+        let handle = rt.run().unwrap();
+        let received = wait_for_count(&counter, 200, Duration::from_secs(5));
+        handle.shutdown();
+        handle.join();
+        assert_eq!(received, 200, "all 200 spawn+send pairs complete");
+    }
+
+    // ── Part C: 50-level chain across workers ──
+    {
+        let rt = std_runtime(RuntimeConfig { num_threads: 2, max_actors: 5_000, ..Default::default() });
+        let addr = rt.spawn(ChainActor).unwrap();
+        let inbox = rt.new_inbox::<Done>().unwrap();
+        rt.send_to(addr, ChainMsg { remaining: 50, depth: 0, reply_to: *inbox.addr() }).unwrap();
+
+        let handle = rt.run().unwrap();
+        let reply = poll_inbox(&inbox, Duration::from_secs(5));
+        handle.shutdown();
+        handle.join();
+        assert_eq!(reply, Some(Done(50)), "50-level chain completes across workers");
+    }
+}
+
+/// Panic isolation under multi-threaded load.
+///
+/// Story: 10 panicking actors and 10 healthy actors on 4 threads — every
+/// panic is isolated and all 1000 healthy messages are still processed.
+#[test]
+fn mt_panic_isolation_under_load() {
+    let rt = std_runtime(RuntimeConfig {
+        num_threads: 4,
+        max_actors: 5_000,
+        channel_buffer_size: 10_000,
+        ..Default::default()
+    });
+
+    let counter = Arc::new(AtomicUsize::new(0));
+    let dummy = rt.new_inbox::<Pong>().unwrap();
+
+    let mut panic_addrs = Vec::new();
+    let mut healthy_addrs = Vec::new();
+    for _ in 0..10 {
+        panic_addrs.push(rt.spawn(PanicActor).unwrap());
+        healthy_addrs.push(rt.spawn(CountingPingActor { counter: counter.clone() }).unwrap());
+    }
+
+    // Trigger panics and flood healthy actors.
+    for &addr in &panic_addrs {
+        rt.send_to(addr, PanicMsg).unwrap();
+    }
+    for &addr in &healthy_addrs {
+        for _ in 0..100 {
+            rt.send_to(addr, Ping { reply_to: *dummy.addr() }).unwrap();
+        }
+    }
+
+    let handle = rt.run().unwrap();
+    let expected = 10 * 100;
+    let processed = wait_for_count(&counter, expected, Duration::from_secs(5));
+    handle.shutdown();
+    handle.join();
+
+    assert_eq!(
+        processed, expected,
+        "all {expected} healthy messages processed despite panicking peers"
+    );
+}
+
+/// Worker parking and shutdown latency.
+///
+/// Story: Workers park after idle time. We verify they wake quickly on new
+/// messages, that messages sent after run() are delivered, and that shutdown
+/// wakes all parked workers promptly.
+#[test]
+fn worker_parking_and_shutdown() {
+    // ── Part A: Parked workers wake on send ──
+    let rt = std_runtime(RuntimeConfig { num_threads: 2, ..Default::default() });
+    let addr = rt.spawn(PingPongActor).unwrap();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+
+    let handle = rt.run().unwrap();
+    std::thread::sleep(Duration::from_millis(50)); // Let workers park.
+
+    let before = Instant::now();
+    handle.runtime.send_to(addr, Ping { reply_to: *inbox.addr() }).unwrap();
+    let reply = poll_inbox(&inbox, Duration::from_secs(1));
+    let latency = before.elapsed();
+
+    assert!(reply.is_some(), "parked worker should wake and process");
+    assert!(latency.as_millis() < 100, "wake latency should be <100ms, was {:?}", latency);
+
+    // ── Part B: Send after run() delivers ──
+    let addr2 = handle.runtime.spawn(PingPongActor).unwrap();
+    let inbox2 = handle.runtime.new_inbox::<Pong>().unwrap();
+    std::thread::sleep(Duration::from_millis(10));
+    handle.runtime.send_to(addr2, Ping { reply_to: *inbox2.addr() }).unwrap();
+
+    let reply2 = poll_inbox(&inbox2, Duration::from_secs(5));
+    assert!(reply2.is_some(), "message sent after run() must be delivered");
+
+    handle.shutdown();
+    handle.join();
+
+    // ── Part C: Shutdown wakes parked workers quickly ──
+    let rt2 = std_runtime(RuntimeConfig { num_threads: 4, ..Default::default() });
+    let h2 = rt2.run().unwrap();
+    std::thread::sleep(Duration::from_millis(50)); // Let workers park.
+
+    let before = Instant::now();
+    h2.shutdown();
+    h2.join();
+    let shutdown_time = before.elapsed();
+
+    assert!(
+        shutdown_time.as_millis() < 500,
+        "shutdown should complete quickly with parked workers, took {:?}",
+        shutdown_time
+    );
+}
+
+/// Sustained throughput with no message loss.
+///
+/// Story: We send 10 batches of 100 messages, ticking between batches on a
+/// single-threaded runtime. Each batch must make forward progress, and after
+/// draining, all 1000 messages are accounted for.
+#[test]
+fn sustained_throughput_no_message_loss() {
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter = Arc::new(AtomicUsize::new(0));
+    let dummy = rt.new_inbox::<Pong>().unwrap();
+    let addr = rt.spawn(CountingPingActor { counter: counter.clone() }).unwrap();
+
+    for batch in 0..10 {
+        for _ in 0..100 {
+            rt.send_to(addr, Ping { reply_to: *dummy.addr() }).unwrap();
+        }
+        tick_n(&rt, 5);
+        let processed = counter.load(Ordering::SeqCst);
+        assert!(
+            processed > batch * 50,
+            "batch {batch}: expected progress, only {processed} processed"
+        );
+    }
+
+    // Drain remaining.
+    tick_n(&rt, 100);
+    let total = counter.load(Ordering::SeqCst);
+    assert_eq!(total, 1000, "sustained load should not drop any messages");
+}
+
+/// Load-aware actor placement.
+///
+/// Story: A fresh runtime falls back to round-robin (even distribution).
+/// Under imbalanced load, new actors bias toward the lighter worker.
+/// A single-worker runtime degrades gracefully.
+#[test]
+fn load_aware_actor_placement() {
+    // ── Part A: Round-robin fallback on fresh runtime (4 workers, 100 actors) ──
+    let rt = std_runtime(RuntimeConfig { num_threads: 4, ..Default::default() });
+    for _ in 0..100 {
+        rt.spawn(CounterActor { count: 0 }).unwrap();
+    }
+
+    let handle = rt.run().unwrap();
+    std::thread::sleep(Duration::from_millis(20));
+
+    let stats = handle.runtime.stats();
+    handle.shutdown();
+    handle.join();
+
+    for w in &stats.workers {
+        assert!(
+            w.num_actors >= 20 && w.num_actors <= 30,
+            "worker {} has {} actors, expected ~25 (round-robin)",
+            w.id, w.num_actors
+        );
+    }
+
+    // ── Part B: Imbalanced load biases toward lighter worker ──
+    let rt2 = std_runtime(RuntimeConfig { num_threads: 2, ..Default::default() });
+    let mut addrs = Vec::new();
+    for _ in 0..20 {
+        addrs.push(rt2.spawn(CounterActor { count: 0 }).unwrap());
+    }
+
+    let h2 = rt2.run().unwrap();
+    std::thread::sleep(Duration::from_millis(10));
+
+    // Bombard the first 10 actors (likely worker 0) with messages.
+    for addr in &addrs[..10] {
+        for _ in 0..50 {
+            let _ = h2.runtime.send_to(*addr, Increment { reply_to: *addr });
+        }
+    }
+    std::thread::sleep(Duration::from_millis(20));
+
+    // Spawn 10 more — should bias toward lighter worker.
+    for _ in 0..10 {
+        h2.runtime.spawn(CounterActor { count: 0 }).unwrap();
+    }
+    std::thread::sleep(Duration::from_millis(20));
+
+    let stats2 = h2.runtime.stats();
+    h2.shutdown();
+    h2.join();
+
+    let total_actors: usize = stats2.workers.iter().map(|w| w.num_actors).sum();
+    assert!(total_actors >= 20, "expected at least 20 actors, got {total_actors}");
+    assert!(
+        stats2.workers.iter().all(|w| w.num_actors > 0),
+        "both workers should have actors: {:?}",
+        stats2.workers.iter().map(|w| w.num_actors).collect::<Vec<_>>()
+    );
+
+    // ── Part C: Single-worker degrades gracefully ──
+    let rt3 = std_runtime(RuntimeConfig::default());
+    for _ in 0..50 {
+        rt3.spawn(CounterActor { count: 0 }).unwrap();
+    }
+    tick_n(&rt3, 10);
+
+    let stats3 = rt3.stats();
+    assert_eq!(stats3.workers.len(), 1);
+    assert_eq!(stats3.workers[0].num_actors, 50);
+}
--- a/tests/runtime_supervision.rs
+++ b/tests/runtime_supervision.rs
@ -1,866 +0,0 @@
-mod common;
-use common::*;
-
-use std::sync::atomic::{AtomicUsize, Ordering};
-use std::sync::Arc;
-
-// ─── Supervisor Helpers ────────────────────────────────────────────────────
-
-/// Actor that panics after receiving a configurable number of messages.
-struct PanicAfterN {
-    trigger: usize,
-    count: usize,
-    counter: Arc<AtomicUsize>,
-}
-
-impl ActorInterface for PanicAfterN {
-    type Incoming = Ping;
-    type Response = ();
-    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
-        self.count += 1;
-        self.counter.fetch_add(1, Ordering::SeqCst);
-        let _ = ctx.send(msg.reply_to, Pong);
-        if self.count >= self.trigger {
-            panic!("intentional panic at message {}", self.count);
-        }
-    }
-}
-
-// --- handle_down tests ---
-
-/// Given an actor with handle_down and a monitored target,
-/// when the target dies, the watcher receives a Down via handle_down.
-#[test]
-fn handle_down_receives_death_notification() {
-    struct MonitoringTracker {
-        target: ActorAddress,
-        downs: Vec<Down>,
-        inbox: ActorAddress,
-    }
-    impl ActorInterface for MonitoringTracker {
-        type Incoming = Ping;
-        type Response = ();
-        fn on_start(&mut self, ctx: &Ctx) {
-            ctx.monitor(self.target);
-        }
-        fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
-            let _ = ctx.send(self.inbox, Count(self.downs.len()));
-        }
-        fn handle_down(&mut self, _ctx: &Ctx, down: Down) {
-            self.downs.push(down);
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Count>().unwrap();
-    let inbox_addr = *inbox.addr();
-    let target = rt.spawn(PanicActor).unwrap();
-    let tracker = rt.spawn(MonitoringTracker {
-        target,
-        downs: vec![],
-        inbox: inbox_addr,
-    }).unwrap();
-    rt.tick(); // on_start for both
-
-    // Kill the target
-    rt.send_to(target, PanicMsg).unwrap();
-    rt.tick(); // target panics
-    rt.tick(); // Down delivered to tracker via handle_down
-
-    // Ask tracker how many downs it saw
-    rt.send_to(tracker, Ping { reply_to: inbox_addr }).unwrap();
-    rt.tick();
-    assert_eq!(inbox.try_recv(), Some(Count(1)));
-}
-
-/// Given an actor whose Incoming type IS Down, handle_down is NOT called --
-/// the Down goes through the normal handle() method (backward compatibility).
-#[test]
-fn handle_down_skipped_when_incoming_is_down() {
-    struct DownAsIncoming {
-        target: ActorAddress,
-        inbox: ActorAddress,
-    }
-    impl ActorInterface for DownAsIncoming {
-        type Incoming = Down;
-        type Response = ();
-        fn on_start(&mut self, ctx: &Ctx) {
-            ctx.monitor(self.target);
-        }
-        fn handle(&mut self, ctx: &Ctx, msg: Down) {
-            let _ = ctx.send(self.inbox, msg);
-        }
-        fn handle_down(&mut self, _ctx: &Ctx, _down: Down) {
-            panic!("handle_down must not be called when Incoming=Down");
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Down>().unwrap();
-    let inbox_addr = *inbox.addr();
-    let target = rt.spawn(PanicActor).unwrap();
-    let _watcher = rt.spawn(DownAsIncoming { target, inbox: inbox_addr }).unwrap();
-    rt.tick(); // on_start
-
-    rt.send_to(target, PanicMsg).unwrap();
-    rt.tick(); // panic
-    rt.tick(); // Down delivered through handle(), not handle_down
-
-    let received = inbox.try_recv().expect("Down should be delivered via handle()");
-    assert_eq!(received.reason, StopReason::Panicked);
-}
-
-// --- ctx.stop_actor tests ---
-
-/// Given two actors, one can stop the other via ctx.stop_actor().
-#[test]
-fn ctx_stop_actor_stops_target() {
-    #[derive(Clone)]
-    struct StopCmd {
-        target: ActorAddress,
-    }
-    struct Stopper;
-    impl ActorInterface for Stopper {
-        type Incoming = StopCmd;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, msg: StopCmd) {
-            let _ = ctx.stop_actor(msg.target);
-        }
-    }
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let target = rt.spawn(PingPongActor).unwrap();
-    let stopper = rt.spawn(Stopper).unwrap();
-    rt.tick(); // on_start
-
-    rt.send_to(stopper, StopCmd { target }).unwrap();
-    rt.tick(); // stopper handles StopCmd -> stop_actor(target)
-    rt.tick(); // StopSignal delivered to target, target stops
-    rt.tick(); // cleanup
-
-    assert!(rt.send_to(target, Ping { reply_to: ActorAddress::default() }).is_err());
-    // Stopper should still be alive
-    assert!(rt.send_to(stopper, StopCmd { target }).is_ok());
-}
-
-// --- Supervisor tests ---
-
-/// Given a supervisor with one permanent child,
-/// when the child panics, the supervisor restarts it.
-#[test]
-fn supervisor_restarts_permanent_child_on_panic() {
-    let counter = Arc::new(AtomicUsize::new(0));
-    let counter_c = counter.clone();
-    let inbox_holder: Arc<std::sync::Mutex<Option<ActorAddress>>> =
-        Arc::new(std::sync::Mutex::new(None));
-
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    let inbox_addr = *inbox.addr();
-    *inbox_holder.lock().unwrap() = Some(inbox_addr);
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![ChildSpec::new("worker", RestartPolicy::Permanent, move |ctx| {
-            ctx.spawn(PanicAfterN {
-                trigger: 2, // panics on 2nd message
-                count: 0,
-                counter: counter_c.clone(),
-            })
-        })],
-    );
-    let _sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); // supervisor on_start -> spawns child
-    rt.tick(); // child on_start
-
-    // Find the child by checking stats
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 2); // supervisor + child
-
-    // Discover child address from stats
-    let child_addr = stats.actors.iter()
-        .find(|(addr, _)| *addr != _sup_addr)
-        .map(|(addr, _)| *addr)
-        .unwrap();
-
-    // First message: child processes, increments counter
-    rt.send_to(child_addr, Ping { reply_to: inbox_addr }).unwrap();
-    rt.tick();
-    assert_eq!(counter.load(Ordering::SeqCst), 1);
-
-    // Second message: child panics (trigger=2)
-    rt.send_to(child_addr, Ping { reply_to: inbox_addr }).unwrap();
-    rt.tick(); // child panics and is poisoned
-    rt.tick(); // cleanup: Down delivered to supervisor via handle_down
-    rt.tick(); // supervisor restarts child (spawns new one)
-    rt.tick(); // new child on_start
-
-    // Supervisor is still alive, and a new child exists
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 2); // supervisor + new child
-}
-
-/// Given a supervisor with a transient child,
-/// when the child stops normally, it is NOT restarted.
-#[test]
-fn supervisor_does_not_restart_transient_child_on_normal_stop() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    struct StopsAfterFirst;
-    impl ActorInterface for StopsAfterFirst {
-        type Incoming = Ping;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
-            ctx.stop_self();
-        }
-    }
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![ChildSpec::new("worker", RestartPolicy::Transient, |ctx| {
-            ctx.spawn(StopsAfterFirst)
-        })],
-    );
-    let sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); // supervisor on_start -> child spawned
-    rt.tick(); // child on_start
-
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 2); // sup + child
-
-    // Find child address
-    let child_addr = stats.actors.iter()
-        .find(|(addr, _)| *addr != sup_addr)
-        .map(|(addr, _)| *addr)
-        .unwrap();
-
-    // Send message -- child stops itself
-    rt.send_to(child_addr, Ping { reply_to: ActorAddress::default() }).unwrap();
-    rt.tick(); // child handles, stops self
-    rt.tick(); // cleanup: Down(Normal) delivered to supervisor
-    rt.tick(); // supervisor sees Transient + Normal -> no restart
-
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 1); // only supervisor remains
-}
-
-/// Given a supervisor with a transient child,
-/// when the child panics, it IS restarted.
-#[test]
-fn supervisor_restarts_transient_child_on_panic() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let counter = Arc::new(AtomicUsize::new(0));
-    let counter_c = counter.clone();
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![ChildSpec::new("worker", RestartPolicy::Transient, move |ctx| {
-            ctx.spawn(PanicAfterN {
-                trigger: 1, // panics on first message
-                count: 0,
-                counter: counter_c.clone(),
-            })
-        })],
-    );
-    let sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); // supervisor starts, spawns child
-    rt.tick(); // child on_start
-
-    let child_addr = rt.stats().actors.iter()
-        .find(|(addr, _)| *addr != sup_addr)
-        .map(|(addr, _)| *addr)
-        .unwrap();
-
-    // Send message -- child panics
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    rt.send_to(child_addr, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick(); // child panics
-    rt.tick(); // Down(Panicked) -> supervisor restarts
-    rt.tick(); // new child spawned
-    rt.tick(); // new child on_start
-
-    // Supervisor + new child alive
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 2);
-}
-
-/// Given a supervisor with a temporary child,
-/// when the child dies (any reason), it is never restarted.
-#[test]
-fn supervisor_never_restarts_temporary_child() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![ChildSpec::new("worker", RestartPolicy::Temporary, |ctx| {
-            ctx.spawn(PanicActor)
-        })],
-    );
-    let sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); // supervisor starts, spawns child
-    rt.tick(); // child on_start
-
-    let child_addr = rt.stats().actors.iter()
-        .find(|(addr, _)| *addr != sup_addr)
-        .map(|(addr, _)| *addr)
-        .unwrap();
-
-    // Kill the child
-    rt.send_to(child_addr, PanicMsg).unwrap();
-    rt.tick(); // panic
-    rt.tick(); // Down -> supervisor sees Temporary -> no restart
-    rt.tick(); // settle
-
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 1); // only supervisor
-}
-
-/// Given a supervisor with max_restarts=2,
-/// when more than 2 restarts occur, the supervisor stops itself (meltdown).
-#[test]
-fn supervisor_meltdown_after_max_restarts() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let counter = Arc::new(AtomicUsize::new(0));
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        2, // only 2 restarts allowed
-        vec![ChildSpec::new("crasher", RestartPolicy::Permanent, {
-            let counter = counter.clone();
-            move |ctx| {
-                ctx.spawn(PanicAfterN {
-                    trigger: 1,
-                    count: 0,
-                    counter: counter.clone(),
-                })
-            }
-        })],
-    );
-    let sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); rt.tick(); // supervisor + child started
-
-    // Crash the child 3 times
-    for _ in 0..3 {
-        if let Some((child_addr, _)) = rt.stats().actors.iter()
-            .find(|(addr, _)| *addr != sup_addr)
-        {
-            let inbox = rt.new_inbox::<Pong>().unwrap();
-            let _ = rt.send_to(*child_addr, Ping { reply_to: *inbox.addr() });
-            rt.tick(); // child panics
-            rt.tick(); // Down delivered -> restart or meltdown
-            rt.tick(); // new child spawned (or supervisor stopped)
-            rt.tick(); // settle
-        }
-    }
-
-    // After 3 crashes with max_restarts=2, supervisor should have stopped itself
-    let stats = rt.stats();
-    let sup_alive = stats.actors.iter().any(|(addr, _)| *addr == sup_addr);
-    assert!(!sup_alive, "supervisor should have stopped after exceeding max_restarts");
-}
-
-/// Given a supervisor with multiple children,
-/// when one child panics, only that child is restarted (OneForOne).
-#[test]
-fn supervisor_one_for_one_only_restarts_failed_child() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let counter_a = Arc::new(AtomicUsize::new(0));
-    let counter_b = Arc::new(AtomicUsize::new(0));
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![
-            ChildSpec::new("crasher", RestartPolicy::Permanent, {
-                let c = counter_a.clone();
-                move |ctx| ctx.spawn_named("child_a", PanicAfterN {
-                    trigger: 1, count: 0, counter: c.clone(),
-                })
-            }),
-            ChildSpec::new("stable", RestartPolicy::Permanent, {
-                let c = counter_b.clone();
-                move |ctx| ctx.spawn_named("child_b", CountingPingActor { counter: c.clone() })
-            }),
-        ],
-    );
-    let _sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); rt.tick(); // start up
-
-    let child_a = rt.where_is("child_a").expect("child_a should be named");
-    let child_b = rt.where_is("child_b").expect("child_b should be named");
-
-    // Send to child_b to prove it's alive
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    rt.send_to(child_b, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick();
-    let b_processed_before = counter_b.load(Ordering::SeqCst);
-    assert!(b_processed_before >= 1);
-
-    // Crash child_a
-    rt.send_to(child_a, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick(); // child_a panics
-    rt.tick(); // Down -> supervisor restarts child_a
-    rt.tick(); rt.tick(); // new child spawned + on_start
-
-    // child_b should still be alive (same address, same name)
-    let child_b_after = rt.where_is("child_b").expect("child_b should still exist");
-    assert_eq!(child_b, child_b_after, "child_b address should be unchanged");
-
-    rt.send_to(child_b, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick();
-    assert!(counter_b.load(Ordering::SeqCst) > b_processed_before,
-        "child_b should still be processing messages");
-
-    // Supervisor + 2 children should be alive
-    assert_eq!(rt.stats().workers[0].num_actors, 3);
-}
-
-/// Given a OneForAll supervisor with 3 children,
-/// when one child panics, ALL children are stopped and restarted in spec order.
-#[test]
-fn supervisor_one_for_all_restarts_all_on_single_failure() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let counter_a = Arc::new(AtomicUsize::new(0));
-    let counter_b = Arc::new(AtomicUsize::new(0));
-    let counter_c = Arc::new(AtomicUsize::new(0));
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForAll,
-        5,
-        vec![
-            ChildSpec::new("a", RestartPolicy::Permanent, {
-                let c = counter_a.clone();
-                move |ctx| ctx.spawn_named("ofa_a", PanicAfterN {
-                    trigger: 1, count: 0, counter: c.clone(),
-                })
-            }),
-            ChildSpec::new("b", RestartPolicy::Permanent, {
-                let c = counter_b.clone();
-                move |ctx| ctx.spawn_named("ofa_b", CountingPingActor { counter: c.clone() })
-            }),
-            ChildSpec::new("c", RestartPolicy::Permanent, {
-                let c = counter_c.clone();
-                move |ctx| ctx.spawn_named("ofa_c", CountingPingActor { counter: c.clone() })
-            }),
-        ],
-    );
-    let _sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); rt.tick(); // startup
-
-    let old_b = rt.where_is("ofa_b").expect("ofa_b exists");
-    let old_c = rt.where_is("ofa_c").expect("ofa_c exists");
-    let child_a = rt.where_is("ofa_a").expect("ofa_a exists");
-
-    // Crash child_a
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    rt.send_to(child_a, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick(); // child_a panics
-    // supervisor receives Down(a) -> OneForAll -> stops b and c
-    for _ in 0..8 { rt.tick(); }
-
-    // All 3 children should be alive with NEW addresses
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 4); // sup + 3 new children
-
-    let new_b = rt.where_is("ofa_b").expect("ofa_b re-registered after restart");
-    let new_c = rt.where_is("ofa_c").expect("ofa_c re-registered after restart");
-    assert_ne!(old_b, new_b, "child_b should have a new address after restart");
-    assert_ne!(old_c, new_c, "child_c should have a new address after restart");
-}
-
-/// Given a RestForOne supervisor with children [a, b, c],
-/// when child b panics, children b and c are restarted.
-/// Child a is unaffected.
-#[test]
-fn supervisor_rest_for_one_restarts_rest_after_failed() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let counter_a = Arc::new(AtomicUsize::new(0));
-    let counter_b = Arc::new(AtomicUsize::new(0));
-    let counter_c = Arc::new(AtomicUsize::new(0));
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::RestForOne,
-        5,
-        vec![
-            ChildSpec::new("a", RestartPolicy::Permanent, {
-                let c = counter_a.clone();
-                move |ctx| ctx.spawn_named("rfo_a", CountingPingActor { counter: c.clone() })
-            }),
-            ChildSpec::new("b", RestartPolicy::Permanent, {
-                let c = counter_b.clone();
-                move |ctx| ctx.spawn_named("rfo_b", PanicAfterN {
-                    trigger: 1, count: 0, counter: c.clone(),
-                })
-            }),
-            ChildSpec::new("c", RestartPolicy::Permanent, {
-                let c = counter_c.clone();
-                move |ctx| ctx.spawn_named("rfo_c", CountingPingActor { counter: c.clone() })
-            }),
-        ],
-    );
-    let _sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); rt.tick(); // startup
-
-    let old_a = rt.where_is("rfo_a").expect("rfo_a exists");
-    let old_c = rt.where_is("rfo_c").expect("rfo_c exists");
-    let child_b = rt.where_is("rfo_b").expect("rfo_b exists");
-
-    // Crash child_b
-    let inbox = rt.new_inbox::<Pong>().unwrap();
-    rt.send_to(child_b, Ping { reply_to: *inbox.addr() }).unwrap();
-    rt.tick(); // child_b panics
-    for _ in 0..8 { rt.tick(); }
-
-    // All 3 children should be alive
-    let stats = rt.stats();
-    assert_eq!(stats.workers[0].num_actors, 4); // sup + 3 children
-
-    // child_a should be UNCHANGED
-    let new_a = rt.where_is("rfo_a").expect("rfo_a still exists");
-    assert_eq!(old_a, new_a, "child_a should not be restarted in RestForOne when b fails");
-
-    // child_c should have a NEW address
-    let new_c = rt.where_is("rfo_c").expect("rfo_c re-registered");
-    assert_ne!(old_c, new_c, "child_c should have a new address after RestForOne restart");
-}
-
-/// Given a OneForAll supervisor, when the last child of the failed set confirms death,
-/// all children are restarted in spec order.
-#[test]
-fn supervisor_one_for_all_waits_for_all_downs_before_restart() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForAll,
-        5,
-        vec![
-            ChildSpec::new("x", RestartPolicy::Permanent, |ctx| ctx.spawn(PingPongActor)),
-            ChildSpec::new("y", RestartPolicy::Permanent, |ctx| ctx.spawn(PingPongActor)),
-        ],
-    );
-    let sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); rt.tick(); // startup
-
-    assert_eq!(rt.stats().workers[0].num_actors, 3); // sup + 2 children
-
-    // Stop one child
-    let actors: Vec<_> = rt.stats().actors.iter()
-        .filter(|(addr, _)| *addr != sup_addr)
-        .map(|(addr, _)| *addr)
-        .collect();
-    rt.stop_actor(actors[0]).unwrap();
-
-    // Tick enough times for full cycle
-    for _ in 0..10 { rt.tick(); }
-
-    // Should have supervisor + 2 new children
-    assert_eq!(rt.stats().workers[0].num_actors, 3);
-}
-
-/// Given a supervisor that stops, its children also stop.
-#[test]
-fn supervisor_on_stop_kills_children() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    let sup = Supervisor::new(
-        SupervisorStrategy::OneForOne,
-        5,
-        vec![
-            ChildSpec::new("a", RestartPolicy::Permanent, |ctx| ctx.spawn(PingPongActor)),
-            ChildSpec::new("b", RestartPolicy::Permanent, |ctx| ctx.spawn(PingPongActor)),
-        ],
-    );
-    let sup_addr = rt.spawn(sup).unwrap();
-    rt.tick(); rt.tick(); // start up
-
-    assert_eq!(rt.stats().workers[0].num_actors, 3); // sup + 2 children
-
-    // Stop the supervisor
-    rt.stop_actor(sup_addr).unwrap();
-    rt.tick(); // StopSignal delivered to supervisor, on_stop sends stop to children
-    rt.tick(); // supervisor cleaned up, stop signals delivered to children
-    rt.tick(); // children stop
-    rt.tick(); // children cleaned up
-
-    assert_eq!(rt.stats().workers[0].num_actors, 0);
-}
-
-// ── Router tests ─────────────────────────────────────────────────────────────
-
-#[test]
-fn router_round_robin_distributes_across_workers() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let collected = Arc::new(std::sync::Mutex::new(Vec::new()));
-
-    struct Collector(Arc<std::sync::Mutex<Vec<(ActorAddress, usize)>>>);
-    #[derive(Clone)]
-    struct Work(usize);
-    impl ActorInterface for Collector {
-        type Incoming = Work;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, msg: Work) {
-            self.0.lock().unwrap().push((ctx.self_addr(), msg.0));
-        }
-    }
-
-    let c = collected.clone();
-    let router = Router::<Work>::new(
-        RoutingStrategy::RoundRobin,
-        3,
-        move |ctx| ctx.spawn(Collector(c.clone())),
-        10,
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick(); // on_start spawns 3 workers
-
-    for i in 0..6 {
-        rt.send_to(router_addr, Work(i)).unwrap();
-    }
-    rt.tick(); // router receives 6 Work messages, forwards to workers
-    rt.tick(); // workers process their messages
-
-    let data = collected.lock().unwrap();
-    assert_eq!(data.len(), 6);
-
-    // Count how many unique workers received messages
-    let mut per_worker = std::collections::HashMap::new();
-    for (addr, _) in data.iter() {
-        *per_worker.entry(*addr).or_insert(0usize) += 1;
-    }
-    // All 3 workers should have received exactly 2 messages each
-    assert_eq!(per_worker.len(), 3);
-    for count in per_worker.values() {
-        assert_eq!(*count, 2);
-    }
-}
-
-#[test]
-fn router_broadcast_sends_to_all_workers() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let count = Arc::new(AtomicUsize::new(0));
-
-    struct Counter(Arc<AtomicUsize>);
-    #[derive(Clone)]
-    struct Ping;
-    impl ActorInterface for Counter {
-        type Incoming = Ping;
-        type Response = ();
-        fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {
-            self.0.fetch_add(1, Ordering::Relaxed);
-        }
-    }
-
-    let c = count.clone();
-    let router = Router::<Ping>::new(
-        RoutingStrategy::Broadcast,
-        3,
-        move |ctx| ctx.spawn(Counter(c.clone())),
-        10,
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick(); // on_start spawns workers
-
-    rt.send_to(router_addr, Ping).unwrap();
-    rt.tick(); // router broadcasts
-    rt.tick(); // workers process
-
-    assert_eq!(count.load(Ordering::Relaxed), 3);
-}
-
-#[test]
-fn router_random_delivers_to_some_worker() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let collected = Arc::new(std::sync::Mutex::new(Vec::new()));
-
-    struct Collector(Arc<std::sync::Mutex<Vec<ActorAddress>>>);
-    #[derive(Clone)]
-    struct Work;
-    impl ActorInterface for Collector {
-        type Incoming = Work;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, _msg: Work) {
-            self.0.lock().unwrap().push(ctx.self_addr());
-        }
-    }
-
-    let c = collected.clone();
-    let router = Router::<Work>::new(
-        RoutingStrategy::Random,
-        3,
-        move |ctx| ctx.spawn(Collector(c.clone())),
-        10,
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick();
-
-    for _ in 0..30 {
-        rt.send_to(router_addr, Work).unwrap();
-    }
-    rt.tick();
-    rt.tick();
-
-    let data = collected.lock().unwrap();
-    assert_eq!(data.len(), 30);
-
-    let unique: std::collections::HashSet<_> = data.iter().collect();
-    assert!(unique.len() >= 2, "expected at least 2 workers used, got {}", unique.len());
-}
-
-#[test]
-fn router_replaces_dead_worker() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let spawn_count = Arc::new(AtomicUsize::new(0));
-
-    struct PanicOnFirst {
-        first: bool,
-    }
-    #[derive(Clone)]
-    struct Work;
-    impl ActorInterface for PanicOnFirst {
-        type Incoming = Work;
-        type Response = ();
-        fn handle(&mut self, _ctx: &Ctx, _msg: Work) {
-            if self.first {
-                self.first = false;
-                panic!("first message panic");
-            }
-        }
-    }
-
-    let sc = spawn_count.clone();
-    let router = Router::<Work>::new(
-        RoutingStrategy::RoundRobin,
-        3,
-        move |ctx| {
-            let n = sc.fetch_add(1, Ordering::Relaxed);
-            ctx.spawn(PanicOnFirst { first: n == 0 })
-        },
-        10,
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick(); // spawn workers (3 spawned)
-    assert_eq!(spawn_count.load(Ordering::Relaxed), 3);
-
-    // Send a message that will hit worker 0
-    rt.send_to(router_addr, Work).unwrap();
-    rt.tick(); // router forwards to worker 0
-    rt.tick(); // worker 0 panics
-    rt.tick(); // cleanup + Down delivered to router
-    rt.tick(); // router spawns replacement
-    rt.tick(); // replacement starts
-
-    // Should have spawned 4 total (3 original + 1 replacement)
-    assert_eq!(spawn_count.load(Ordering::Relaxed), 4);
-
-    // Verify all 3 slots are live
-    assert_eq!(rt.stats().workers[0].num_actors, 4);
-}
-
-#[test]
-fn router_meltdown_after_max_restarts() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    struct AlwaysPanics;
-    #[derive(Clone)]
-    struct Work;
-    impl ActorInterface for AlwaysPanics {
-        type Incoming = Work;
-        type Response = ();
-        fn handle(&mut self, _ctx: &Ctx, _msg: Work) {
-            panic!("always");
-        }
-    }
-
-    let router = Router::<Work>::new(
-        RoutingStrategy::RoundRobin,
-        1,
-        |ctx| ctx.spawn(AlwaysPanics),
-        2, // max 2 restarts
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick(); // on_start
-
-    // Kill the worker 3 times (> max_restarts=2)
-    for _ in 0..3 {
-        rt.send_to(router_addr, Work).unwrap();
-        for _ in 0..5 {
-            rt.tick();
-        }
-    }
-
-    // After 3 restarts, router should have shut down
-    for _ in 0..5 {
-        rt.tick();
-    }
-    assert_eq!(rt.stats().workers[0].num_actors, 0);
-}
-
-#[test]
-fn router_on_stop_kills_workers() {
-    let rt = std_runtime(RuntimeConfig::default());
-
-    struct Dummy;
-    #[derive(Clone)]
-    struct Work;
-    impl ActorInterface for Dummy {
-        type Incoming = Work;
-        type Response = ();
-        fn handle(&mut self, _ctx: &Ctx, _msg: Work) {}
-    }
-
-    let router = Router::<Work>::new(
-        RoutingStrategy::RoundRobin,
-        3,
-        |ctx| ctx.spawn(Dummy),
-        10,
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick(); // on_start
-    assert_eq!(rt.stats().workers[0].num_actors, 4); // router + 3 workers
-
-    rt.stop_actor(router_addr).unwrap();
-    for _ in 0..5 {
-        rt.tick();
-    }
-
-    assert_eq!(rt.stats().workers[0].num_actors, 0);
-}
-
-#[test]
-fn router_broadcast_multiple_messages_all_received() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let total = Arc::new(AtomicUsize::new(0));
-
-    struct Sink(Arc<AtomicUsize>);
-    #[derive(Clone)]
-    struct Tick;
-    impl ActorInterface for Sink {
-        type Incoming = Tick;
-        type Response = ();
-        fn handle(&mut self, _ctx: &Ctx, _msg: Tick) {
-            self.0.fetch_add(1, Ordering::Relaxed);
-        }
-    }
-
-    let t = total.clone();
-    let router = Router::<Tick>::new(
-        RoutingStrategy::Broadcast,
-        3,
-        move |ctx| ctx.spawn(Sink(t.clone())),
-        10,
-    );
-    let router_addr = rt.spawn(router).unwrap();
-    rt.tick();
-
-    for _ in 0..5 {
-        rt.send_to(router_addr, Tick).unwrap();
-    }
-    rt.tick(); // router broadcasts
-    rt.tick(); // workers process
-
-    assert_eq!(total.load(Ordering::Relaxed), 15);
-}
--- a/tests/runtime_timers.rs
+++ b/tests/runtime_timers.rs
@ -1,213 +0,0 @@
-mod common;
-use common::*;
-
-// ── Timer Helpers ─────────────────────────────────────────────────────────
-
-/// Actor that schedules a one-shot timer in on_start: sends a Ping to target after N ticks.
-struct TimerStartActor {
-    target: ActorAddress,
-    delay_ticks: u64,
-}
-
-impl ActorInterface for TimerStartActor {
-    type Incoming = Ping;
-    type Response = Pong;
-
-    fn on_start(&mut self, ctx: &Ctx) {
-        ctx.send_after_ticks(self.target, Ping { reply_to: ctx.self_addr() }, self.delay_ticks);
-    }
-
-    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {}
-}
-
-/// Actor that schedules a one-shot timer when it receives a Forward message.
-struct DelayPingPongActor;
-
-impl ActorInterface for DelayPingPongActor {
-    type Incoming = Forward;
-    type Response = Done;
-
-    fn handle(&mut self, ctx: &Ctx, msg: Forward) {
-        ctx.send_after_ticks(msg.reply_to, Done(msg.value), 3);
-    }
-}
-
-/// Actor that schedules an interval timer on start: sends Ping every N ticks.
-struct HeartbeatActor {
-    target: ActorAddress,
-    period: u64,
-}
-
-impl ActorInterface for HeartbeatActor {
-    type Incoming = Ping;
-    type Response = Pong;
-
-    fn on_start(&mut self, ctx: &Ctx) {
-        ctx.send_interval_ticks(self.target, Ping { reply_to: ctx.self_addr() }, self.period);
-    }
-
-    fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {}
-}
-
-// ── Timer Tests ───────────────────────────────────────────────────────────
-
-/// Given an actor that schedules a one-shot timer in on_start,
-/// when enough ticks pass,
-/// then the timer message is delivered to the target.
-#[test]
-fn one_shot_timer_fires_after_n_ticks() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Ping>().unwrap();
-
-    let _timer_actor = rt.spawn(TimerStartActor {
-        target: *inbox.addr(),
-        delay_ticks: 3,
-    }).unwrap();
-
-    // Tick 1: on_start schedules timer (fire_at = current_tick + 3 = 4)
-    rt.tick(); // tick 1: on_start, timer scheduled
-    assert!(inbox.try_recv().is_none(), "no delivery before delay");
-
-    rt.tick(); // tick 2
-    assert!(inbox.try_recv().is_none(), "no delivery on tick 2");
-
-    rt.tick(); // tick 3
-    assert!(inbox.try_recv().is_none(), "no delivery on tick 3");
-
-    rt.tick(); // tick 4: timer fires
-    let msg = inbox.try_recv();
-    assert!(msg.is_some(), "timer message delivered after 3-tick delay");
-}
-
-/// Given an actor that schedules a one-shot timer from a message handler,
-/// when enough ticks pass after the triggering message,
-/// then the delayed response arrives.
-#[test]
-fn handler_can_schedule_one_shot_timer() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Done>().unwrap();
-
-    let addr = rt.spawn(DelayPingPongActor).unwrap();
-
-    let _ = rt.send_to(addr, Forward { value: 42, reply_to: *inbox.addr() });
-    rt.tick(); // process Forward, schedule timer (delay=3)
-
-    assert!(inbox.try_recv().is_none(), "no immediate reply");
-
-    rt.tick(); // tick 2
-    rt.tick(); // tick 3
-    assert!(inbox.try_recv().is_none(), "not yet");
-
-    rt.tick(); // tick 4: timer fires
-    let reply = inbox.try_recv();
-    assert_eq!(reply, Some(Done(42)), "delayed reply arrives after 3 ticks");
-}
-
-/// Given a one-shot timer,
-/// when it fires,
-/// then it does NOT fire again on subsequent ticks (consumed).
-#[test]
-fn one_shot_timer_fires_only_once() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Ping>().unwrap();
-
-    let _timer_actor = rt.spawn(TimerStartActor {
-        target: *inbox.addr(),
-        delay_ticks: 1,
-    }).unwrap();
-
-    rt.tick(); // on_start schedules timer
-    rt.tick(); // timer fires
-    assert!(inbox.try_recv().is_some(), "first fire");
-
-    // Subsequent ticks should NOT fire again
-    for _ in 0..5 { rt.tick(); }
-    assert!(inbox.try_recv().is_none(), "one-shot does not repeat");
-}
-
-/// Given an interval timer with period 2,
-/// when multiple ticks pass,
-/// then the timer fires repeatedly every 2 ticks.
-#[test]
-fn interval_timer_fires_repeatedly() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Ping>().unwrap();
-
-    let _heartbeat = rt.spawn(HeartbeatActor {
-        target: *inbox.addr(),
-        period: 2,
-    }).unwrap();
-
-    rt.tick(); // tick 1: on_start, interval scheduled (next_fire = current + 2 = 3)
-    assert!(inbox.try_recv().is_none(), "no fire on tick 1");
-
-    rt.tick(); // tick 2
-    assert!(inbox.try_recv().is_none(), "no fire on tick 2");
-
-    rt.tick(); // tick 3: first fire
-    assert!(inbox.try_recv().is_some(), "fire on tick 3");
-
-    rt.tick(); // tick 4
-    assert!(inbox.try_recv().is_none(), "no fire on tick 4");
-
-    rt.tick(); // tick 5: second fire
-    assert!(inbox.try_recv().is_some(), "fire on tick 5");
-
-    rt.tick(); // tick 6
-    assert!(inbox.try_recv().is_none(), "no fire on tick 6");
-
-    rt.tick(); // tick 7: third fire
-    assert!(inbox.try_recv().is_some(), "fire on tick 7");
-}
-
-/// Given an interval timer targeting an actor that gets stopped,
-/// when the actor is removed,
-/// then the interval timer is cleaned up (no orphan timers).
-#[test]
-fn interval_timer_cleaned_up_when_actor_dies() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let _inbox = rt.new_inbox::<Ping>().unwrap();
-
-    // Heartbeat sends to a counter that we'll kill
-    let counter_addr = rt.spawn(CounterActor { count: 0 }).unwrap();
-
-    // HeartbeatActor sends Ping to counter every tick
-    let _hb = rt.spawn(HeartbeatActor {
-        target: counter_addr,
-        period: 1,
-    }).unwrap();
-
-    // Let it run a few ticks
-    for _ in 0..3 { rt.tick(); }
-
-    // Stop the counter
-    rt.stop_actor(counter_addr).unwrap();
-    for _ in 0..5 { rt.tick(); }
-
-    // Counter is gone, interval timer should be GC'd.
-    let stats = rt.stats();
-    // Only the heartbeat actor should remain
-    assert_eq!(stats.workers[0].num_actors, 1);
-}
-
-/// Given a timer with delay 0,
-/// when the next tick fires,
-/// then the message is delivered immediately on the next tick.
-#[test]
-fn timer_with_zero_delay_fires_next_tick() {
-    let rt = std_runtime(RuntimeConfig::default());
-    let inbox = rt.new_inbox::<Ping>().unwrap();
-
-    let _timer_actor = rt.spawn(TimerStartActor {
-        target: *inbox.addr(),
-        delay_ticks: 0,
-    }).unwrap();
-
-    rt.tick(); // on_start schedules timer with delay=0
-    // Timer requests are processed after tick_all (phase 5.5)
-    // Timer fires on the NEXT tick (phase 2.5)
-    assert!(inbox.try_recv().is_none(), "not yet -- timer fires next tick");
-
-    rt.tick(); // timer fires
-    assert!(inbox.try_recv().is_some(), "zero-delay timer fires on next tick");
-}
--- a/tests/std_extension.rs
+++ b/tests/std_extension.rs
@ -0,0 +1,742 @@
+//! StdExtension Tests — higher-level patterns from swactor-std.
+//!
+//! Covers: naming registry, groups/pub-sub, ask pattern, supervision
+//! strategies and restart policies, and router work distribution.
+
+mod common;
+use common::*;
+
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::Arc;
+
+// ── Local actors ────────────────────────────────────────────────────────────
+
+/// Looks up a peer by name using ctx.where_is().
+struct NameLookupActor {
+    target_name: &'static str,
+    reply_to: ActorAddress,
+}
+
+impl ActorInterface for NameLookupActor {
+    type Incoming = Ping;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
+        if let Some(peer) = ctx.where_is(self.target_name) {
+            ctx.send(self.reply_to, MyAddr(peer)).unwrap();
+        }
+    }
+}
+
+/// Spawns a named child from a handler.
+struct NamedSpawnerActor {
+    reply_to: ActorAddress,
+}
+
+impl ActorInterface for NamedSpawnerActor {
+    type Incoming = Ping;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
+        if let Ok(addr) = ctx.spawn_named("child", PingPongActor) {
+            ctx.send(self.reply_to, MyAddr(addr)).unwrap();
+        }
+    }
+}
+
+/// Panics after `trigger` messages.
+struct PanicAfterN {
+    trigger: usize,
+    count: usize,
+    counter: Arc<AtomicUsize>,
+}
+
+impl ActorInterface for PanicAfterN {
+    type Incoming = Ping;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, msg: Ping) {
+        self.count += 1;
+        self.counter.fetch_add(1, Ordering::SeqCst);
+        let _ = ctx.send(msg.reply_to, Pong);
+        if self.count >= self.trigger {
+            panic!("intentional panic at message {}", self.count);
+        }
+    }
+}
+
+/// Stops itself on first message.
+struct StopsAfterFirst;
+impl ActorInterface for StopsAfterFirst {
+    type Incoming = Ping;
+    type Response = ();
+    fn handle(&mut self, ctx: &Ctx, _msg: Ping) {
+        ctx.stop_self();
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Naming Registry
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Full naming lifecycle: register, lookup, send, duplicate fails, auto-unregister
+/// on stop and panic, name reuse, registered_names list, manual unregister.
+#[test]
+fn naming_registry_lifecycle() {
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+
+    // Register "alice", lookup, send Ping → Pong
+    let alice = rt.spawn_named("alice", PingPongActor).unwrap();
+    assert_eq!(rt.where_is("alice"), Some(alice));
+    rt.send_to(alice, Ping { reply_to: *inbox.addr() }).unwrap();
+    rt.tick();
+    assert!(inbox.try_recv().is_some(), "named actor processes messages");
+
+    // Duplicate fails, original binding preserved
+    assert!(rt.spawn_named("alice", PingPongActor).is_err());
+    assert_eq!(rt.where_is("alice"), Some(alice));
+
+    // Unknown name → None
+    assert_eq!(rt.where_is("ghost"), None);
+
+    // Stop "alice" → name freed
+    rt.stop_actor(alice).unwrap();
+    rt.tick();
+    assert_eq!(rt.where_is("alice"), None, "name freed after stop");
+
+    // Reuse the name
+    let alice2 = rt.spawn_named("alice", PingPongActor).unwrap();
+    assert_ne!(alice, alice2);
+    assert_eq!(rt.where_is("alice"), Some(alice2));
+
+    // Panic also frees the name
+    let bob = rt.spawn_named("bob", PanicActor).unwrap();
+    rt.tick();
+    rt.send_to(bob, PanicMsg).unwrap();
+    rt.tick();
+    assert_eq!(rt.where_is("bob"), None, "name freed after panic");
+    let _bob2 = rt.spawn_named("bob", PingPongActor).unwrap();
+    assert!(rt.where_is("bob").is_some());
+
+    // registered_names enumerates all
+    rt.spawn_named("gamma", PingPongActor).unwrap();
+    let mut names = rt.registered_names();
+    names.sort();
+    assert!(names.contains(&"alice".to_string()));
+    assert!(names.contains(&"bob".to_string()));
+    assert!(names.contains(&"gamma".to_string()));
+
+    // Manual unregister: name freed but actor lives
+    let charlie_inbox = rt.new_inbox::<Pong>().unwrap();
+    let charlie = rt.spawn_named("charlie", PingPongActor).unwrap();
+    rt.tick();
+    let removed = rt.unregister("charlie");
+    assert_eq!(removed, Some(charlie));
+    assert_eq!(rt.where_is("charlie"), None, "name freed by unregister");
+    rt.send_to(charlie, Ping { reply_to: *charlie_inbox.addr() }).unwrap();
+    rt.tick();
+    assert!(charlie_inbox.try_recv().is_some(), "actor still alive after name unregistered");
+}
+
+/// Actors resolve and register names from handlers using ctx.
+#[test]
+fn naming_from_actor_handlers() {
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<MyAddr>().unwrap();
+
+    // ctx.where_is from handler
+    let target = rt.spawn_named("target", PingPongActor).unwrap();
+    let looker = rt.spawn(NameLookupActor {
+        target_name: "target",
+        reply_to: *inbox.addr(),
+    }).unwrap();
+    rt.tick();
+    rt.send_to(looker, Ping { reply_to: ActorAddress::default() }).unwrap();
+    tick_n(&rt, 3);
+    assert_eq!(inbox.try_recv(), Some(MyAddr(target)), "ctx.where_is resolves");
+
+    // ctx.spawn_named from handler
+    let spawner = rt.spawn(NamedSpawnerActor { reply_to: *inbox.addr() }).unwrap();
+    rt.tick();
+    rt.send_to(spawner, Ping { reply_to: ActorAddress::default() }).unwrap();
+    tick_n(&rt, 3);
+    let child_addr = inbox.try_recv().expect("child address returned");
+    assert_eq!(rt.where_is("child"), Some(child_addr.0), "name registered from handler");
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Groups / Pub-Sub
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Full groups lifecycle: join, publish broadcasts, leave stops delivery,
+/// dead actor auto-removed, multi-group cleanup, empty group deleted,
+/// join and publish from handlers.
+#[test]
+fn groups_pub_sub_lifecycle() {
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+
+    // Join 3 actors, publish → all 3 get it
+    let a = rt.spawn(PingPongActor).unwrap();
+    let b = rt.spawn(PingPongActor).unwrap();
+    let c = rt.spawn(PingPongActor).unwrap();
+    rt.join_group(a, "workers");
+    rt.join_group(b, "workers");
+    rt.join_group(c, "workers");
+    rt.tick();
+
+    let count = rt.publish_to("workers", Ping { reply_to: *inbox.addr() });
+    assert_eq!(count, 3, "3 members, 3 messages sent");
+    rt.tick();
+    let mut pongs = 0;
+    while inbox.try_recv().is_some() { pongs += 1; }
+    assert_eq!(pongs, 3, "all 3 received");
+
+    // Leave stops delivery
+    rt.leave_group(c, "workers");
+    let count = rt.publish_to("workers", Ping { reply_to: *inbox.addr() });
+    assert_eq!(count, 2, "2 after leave");
+    rt.tick();
+    let mut pongs = 0;
+    while inbox.try_recv().is_some() { pongs += 1; }
+    assert_eq!(pongs, 2);
+
+    // Dead actor auto-removed
+    rt.stop_actor(b).unwrap();
+    rt.tick();
+    let count = rt.publish_to("workers", Ping { reply_to: *inbox.addr() });
+    assert_eq!(count, 1, "dead actor removed");
+
+    // Multi-group cleanup: actor in alpha/beta/gamma dies → all cleaned
+    let rt = std_runtime(RuntimeConfig::default());
+    let actor = rt.spawn(PingPongActor).unwrap();
+    rt.join_group(actor, "alpha");
+    rt.join_group(actor, "beta");
+    rt.join_group(actor, "gamma");
+    rt.tick();
+    rt.stop_actor(actor).unwrap();
+    rt.tick();
+    assert!(rt.group_members("alpha").is_empty());
+    assert!(rt.group_members("beta").is_empty());
+    assert!(rt.group_members("gamma").is_empty());
+
+    // Empty group auto-deleted
+    let rt = std_runtime(RuntimeConfig::default());
+    let actor = rt.spawn(PingPongActor).unwrap();
+    rt.join_group(actor, "temp");
+    assert!(rt.groups().contains(&"temp".to_string()));
+    rt.leave_group(actor, "temp");
+    assert!(!rt.groups().contains(&"temp".to_string()), "empty group removed");
+
+    // Empty group query
+    let rt = std_runtime(RuntimeConfig::default());
+    assert!(rt.group_members("nonexistent").is_empty());
+
+    // ctx.join_group from on_start
+    struct GroupJoiner;
+    impl ActorInterface for GroupJoiner {
+        type Incoming = Ping;
+        type Response = ();
+        fn on_start(&mut self, ctx: &Ctx) {
+            ctx.join_group("auto-joined");
+        }
+        fn handle(&mut self, _ctx: &Ctx, _msg: Ping) {}
+    }
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let x = rt.spawn(GroupJoiner).unwrap();
+    let y = rt.spawn(GroupJoiner).unwrap();
+    rt.tick();
+    let members = rt.group_members("auto-joined");
+    assert_eq!(members.len(), 2);
+    assert!(members.contains(&x));
+    assert!(members.contains(&y));
+
+    // ctx.publish from handler
+    #[derive(Clone)]
+    struct BroadcastCmd { reply_to: ActorAddress }
+
+    struct Broadcaster;
+    impl ActorInterface for Broadcaster {
+        type Incoming = BroadcastCmd;
+        type Response = ();
+        fn on_start(&mut self, ctx: &Ctx) {
+            ctx.join_group("bcast");
+        }
+        fn handle(&mut self, ctx: &Ctx, msg: BroadcastCmd) {
+            ctx.publish("bcast", Ping { reply_to: msg.reply_to });
+        }
+    }
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let p1 = rt.spawn(PingPongActor).unwrap();
+    let p2 = rt.spawn(PingPongActor).unwrap();
+    rt.join_group(p1, "bcast");
+    rt.join_group(p2, "bcast");
+    let broadcaster = rt.spawn(Broadcaster).unwrap();
+    rt.tick();
+    rt.send_to(broadcaster, BroadcastCmd { reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 3);
+    let mut pongs = 0;
+    while inbox.try_recv().is_some() { pongs += 1; }
+    assert!(pongs >= 2, "at least 2 PingPong members replied, got {pongs}");
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Ask Pattern
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Ask pattern: basic ask, repeated asks track state, try_recv before/after
+/// tick, dead actor times out.
+#[test]
+fn ask_pattern() {
+    let rt = std_runtime(RuntimeConfig::default());
+
+    // Basic ask
+    let actor = rt.spawn(PingPongActor).unwrap();
+    rt.tick();
+    let pong: Pong = rt.ask(actor, |reply_to| Ping { reply_to })
+        .unwrap().recv_ticking(&rt, 10).unwrap();
+    assert_eq!(pong, Pong);
+
+    // Repeated asks track state
+    let counter = rt.spawn(CounterActor { count: 0 }).unwrap();
+    rt.tick();
+    let c1: Count = rt.ask(counter, |reply_to| Increment { reply_to })
+        .unwrap().recv_ticking(&rt, 10).unwrap();
+    let c2: Count = rt.ask(counter, |reply_to| Increment { reply_to })
+        .unwrap().recv_ticking(&rt, 10).unwrap();
+    let c3: Count = rt.ask(counter, |reply_to| Increment { reply_to })
+        .unwrap().recv_ticking(&rt, 10).unwrap();
+    assert_eq!((c1, c2, c3), (Count(1), Count(2), Count(3)));
+
+    // try_recv: None before tick, Some after
+    let rt = std_runtime(RuntimeConfig::default());
+    let actor = rt.spawn(PingPongActor).unwrap();
+    rt.tick();
+    let ask = rt.ask::<Ping, Pong>(actor, |reply_to| Ping { reply_to }).unwrap();
+    assert!(ask.try_recv().is_none(), "no response before tick");
+    rt.tick();
+    assert_eq!(ask.try_recv(), Some(Pong));
+
+    // Dead actor → timeout
+    let rt = std_runtime(RuntimeConfig::default());
+    let actor = rt.spawn(PingPongActor).unwrap();
+    rt.tick();
+    rt.stop_actor(actor).unwrap();
+    rt.tick();
+    if let Ok(ask) = rt.ask::<Ping, Pong>(actor, |reply_to| Ping { reply_to }) {
+        assert!(ask.recv_ticking(&rt, 5).is_err(), "timeout with dead actor");
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Supervision
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Restart policies: permanent always restarts, transient only on panic,
+/// temporary never restarts, meltdown after max_restarts.
+#[test]
+fn supervision_restart_policies() {
+    // Permanent child panics → restarted
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter = Arc::new(AtomicUsize::new(0));
+    let counter_c = counter.clone();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 5,
+        vec![ChildSpec::new("worker", RestartPolicy::Permanent, move |ctx| {
+            ctx.spawn(PanicAfterN { trigger: 2, count: 0, counter: counter_c.clone() })
+        })],
+    );
+    let sup_addr = rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let child = rt.stats().actors.iter()
+        .find(|(a, _)| *a != sup_addr).map(|(a, _)| *a).unwrap();
+    rt.send_to(child, Ping { reply_to: *inbox.addr() }).unwrap();
+    rt.tick();
+    assert_eq!(counter.load(Ordering::SeqCst), 1);
+    rt.send_to(child, Ping { reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 5); // panics, supervisor restarts
+    assert_eq!(rt.stats().workers[0].num_actors, 2, "supervisor + restarted child");
+
+    // Transient stops normally → NOT restarted
+    let rt = std_runtime(RuntimeConfig::default());
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 5,
+        vec![ChildSpec::new("worker", RestartPolicy::Transient, |ctx| {
+            ctx.spawn(StopsAfterFirst)
+        })],
+    );
+    let sup_addr = rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let child = rt.stats().actors.iter()
+        .find(|(a, _)| *a != sup_addr).map(|(a, _)| *a).unwrap();
+    rt.send_to(child, Ping { reply_to: ActorAddress::default() }).unwrap();
+    tick_n(&rt, 4);
+    assert_eq!(rt.stats().workers[0].num_actors, 1, "transient+normal → no restart");
+
+    // Transient panics → restarted
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter = Arc::new(AtomicUsize::new(0));
+    let counter_c = counter.clone();
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 5,
+        vec![ChildSpec::new("worker", RestartPolicy::Transient, move |ctx| {
+            ctx.spawn(PanicAfterN { trigger: 1, count: 0, counter: counter_c.clone() })
+        })],
+    );
+    let sup_addr = rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let child = rt.stats().actors.iter()
+        .find(|(a, _)| *a != sup_addr).map(|(a, _)| *a).unwrap();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    rt.send_to(child, Ping { reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(rt.stats().workers[0].num_actors, 2, "transient+panic → restarted");
+
+    // Temporary never restarts
+    let rt = std_runtime(RuntimeConfig::default());
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 5,
+        vec![ChildSpec::new("worker", RestartPolicy::Temporary, |ctx| ctx.spawn(PanicActor))],
+    );
+    let sup_addr = rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let child = rt.stats().actors.iter()
+        .find(|(a, _)| *a != sup_addr).map(|(a, _)| *a).unwrap();
+    rt.send_to(child, PanicMsg).unwrap();
+    tick_n(&rt, 4);
+    assert_eq!(rt.stats().workers[0].num_actors, 1, "temporary → no restart");
+
+    // Meltdown: max_restarts=2, crash 3 times → supervisor stops
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter = Arc::new(AtomicUsize::new(0));
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 2,
+        vec![ChildSpec::new("crasher", RestartPolicy::Permanent, {
+            let c = counter.clone();
+            move |ctx| ctx.spawn(PanicAfterN { trigger: 1, count: 0, counter: c.clone() })
+        })],
+    );
+    let sup_addr = rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    for _ in 0..3 {
+        if let Some((child, _)) = rt.stats().actors.iter()
+            .find(|(a, _)| *a != sup_addr)
+        {
+            let inbox = rt.new_inbox::<Pong>().unwrap();
+            let _ = rt.send_to(*child, Ping { reply_to: *inbox.addr() });
+            tick_n(&rt, 5);
+        }
+    }
+    let sup_alive = rt.stats().actors.iter().any(|(a, _)| *a == sup_addr);
+    assert!(!sup_alive, "supervisor stopped after exceeding max_restarts");
+}
+
+/// Strategies: OneForOne, OneForAll, RestForOne. Stopping supervisor kills children.
+#[test]
+fn supervision_strategies() {
+    // OneForOne: only failed child restarted
+    let rt = std_runtime(RuntimeConfig::default());
+    let counter_a = Arc::new(AtomicUsize::new(0));
+    let counter_b = Arc::new(AtomicUsize::new(0));
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 5,
+        vec![
+            ChildSpec::new("crasher", RestartPolicy::Permanent, {
+                let c = counter_a.clone();
+                move |ctx| ctx.spawn_named("ofo_a", PanicAfterN {
+                    trigger: 1, count: 0, counter: c.clone(),
+                })
+            }),
+            ChildSpec::new("stable", RestartPolicy::Permanent, {
+                let c = counter_b.clone();
+                move |ctx| ctx.spawn_named("ofo_b", CountingPingActor { counter: c.clone() })
+            }),
+        ],
+    );
+    rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let child_a = rt.where_is("ofo_a").unwrap();
+    let child_b = rt.where_is("ofo_b").unwrap();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    rt.send_to(child_a, Ping { reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 5);
+    let child_b_after = rt.where_is("ofo_b").unwrap();
+    assert_eq!(child_b, child_b_after, "child_b unchanged in OneForOne");
+    rt.send_to(child_b, Ping { reply_to: *inbox.addr() }).unwrap();
+    rt.tick();
+    assert!(counter_b.load(Ordering::SeqCst) >= 1, "child_b still processing");
+
+    // OneForAll: all children restarted
+    let rt = std_runtime(RuntimeConfig::default());
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForAll, 5,
+        vec![
+            ChildSpec::new("a", RestartPolicy::Permanent, {
+                let c = Arc::new(AtomicUsize::new(0));
+                move |ctx| ctx.spawn_named("ofa_a", PanicAfterN {
+                    trigger: 1, count: 0, counter: c.clone(),
+                })
+            }),
+            ChildSpec::new("b", RestartPolicy::Permanent, {
+                let c = Arc::new(AtomicUsize::new(0));
+                move |ctx| ctx.spawn_named("ofa_b", CountingPingActor { counter: c.clone() })
+            }),
+        ],
+    );
+    rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let old_b = rt.where_is("ofa_b").unwrap();
+    let child_a = rt.where_is("ofa_a").unwrap();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    rt.send_to(child_a, Ping { reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 8);
+    let new_b = rt.where_is("ofa_b").expect("ofa_b re-registered");
+    assert_ne!(old_b, new_b, "child_b restarted in OneForAll");
+
+    // RestForOne: failed child + later children restarted, earlier unaffected
+    let rt = std_runtime(RuntimeConfig::default());
+    let sup = Supervisor::new(
+        SupervisorStrategy::RestForOne, 5,
+        vec![
+            ChildSpec::new("a", RestartPolicy::Permanent, {
+                let c = Arc::new(AtomicUsize::new(0));
+                move |ctx| ctx.spawn_named("rfo_a", CountingPingActor { counter: c.clone() })
+            }),
+            ChildSpec::new("b", RestartPolicy::Permanent, {
+                let c = Arc::new(AtomicUsize::new(0));
+                move |ctx| ctx.spawn_named("rfo_b", PanicAfterN {
+                    trigger: 1, count: 0, counter: c.clone(),
+                })
+            }),
+            ChildSpec::new("c", RestartPolicy::Permanent, {
+                let c = Arc::new(AtomicUsize::new(0));
+                move |ctx| ctx.spawn_named("rfo_c", CountingPingActor { counter: c.clone() })
+            }),
+        ],
+    );
+    rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    let old_a = rt.where_is("rfo_a").unwrap();
+    let old_c = rt.where_is("rfo_c").unwrap();
+    let child_b = rt.where_is("rfo_b").unwrap();
+    let inbox = rt.new_inbox::<Pong>().unwrap();
+    rt.send_to(child_b, Ping { reply_to: *inbox.addr() }).unwrap();
+    tick_n(&rt, 8);
+    let new_a = rt.where_is("rfo_a").unwrap();
+    let new_c = rt.where_is("rfo_c").expect("rfo_c re-registered");
+    assert_eq!(old_a, new_a, "child_a unchanged in RestForOne");
+    assert_ne!(old_c, new_c, "child_c restarted in RestForOne");
+
+    // Stopping supervisor kills children
+    let rt = std_runtime(RuntimeConfig::default());
+    let sup = Supervisor::new(
+        SupervisorStrategy::OneForOne, 5,
+        vec![
+            ChildSpec::new("a", RestartPolicy::Permanent, |ctx| ctx.spawn(PingPongActor)),
+            ChildSpec::new("b", RestartPolicy::Permanent, |ctx| ctx.spawn(PingPongActor)),
+        ],
+    );
+    let sup_addr = rt.spawn(sup).unwrap();
+    tick_n(&rt, 2);
+    assert_eq!(rt.stats().workers[0].num_actors, 3);
+    rt.stop_actor(sup_addr).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(rt.stats().workers[0].num_actors, 0, "stopping supervisor kills children");
+}
+
+/// handle_down dispatch and ctx.stop_actor from handler.
+#[test]
+fn handle_down_dispatch() {
+    // ctx.stop_actor from handler stops target
+    #[derive(Clone)]
+    struct StopCmd { target: ActorAddress }
+    struct Stopper;
+    impl ActorInterface for Stopper {
+        type Incoming = StopCmd;
+        type Response = ();
+        fn handle(&mut self, ctx: &Ctx, msg: StopCmd) {
+            let _ = ctx.stop_actor(msg.target);
+        }
+    }
+
+    let rt = std_runtime(RuntimeConfig::default());
+    let target = rt.spawn(PingPongActor).unwrap();
+    let stopper = rt.spawn(Stopper).unwrap();
+    rt.tick();
+    rt.send_to(stopper, StopCmd { target }).unwrap();
+    tick_n(&rt, 4);
+    assert!(rt.send_to(target, Ping { reply_to: ActorAddress::default() }).is_err(),
+        "target stopped by ctx.stop_actor");
+    assert!(rt.send_to(stopper, StopCmd { target }).is_ok(), "stopper still alive");
+}
+
+// ═══════════════════════════════════════════════════════════════════════════
+// Router
+// ═══════════════════════════════════════════════════════════════════════════
+
+/// Router distributes work: round-robin is even, broadcast hits all, random
+/// uses multiple workers. Dead workers replaced. Stop router kills workers.
+/// Meltdown after max restarts.
+#[test]
+fn router_work_distribution() {
+    // Round-robin: 3 workers, 6 msgs → 2 each
+    let rt = std_runtime(RuntimeConfig::default());
+    let collected = Arc::new(std::sync::Mutex::new(Vec::new()));
+    struct Collector(Arc<std::sync::Mutex<Vec<(ActorAddress, usize)>>>);
+    #[derive(Clone)]
+    struct Work(usize);
+    impl ActorInterface for Collector {
+        type Incoming = Work;
+        type Response = ();
+        fn handle(&mut self, ctx: &Ctx, msg: Work) {
+            self.0.lock().unwrap().push((ctx.self_addr(), msg.0));
+        }
+    }
+
+    let c = collected.clone();
+    let router = Router::<Work>::new(
+        RoutingStrategy::RoundRobin, 3,
+        move |ctx| ctx.spawn(Collector(c.clone())), 10,
+    );
+    let router_addr = rt.spawn(router).unwrap();
+    rt.tick();
+    for i in 0..6 {
+        rt.send_to(router_addr, Work(i)).unwrap();
+    }
+    tick_n(&rt, 3);
+    let data = collected.lock().unwrap();
+    assert_eq!(data.len(), 6);
+    let mut per_worker = std::collections::HashMap::new();
+    for (addr, _) in data.iter() {
+        *per_worker.entry(*addr).or_insert(0usize) += 1;
+    }
+    assert_eq!(per_worker.len(), 3, "3 distinct workers");
+    for count in per_worker.values() {
+        assert_eq!(*count, 2, "each worker gets exactly 2");
+    }
+
+    // Broadcast: 5 msgs to 3 workers → 15 total
+    let rt = std_runtime(RuntimeConfig::default());
+    let total = Arc::new(AtomicUsize::new(0));
+    struct BCounter(Arc<AtomicUsize>);
+    #[derive(Clone)]
+    struct BPing;
+    impl ActorInterface for BCounter {
+        type Incoming = BPing;
+        type Response = ();
+        fn handle(&mut self, _ctx: &Ctx, _msg: BPing) {
+            self.0.fetch_add(1, Ordering::Relaxed);
+        }
+    }
+    let t = total.clone();
+    let router = Router::<BPing>::new(
+        RoutingStrategy::Broadcast, 3,
+        move |ctx| ctx.spawn(BCounter(t.clone())), 10,
+    );
+    let router_addr = rt.spawn(router).unwrap();
+    rt.tick();
+    for _ in 0..5 {
+        rt.send_to(router_addr, BPing).unwrap();
+    }
+    tick_n(&rt, 3);
+    assert_eq!(total.load(Ordering::Relaxed), 15, "5 broadcasts × 3 workers = 15");
+
+    // Random: 30 msgs → at least 2 workers used
+    let rt = std_runtime(RuntimeConfig::default());
+    let rcollected = Arc::new(std::sync::Mutex::new(Vec::new()));
+    struct RCollector(Arc<std::sync::Mutex<Vec<ActorAddress>>>);
+    #[derive(Clone)]
+    struct RWork;
+    impl ActorInterface for RCollector {
+        type Incoming = RWork;
+        type Response = ();
+        fn handle(&mut self, ctx: &Ctx, _msg: RWork) {
+            self.0.lock().unwrap().push(ctx.self_addr());
+        }
+    }
+    let c = rcollected.clone();
+    let router = Router::<RWork>::new(
+        RoutingStrategy::Random, 3,
+        move |ctx| ctx.spawn(RCollector(c.clone())), 10,
+    );
+    let router_addr = rt.spawn(router).unwrap();
+    rt.tick();
+    for _ in 0..30 {
+        rt.send_to(router_addr, RWork).unwrap();
+    }
+    tick_n(&rt, 3);
+    let data = rcollected.lock().unwrap();
+    let unique: std::collections::HashSet<_> = data.iter().collect();
+    assert!(unique.len() >= 2, "random uses at least 2 workers");
+
+    // Dead worker replaced
+    let rt = std_runtime(RuntimeConfig::default());
+    let spawn_count = Arc::new(AtomicUsize::new(0));
+    struct PanicOnFirst { first: bool }
+    #[derive(Clone)]
+    struct DWork;
+    impl ActorInterface for PanicOnFirst {
+        type Incoming = DWork;
+        type Response = ();
+        fn handle(&mut self, _ctx: &Ctx, _msg: DWork) {
+            if self.first { self.first = false; panic!("first message panic"); }
+        }
+    }
+    let sc = spawn_count.clone();
+    let router = Router::<DWork>::new(
+        RoutingStrategy::RoundRobin, 3,
+        move |ctx| { sc.fetch_add(1, Ordering::Relaxed); ctx.spawn(PanicOnFirst { first: sc.load(Ordering::Relaxed) == 1 }) },
+        10,
+    );
+    let router_addr = rt.spawn(router).unwrap();
+    rt.tick();
+    rt.send_to(router_addr, DWork).unwrap();
+    tick_n(&rt, 5);
+    assert!(spawn_count.load(Ordering::Relaxed) >= 4, "replacement spawned");
+
+    // Meltdown: max_restarts=2
+    let rt = std_runtime(RuntimeConfig::default());
+    struct AlwaysPanics;
+    #[derive(Clone)]
+    struct MWork;
+    impl ActorInterface for AlwaysPanics {
+        type Incoming = MWork;
+        type Response = ();
+        fn handle(&mut self, _ctx: &Ctx, _msg: MWork) { panic!("always"); }
+    }
+    let router = Router::<MWork>::new(
+        RoutingStrategy::RoundRobin, 1,
+        |ctx| ctx.spawn(AlwaysPanics), 2,
+    );
+    let router_addr = rt.spawn(router).unwrap();
+    rt.tick();
+    for _ in 0..3 {
+        rt.send_to(router_addr, MWork).unwrap();
+        tick_n(&rt, 5);
+    }
+    tick_n(&rt, 5);
+    assert_eq!(rt.stats().workers[0].num_actors, 0, "router melted down");
+
+    // Stop router kills workers
+    let rt = std_runtime(RuntimeConfig::default());
+    struct Dummy;
+    #[derive(Clone)]
+    struct SWork;
+    impl ActorInterface for Dummy {
+        type Incoming = SWork;
+        type Response = ();
+        fn handle(&mut self, _ctx: &Ctx, _msg: SWork) {}
+    }
+    let router = Router::<SWork>::new(
+        RoutingStrategy::RoundRobin, 3,
+        |ctx| ctx.spawn(Dummy), 10,
+    );
+    let router_addr = rt.spawn(router).unwrap();
+    rt.tick();
+    assert_eq!(rt.stats().workers[0].num_actors, 4);
+    rt.stop_actor(router_addr).unwrap();
+    tick_n(&rt, 5);
+    assert_eq!(rt.stats().workers[0].num_actors, 0, "stop router kills workers");
+}
--- a/tests/watch_api.rs
+++ b/tests/watch_api.rs
@ -1,335 +0,0 @@
-use std::sync::atomic::{AtomicUsize, Ordering};
-use std::sync::Arc;
-
-use swactor::actor::{ActorAddress, ActorExited, ActorInterface, ExitReason};
-use swactor::runtime::{Ctx, Runtime, RuntimeConfig};
-use swactor_std::{CtxWatching, RuntimeWatching, StdExtension};
-
-// ── Actors ──────────────────────────────────────────────────────────────────
-
-/// An actor that panics when it receives PanicMsg.
-struct PanicOnCommand;
-
-#[derive(Clone)]
-struct PanicMsg;
-
-impl ActorInterface for PanicOnCommand {
-    type Incoming = PanicMsg;
-    type Response = ();
-    fn handle(&mut self, _ctx: &Ctx, _msg: PanicMsg) {
-        panic!("deliberate panic for test");
-    }
-}
-
-/// An actor that watches targets and counts exit notifications.
-struct ExitWatcher {
-    exit_count: Arc<AtomicUsize>,
-    last_reason: Arc<std::sync::Mutex<Option<ExitReason>>>,
-    last_addr: Arc<std::sync::Mutex<Option<ActorAddress>>>,
-}
-
-#[derive(Clone)]
-enum WatcherCmd {
-    WatchThis(ActorAddress),
-    UnwatchThis(ActorAddress),
-}
-
-impl ActorInterface for ExitWatcher {
-    type Incoming = WatcherCmd;
-    type Response = ();
-    fn handle(&mut self, ctx: &Ctx, msg: WatcherCmd) {
-        match msg {
-            WatcherCmd::WatchThis(target) => {
-                ctx.watch(target);
-            }
-            WatcherCmd::UnwatchThis(target) => {
-                ctx.unwatch(target);
-            }
-        }
-    }
-
-    fn on_actor_exit(&mut self, _ctx: &Ctx, exited: ActorExited) {
-        self.exit_count.fetch_add(1, Ordering::SeqCst);
-        *self.last_reason.lock().unwrap() = Some(exited.reason);
-        *self.last_addr.lock().unwrap() = Some(exited.addr);
-    }
-}
-
-impl ExitWatcher {
-    fn new() -> (Self, WatcherState) {
-        let exit_count = Arc::new(AtomicUsize::new(0));
-        let last_reason = Arc::new(std::sync::Mutex::new(None));
-        let last_addr = Arc::new(std::sync::Mutex::new(None));
-        let state = WatcherState {
-            exit_count: exit_count.clone(),
-            last_reason: last_reason.clone(),
-            last_addr: last_addr.clone(),
-        };
-        (
-            ExitWatcher {
-                exit_count,
-                last_reason,
-                last_addr,
-            },
-            state,
-        )
-    }
-}
-
-/// Shared state for inspecting what ExitWatcher observed.
-struct WatcherState {
-    exit_count: Arc<AtomicUsize>,
-    last_reason: Arc<std::sync::Mutex<Option<ExitReason>>>,
-    last_addr: Arc<std::sync::Mutex<Option<ActorAddress>>>,
-}
-
-impl WatcherState {
-    fn count(&self) -> usize {
-        self.exit_count.load(Ordering::SeqCst)
-    }
-    fn last_reason(&self) -> Option<ExitReason> {
-        self.last_reason.lock().unwrap().clone()
-    }
-    fn last_addr(&self) -> Option<ActorAddress> {
-        *self.last_addr.lock().unwrap()
-    }
-}
-
-/// A silent actor that does nothing (for targets that shouldn't panic).
-struct Sleeper;
-
-#[derive(Clone)]
-struct Noop;
-
-impl ActorInterface for Sleeper {
-    type Incoming = Noop;
-    type Response = ();
-    fn handle(&mut self, _ctx: &Ctx, _msg: Noop) {}
-}
-
-// ── Helper ──────────────────────────────────────────────────────────────────
-
-fn tick_n(rt: &Runtime, n: usize) {
-    for _ in 0..n {
-        rt.tick();
-    }
-}
-
-fn watch_config() -> RuntimeConfig {
-    RuntimeConfig {
-        num_threads: 1,
-        ..RuntimeConfig::default()
-    }
-}
-
-fn watch_runtime() -> Runtime {
-    Runtime::new(watch_config())
-        .with_extension(Arc::new(StdExtension::new()))
-}
-
-// ── Tests ───────────────────────────────────────────────────────────────────
-
-/// Given a watcher and a target actor,
-/// when the target panics,
-/// then the watcher's on_actor_exit fires with ExitReason::Panicked.
-#[test]
-fn watch_receives_notification_on_panic() {
-    let rt = watch_runtime();
-    let (watcher_actor, state) = ExitWatcher::new();
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let watcher = rt.spawn(watcher_actor).unwrap();
-
-    // Tell watcher to watch the target
-    rt.send_to(watcher, WatcherCmd::WatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-
-    // Kill the target
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    assert_eq!(state.count(), 1, "watcher should have received exactly one ActorExited");
-    assert_eq!(state.last_reason(), Some(ExitReason::Panicked));
-    assert_eq!(state.last_addr(), Some(target));
-}
-
-/// Given a watcher that watches then unwatches a target,
-/// when the target panics,
-/// then the watcher receives NO notification.
-#[test]
-fn unwatch_prevents_notification() {
-    let rt = watch_runtime();
-    let (watcher_actor, state) = ExitWatcher::new();
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let watcher = rt.spawn(watcher_actor).unwrap();
-
-    // Watch
-    rt.send_to(watcher, WatcherCmd::WatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-
-    // Unwatch
-    rt.send_to(watcher, WatcherCmd::UnwatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-
-    // Kill target
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    assert_eq!(state.count(), 0, "after unwatch, no notification should be delivered");
-}
-
-/// Given a watcher that dies before the target,
-/// when the target subsequently panics,
-/// then there is no panic or leak.
-#[test]
-fn watcher_dies_before_target_no_panic() {
-    let rt = watch_runtime();
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let target2 = rt.spawn(PanicOnCommand).unwrap();
-
-    // Watch via runtime-level API
-    rt.watch(target2, target);
-    tick_n(&rt, 3);
-
-    // Kill watcher (target2) first
-    rt.send_to(target2, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    // Kill target — should not crash
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    // If we got here, no crash.
-}
-
-/// Given a watcher that calls watch() twice on the same target,
-/// when the target panics,
-/// then the watcher receives exactly one notification.
-#[test]
-fn idempotent_watch_delivers_one_notification() {
-    let rt = watch_runtime();
-    let (watcher_actor, state) = ExitWatcher::new();
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let watcher = rt.spawn(watcher_actor).unwrap();
-
-    // Watch twice
-    rt.send_to(watcher, WatcherCmd::WatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-    rt.send_to(watcher, WatcherCmd::WatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-
-    // Kill target
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    assert_eq!(state.count(), 1, "double watch should produce exactly one notification");
-}
-
-/// Given multiple watchers on the same target,
-/// when the target panics,
-/// then all watchers receive the notification.
-#[test]
-fn multiple_watchers_all_notified() {
-    let rt = watch_runtime();
-    let (w1_actor, s1) = ExitWatcher::new();
-    let (w2_actor, s2) = ExitWatcher::new();
-    let (w3_actor, s3) = ExitWatcher::new();
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let w1 = rt.spawn(w1_actor).unwrap();
-    let w2 = rt.spawn(w2_actor).unwrap();
-    let w3 = rt.spawn(w3_actor).unwrap();
-
-    rt.send_to(w1, WatcherCmd::WatchThis(target)).unwrap();
-    rt.send_to(w2, WatcherCmd::WatchThis(target)).unwrap();
-    rt.send_to(w3, WatcherCmd::WatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    assert_eq!(s1.count(), 1, "watcher 1 should be notified");
-    assert_eq!(s2.count(), 1, "watcher 2 should be notified");
-    assert_eq!(s3.count(), 1, "watcher 3 should be notified");
-}
-
-/// Self-watch doesn't crash the runtime.
-#[test]
-fn self_watch_does_not_crash() {
-    let rt = watch_runtime();
-    let (watcher_actor, _state) = ExitWatcher::new();
-
-    let actor = rt.spawn(watcher_actor).unwrap();
-    rt.send_to(actor, WatcherCmd::WatchThis(actor)).unwrap();
-    tick_n(&rt, 5);
-
-    // No crash = pass
-}
-
-/// Runtime-level watch (outside actor context) delivers notification.
-#[test]
-fn runtime_level_watch_delivers_notification() {
-    let rt = watch_runtime();
-    let (watcher_actor, state) = ExitWatcher::new();
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let watcher = rt.spawn(watcher_actor).unwrap();
-    tick_n(&rt, 2); // ensure both spawned
-
-    rt.watch(watcher, target);
-
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    assert_eq!(state.count(), 1, "runtime-level watch should deliver notification");
-    assert_eq!(state.last_reason(), Some(ExitReason::Panicked));
-}
-
-/// Given a watcher watching target via on_actor_exit,
-/// when target panics,
-/// then the watcher can react by spawning a replacement (supervision pattern).
-#[test]
-fn watcher_can_react_to_death_by_spawning() {
-    let rt = watch_runtime();
-    let spawned = Arc::new(AtomicUsize::new(0));
-
-    struct Supervisor {
-        spawned_count: Arc<AtomicUsize>,
-    }
-
-    #[derive(Clone)]
-    enum SupervisorMsg {
-        WatchThis(ActorAddress),
-    }
-
-    impl ActorInterface for Supervisor {
-        type Incoming = SupervisorMsg;
-        type Response = ();
-        fn handle(&mut self, ctx: &Ctx, msg: SupervisorMsg) {
-            match msg {
-                SupervisorMsg::WatchThis(target) => ctx.watch(target),
-            }
-        }
-
-        fn on_actor_exit(&mut self, ctx: &Ctx, _exited: ActorExited) {
-            // React: spawn a replacement
-            let replacement = ctx.spawn(Sleeper).unwrap();
-            let _ = replacement;
-            self.spawned_count.fetch_add(1, Ordering::SeqCst);
-        }
-    }
-
-    let target = rt.spawn(PanicOnCommand).unwrap();
-    let sup = rt.spawn(Supervisor { spawned_count: spawned.clone() }).unwrap();
-
-    rt.send_to(sup, SupervisorMsg::WatchThis(target)).unwrap();
-    tick_n(&rt, 3);
-
-    rt.send_to(target, PanicMsg).unwrap();
-    tick_n(&rt, 5);
-
-    assert_eq!(spawned.load(Ordering::SeqCst), 1, "supervisor should have spawned a replacement");
-}
Author	SHA1	Message	Date
Claude	d9489ccac2	feat: consolidate tests	2026-02-13 14:48:35 +00:00
Claude	41af749bb4	refactor: decompose tick_once into helper methods Extract drain_spawns() (phases 1+4, dedup) and cleanup_dead_actors() (phase 7, 48 lines of complex death notification + GC logic). tick_once: 196 → 133 lines (−32%), max nesting depth: 5 → 3. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	5a8aeaf428	bench: add allocation decomposition benchmark for send_to Measures Box::new+type-erasure (2-160ns) vs full send_to (1.9-7.2us) to identify the real send latency bottleneck. Finding: allocation is only 1-5% of total cost; the dominant overhead is RwLock address_map lookup and HybridChannel operations. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	d370a41cf1	feat: add fn_complexity.py and loc_analysis.py analysis tools fn_complexity: per-function line count, nesting depth, and unsafe detection. loc_analysis: per-file breakdown of logic vs comment vs blank vs string lines. Both support --json, --top N, directory recursion. Key finding: tick_once is 196 lines (5x runner-up), core is 1,615 logic lines. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	e21a4ae19c	chore: clean up unused imports in test common module Remove redundant AtomicUsize/Ordering/Arc re-exports from common/mod.rs and runtime_mechanics.rs. Zero warnings now. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	133314903a	test: add 11 scenario tests for runtime-dashboard (from 0) EventStore cursor semantics (streaming, overflow, future cursor), recording pipeline (JSON roundtrip, destructive drain), StatsCollector multi-worker aggregation, and event sequence monotonicity. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	803289c7dc	docs: update runtime docs to reflect architecture changes from Cycles 1-3 Worker struct: TimerWheel → generic WorkerExtension. TickContext: registries moved behind RuntimeExtension. Phases 2.5/5.5/7 generalized. Ctx API split into core (6 methods) vs extension traits (swactor-std). Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	33df021afd	refactor: split runtime_api.rs (4,622 lines) into 7 focused test files Split monolithic test file into domain-specific test modules: - common/mod.rs: shared messages, actors, helpers - runtime_lifecycle.rs (57 tests): spawn, FIFO, threading, panic safety - runtime_mechanics.rs (10): placement, backpressure, cleanup, recovery - runtime_hooks.rs (12): on_start/on_stop, graceful stop - runtime_timers.rs (6): one-shot and interval timers - runtime_registry.rs (32): naming, monitoring, groups, ask pattern - runtime_supervision.rs (20): supervisor + router - runtime_hasher.rs (3): identity hasher correctness All 140 tests pass. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	edea0aac2a	refactor: extract 3,432 lines of embedded HTML to include_str!() Move HTML content from Rust raw string literals into proper .html files, replacing each _html.rs with a 1-line include_str!() call. - runtime-dashboard: dashboard.html (511), actors.html (739), distribution.html (746) - simulation-dashboard: dashboard.html (1,436) Enables proper HTML syntax highlighting and linting in editors. Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 14:48:35 +00:00
Claude	00e1fce80d	refactor: deduplicate message routing pattern in tick_once Extract route_to_pool_or_remote() helper that unifies the routing logic (local pool → address_map → inbox_registry) used in phases 2.5 (worker extension delivery) and 7 (death notifications). Completes P0 core debloating: 2636→2319 (−317, −12.0%). worker.rs: 845→598 (−247, −29.2%). Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 21:45:44 +07:00
Claude	09d5c75c39	feat: move TimerWheel from core to std via WorkerExtension Add WorkerExtension trait (on_tick, handle_request, gc_dead) as a per-worker counterpart to the shared RuntimeExtension. This enables per-worker state like timer wheels to live outside core. - TimerWheel, TimerRequest, CloneMsg → crates/std/src/timer_wheel.rs - CtxTimers extension trait replaces Ctx::send_after_ticks/send_interval_ticks - schedule_timer → generic post_worker_request on ContextInner - Worker.timers → Worker.worker_ext (Option<Box<dyn WorkerExtension>>) - StdExtension factory creates TimerWheel per worker Core: 2440→2320 (−120). worker.rs: 696→599 (−97). Cumulative: 2636→2320 (−316, −12.0%). worker.rs: 845→599 (−29.1%). Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 21:45:44 +07:00
Claude	93b1b1fa1e	feat: move WatchRegistry from core to std extension Merge Watch death-notification system into swactor-std, reducing core surface. WatchRegistry now lives alongside MonitorRegistry in StdExtension, with CtxWatching and RuntimeWatching extension traits for the public API. Phase 5b removed from tick_once; notifications delivered via on_actor_death in phase 7. Core shrinks from 2,636 to 2,440 lines (-196, -7.4%). Authored by Claude, lovingly guided by Zachery Aaron Shores-Chmielewski	2026-02-13 21:45:44 +07:00
zacheryasc	1169f0322d	feat: dashboard updates (#38 ) Adds some more polish to the dashboard. Further modification will have to wait until swactor is used for more complex projects.	2026-02-13 14:42:10 +00:00