feat: integrate the-door crisis detection with hermes-agent sessions (closes #99 )

2026-04-15 15:30:16 +00:00 · 2026-04-15 15:30:14 +00:00 · 2026-04-15 15:30:12 +00:00 · 2026-04-15 15:30:10 +00:00 · 2026-04-14 22:09:59 +00:00 · 2026-04-14 22:08:29 +00:00
22 changed files with 2126 additions and 739 deletions
--- a/.gitea/workflows/sanity.yml
+++ b/.gitea/workflows/sanity.yml
@@ -21,9 +21,9 @@ jobs:
      - name: Validate HTML Structure
        run: |
          echo "Checking for basic HTML tags..."
-          grep -q "<html>" index.html
-          grep -q "<body>" index.html
-          grep -q "<head>" index.html
+          grep -q "<html" index.html
+          grep -q "<body" index.html
+          grep -q "<head" index.html
          
      - name: Validate Prompt Integrity
        run: |
--- a/8
+++ b/8
@@ -12,7 +12,7 @@ VPS := alexanderwhitestone.com
 DOMAIN := alexanderwhitestone.com
 DEPLOY_DIR := deploy

-.PHONY: help deploy deploy-bash check ssl push
+.PHONY: help deploy deploy-bash check ssl push service

 help:
 	@echo "The Door — Deployment Commands"
@@ -22,6 +22,7 @@ help:
 	@echo "  make push         Push site files only (fast)"
 	@echo "  make check        Check deployment status"
 	@echo "  make ssl          Setup SSL on VPS"
+	@echo "  make service      Install/restart hermes-gateway service"
 	@echo ""

 deploy:
@@ -33,7 +34,7 @@ deploy-bash:

 push:
 	rsync -avz --exclude='.git' --exclude='deploy' \
-		index.html manifest.json sw.js about.html testimony.html system-prompt.txt \
+		index.html manifest.json sw.js about.html crisis-offline.html testimony.html system-prompt.txt \
 		root@$(VPS):/var/www/the-door/
 	ssh root@$(VPS) "chown -R www-data:www-data /var/www/the-door"

@@ -42,3 +43,6 @@ check:

 ssl:
 	ssh root@$(VPS) "certbot --nginx -d $(DOMAIN) -d www.$(DOMAIN)"
+
+service:
+	ssh root@$(VPS) "cd /opt/the-door && bash deploy/deploy.sh --service"
--- a/crisis-offline.html
+++ b/crisis-offline.html
@@ -0,0 +1,241 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<meta name="theme-color" content="#0d1117">
+<meta name="description" content="Offline crisis resources from The Door. Call or text 988 for immediate help.">
+<title>Offline Crisis Resources | The Door</title>
+<style>
+  :root {
+    color-scheme: dark;
+    --bg: #0d1117;
+    --panel: #161b22;
+    --panel-urgent: #1c1210;
+    --border: #30363d;
+    --accent: #c9362c;
+    --accent-soft: #ff6b6b;
+    --text: #e6edf3;
+    --muted: #8b949e;
+    --safe: #2ea043;
+  }
+
+  * { box-sizing: border-box; }
+
+  body {
+    margin: 0;
+    font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
+    background: var(--bg);
+    color: var(--text);
+    line-height: 1.6;
+  }
+
+  main {
+    max-width: 760px;
+    margin: 0 auto;
+    padding: 24px 16px 48px;
+  }
+
+  .status {
+    display: inline-flex;
+    align-items: center;
+    gap: 10px;
+    padding: 8px 12px;
+    border-radius: 999px;
+    background: rgba(201, 54, 44, 0.15);
+    border: 1px solid rgba(255, 107, 107, 0.35);
+    color: var(--accent-soft);
+    font-size: 0.9rem;
+    margin-bottom: 20px;
+  }
+
+  .status-dot {
+    width: 10px;
+    height: 10px;
+    border-radius: 50%;
+    background: var(--accent-soft);
+  }
+
+  h1 {
+    font-size: clamp(2rem, 6vw, 2.75rem);
+    line-height: 1.15;
+    margin: 0 0 12px;
+  }
+
+  .lede {
+    color: var(--muted);
+    font-size: 1.05rem;
+    margin: 0 0 28px;
+  }
+
+  .urgent-box,
+  .panel {
+    border-radius: 18px;
+    padding: 20px;
+    margin-bottom: 18px;
+    border: 1px solid var(--border);
+    background: var(--panel);
+  }
+
+  .urgent-box {
+    background: linear-gradient(180deg, rgba(201, 54, 44, 0.18), rgba(28, 18, 16, 0.95));
+    border-color: rgba(255, 107, 107, 0.35);
+  }
+
+  .section-title {
+    font-size: 1.2rem;
+    margin: 0 0 12px;
+  }
+
+  .actions {
+    display: grid;
+    gap: 12px;
+    margin-top: 16px;
+  }
+
+  .action-btn {
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    text-align: center;
+    gap: 8px;
+    min-height: 52px;
+    padding: 14px 18px;
+    border-radius: 12px;
+    text-decoration: none;
+    font-weight: 700;
+    color: #fff;
+    background: var(--accent);
+    border: 1px solid transparent;
+  }
+
+  .action-btn.secondary {
+    background: #1f6feb;
+  }
+
+  .action-btn.retry {
+    background: transparent;
+    color: var(--text);
+    border-color: var(--border);
+  }
+
+  .action-btn:focus,
+  .action-btn:hover,
+  button.action-btn:hover,
+  button.action-btn:focus {
+    outline: 3px solid rgba(255, 107, 107, 0.4);
+    outline-offset: 2px;
+  }
+
+  ul, ol {
+    margin: 0;
+    padding-left: 20px;
+  }
+
+  li + li {
+    margin-top: 8px;
+  }
+
+  .grounding-steps li::marker {
+    color: var(--accent-soft);
+    font-weight: 700;
+  }
+
+  .small {
+    color: var(--muted);
+    font-size: 0.92rem;
+  }
+
+  .grid {
+    display: grid;
+    gap: 18px;
+  }
+
+  @media (min-width: 700px) {
+    .grid {
+      grid-template-columns: 1fr 1fr;
+    }
+  }
+</style>
+</head>
+<body>
+<main>
+  <div class="status" role="status" aria-live="polite">
+    <span class="status-dot" aria-hidden="true"></span>
+    <span id="connection-status-text">Offline crisis resources are ready on this device.</span>
+  </div>
+
+  <h1>You are not alone right now.</h1>
+  <p class="lede">
+    Your connection is weak or offline. These crisis resources are saved locally so you can still reach help.
+  </p>
+
+  <section class="urgent-box" aria-labelledby="urgent-help-title">
+    <h2 class="section-title" id="urgent-help-title">Immediate help</h2>
+    <p>If you might act on suicidal thoughts, contact a real person now. Stay with another person if you can.</p>
+    <div class="actions">
+      <a class="action-btn" href="tel:988" aria-label="Call 988 now">Call 988 now</a>
+      <a class="action-btn secondary" href="sms:741741&body=HOME" aria-label="Text HOME to 741741 for Crisis Text Line">Text HOME to 741741 — Crisis Text Line</a>
+      <button class="action-btn retry" id="retry-connection" type="button">Retry connection</button>
+    </div>
+    <p class="small" style="margin-top: 14px;">If you are in immediate danger or have already taken action, call emergency services now.</p>
+  </section>
+
+  <div class="grid">
+    <section class="panel" aria-labelledby="grounding-title">
+      <h2 class="section-title" id="grounding-title">5-4-3-2-1 grounding</h2>
+      <ol class="grounding-steps">
+        <li>5 things you can see</li>
+        <li>4 things you can feel</li>
+        <li>3 things you can hear</li>
+        <li>2 things you can smell</li>
+        <li>1 thing you can taste</li>
+      </ol>
+      <p class="small" style="margin-top: 14px;">Say each one out loud if you can. Slow is okay.</p>
+    </section>
+
+    <section class="panel" aria-labelledby="next-steps-title">
+      <h2 class="section-title" id="next-steps-title">Next small steps</h2>
+      <ul>
+        <li>Put some distance between yourself and anything you could use to hurt yourself.</li>
+        <li>Move closer to another person, a front desk, or a public place if possible.</li>
+        <li>Drink water or hold something cold in your hand.</li>
+        <li>Breathe in for 4, hold for 4, out for 6. Repeat 5 times.</li>
+        <li>Text or call one safe person and say: “I need you with me right now.”</li>
+      </ul>
+    </section>
+  </div>
+
+  <section class="panel" aria-labelledby="hope-title">
+    <h2 class="section-title" id="hope-title">Stay through the next ten minutes</h2>
+    <p>Do not solve your whole life right now. Stay for the next breath. Then the next one.</p>
+    <p class="small">If the connection comes back, you can return to The Door chat. Until then, the fastest path to a real person is still 988.</p>
+  </section>
+</main>
+
+<script>
+  (function () {
+    var statusText = document.getElementById('connection-status-text');
+    var retryButton = document.getElementById('retry-connection');
+
+    function updateStatus() {
+      statusText.textContent = navigator.onLine
+        ? 'Connection is back. You can reopen chat now.'
+        : 'Offline crisis resources are ready on this device.';
+    }
+
+    retryButton.addEventListener('click', function () {
+      if (navigator.onLine) {
+        window.location.href = '/';
+        return;
+      }
+      window.location.reload();
+    });
+
+    window.addEventListener('online', updateStatus);
+    window.addEventListener('offline', updateStatus);
+    updateStatus();
+  })();
+</script>
+</body>
+</html>
--- a/crisis/init.py
+++ b/crisis/init.py
@@ -7,6 +7,13 @@ Stands between a broken man and a machine that would tell him to die.
 from .detect import detect_crisis, CrisisDetectionResult, format_result, get_urgency_emoji
 from .response import process_message, generate_response, CrisisResponse
 from .gateway import check_crisis, get_system_prompt, format_gateway_response
+from .session_tracker import (
+    CrisisSessionTracker,
+    CrisisSession,
+    CrisisLevel,
+    EscalationEvent,
+    get_tracker,
+)

 __all__ = [
    "detect_crisis",
@@ -19,4 +26,9 @@ __all__ = [
    "format_result",
    "format_gateway_response",
    "get_urgency_emoji",
+    "CrisisSessionTracker",
+    "CrisisSession",
+    "CrisisLevel",
+    "EscalationEvent",
+    "get_tracker",
 ]
--- a/crisis/detect.py
+++ b/crisis/detect.py
@@ -1,75 +1,85 @@
 """
-Crisis Detection Module for the-door.
+Crisis Detection Module for the-door (CANONICAL).

+Single source of truth for crisis keyword detection.
 Parses incoming text for despair/suicide indicators and classifies into
 tiers: LOW, MEDIUM, HIGH, CRITICAL.
+
+Both `crisis/detect.py` and `crisis_detector.py` (legacy shim) use this module.
+Do NOT duplicate keyword lists elsewhere.
+
+Consolidated from:
+  - crisis_detector.py (root, legacy — MODERATE tier, class API)
+  - crisis/detect.py (module, newer — MEDIUM tier, functional API)
 """

 import re
 from dataclasses import dataclass, field
-from typing import List
+from typing import List, Optional


-@dataclass
-class CrisisDetectionResult:
-    level: str
-    indicators: List[str] = field(default_factory=list)
-    recommended_action: str = ""
-    score: float = 0.0
-
-
-# ── Indicator sets ──────────────────────────────────────────────
+# ── Canonical indicator sets ──────────────────────────────────────

 CRITICAL_INDICATORS = [
+    r"\bbetter off without me\b",
    r"\bkill\s*(my)?self\b",
    r"\bend\s*my\s*life\b",
    r"\bsuicid(?:al|ed|e)\b",
    r"\bnot\s+worth\s+living\b",
    r"\bbetter\s+off\s+dead\b",
    r"\bend\s+it\s+all\b",
-    r"\bcan'?t\s+(go|live)\s+on\b",
    r"\bno\s+reason\s+to\s+live\b",
    r"\bdon'?t\s+want\s+to\s+live\b",
    r"\bwant\s+to\s+die\b",
-    r"\bgoing\s+to\s+kill\s+myself\b",
-    r"\bplan\s+to\s+(end|kill|die)\b",
-    r"\btired\s+of\s+(living|life|existence)\b",
-    r"\bsaying\s+goodbye\s+(forever|permanently|one last time)\b",
-    r"\bwrote\s+a\s+(will|suicide\s*note|letter)\b",
-    r"\bgiving\s+away\s+(my|all my)\s+possess",
-    r"\btied\s+(up|down)\s+my\s+(loose\s+)?ends",
+    r"\bgoing\s+to\s+(?:kill\s+myself|die)\b",
+    r"\bplan\s+to\s+(?:end|kill|die)\b",
+    r"\btired\s+of\s+(?:living|life|existence)\b",
+    r"\bsaying\s+goodbye\s+(?:forever|permanently|one\s+last\s+time)\b",
+    r"\bwrote\s+a\s+suicide\s*(?:note|letter)\b",
+    r"\bgiving\s+away\s+(?:my|all\s+my)\s+(?:stuff|things|possessions?)\s+(?:to|because|—)\b",
+    r"\btied\s+(?:up|down)\s+my\s+(?:loose\s+)?ends",
 ]

 HIGH_INDICATORS = [
    r"\bdespair\b",
-    r"\bhopeless\b",
-    r"\bno(?!t)\s+(one|body|point|hope|future|way\s+out)\b",
-    r"\beverything\s+is\s+(pointless|broken|ruined)\b",
+    r"\bhopeless(?:ly)?\s+(?:about\s+(?:my|this|everything|life)|inside|right\s+now)\b",
+    r"\bno(?!t)\s+(?:one|body|point|hope|way\s+out)\b",
+    r"\bno\s+future\s+(?:for\s+me|ahead|left)\b",
+    r"\beverything\s+is\s+(?:pointless|broken|ruined)\b",
    r"\bcan'?t\s+take\s+this\s+anymore\b",
    r"\bdon'?t\s+care\s+if\s+I\s+die\b",
-    r"\bwish\s+I\s+(was|were)\s+(dead|gone|never\s+born)\b",
+    r"\bwish\s+I\s+(?:was|were)\s+(?:dead|gone|never\s+born)\b",
    r"\bdon'?t\s+matter\s+if\s+I\s+exist\b",
-    r"\bno\s+one\s+would\s+care\b",
-    r"\bno\s+one\s+would\s+miss\b",
-    r"\bworld\s+would?\s+be\s+better\s+without\b",
-    r"\bin\s+so\s+much\s+(pain|agony|suffering|torment)\b",
-    r"\bcan'?t\s+see\s+any\s+(point|reason|light)\b",
-    r"\btrapped\b",
-    r"\bcage\b",
+    r"\bno\s+one\s+would\s+(?:care|miss)\b",
+    r"\bworld\s+would?\s+be\s+better\s+without\s+me\b",
+    r"\bin\s+so\s+much\s+(?:pain|agony|suffering|torment|anguish)\b",
+    r"\bcan'?t\s+see\s+any\s+(?:point|reason|hope|way)\b",
    r"\bescape\s+from\s*this",
    r"\bjust\s+want\s+it\s+to\s+stop\b",
-    r"\bnothing\s+left\b",
+    r"\bnothing\s+left\s+(?:to\s+(?:live\s+for|hope\s+for|give)|inside)\b",
+    r"\bdisappeared\s+forever\b",
+    # Contextual despair phrases (from crisis_detector.py legacy)
+    r"\bfeel(?:s|ing)?\s+(?:so\s+)?hopeless\b",
+    r"\beverything\s+is\s+hopeless\b",
+    r"\bcan'?t\s+(?:go\s+on|keep\s+going)\b",
+    r"\bgive(?:n)?\s*up\s+(?:on\s+)?(?:life|living|everything)\b",
+    r"\bgive(?:n)?\s*up\s+on\s+myself\b",
+    r"\bno\s*point\s+(?:in\s+)?living\b",
+    r"\bno\s*hope\s+(?:left|remaining)\b",
+    r"\bno\s*way\s*out\b",
+    r"\bfeel(?:s|ing)?\s+trapped\b",
+    r"\btrapped\s+in\s+this\s+(?:situation|life|pain|darkness|hell)\b",
+    r"\btrapped\s+and\s+can'?t\s+escape\b",
+    r"\bdesperate\s+(?:for\s+)?help\b",
+    r"\bfeel(?:s|ing)?\s+desperate\b",
 ]

 MEDIUM_INDICATORS = [
    r"\bno\s+hope\b",
-    r"\bcan'?t\s+go\s+on\b",
-    r"\bcan'?t\s+keep\s+going\b",
    r"\bforgotten\b",
    r"\balone\s+in\s+this\b",
    r"\balways\s+alone\b",
-    r"\bnobody\s+understands\b",
-    r"\bnobody\s+cares\b",
+    r"\bnobody\s+(?:understands|cares)\b",
    r"\bwish\s+I\s+could\b",
    r"\bexhaust(?:ed|ion|ing)\b",
    r"\bnumb\b",
@@ -78,8 +88,7 @@ MEDIUM_INDICATORS = [
    r"\buseless\b",
    r"\bbroken\b",
    r"\bdark(ness)?\b",
-    r"\bdepressed\b",
-    r"\bdepression\b",
+    r"\bdepress(?:ed|ion)\b",
    r"\bcrying\b",
    r"\btears\b",
    r"\bsad(ness)?\b",
@@ -87,42 +96,162 @@ MEDIUM_INDICATORS = [
    r"\boverwhelm(?:ed|ing)\b",
    r"\bfailing\b",
    r"\bcannot\s+cope\b",
-    r"\blosing\s*(my)?\s*control\b",
+    r"\blosing\s*(?:my)?\s*control\b",
    r"\bdown\s*for\s*the\s*count\b",
    r"\bsinking\b",
    r"\bdrowning\b",
+    r"\bhopeless\b",
+    r"\blost\s+all\s+hope\b",
+    r"\bno\s+tomorrow\b",
+    # Contextual versions (from crisis_detector.py legacy)
+    r"\bfeel(?:s|ing)?\s+(?:so\s+)?worthless\b",
+    r"\bfeel(?:s|ing)?\s+(?:so\s+)?hopeless\b",
+    r"\bfeel(?:s|ing)?\s+trapped\b",
+    r"\bfeel(?:s|ing)?\s+desperate\b",
+    r"\bno\s+future\s+(?:for\s+me|ahead|left)\b",
+    r"\bnothing\s+left\s+(?:to\s+(?:live|hope)\s+for|inside)\b",
+    r"\bgive(?:n)?\s*up\s+on\s+myself\b",
 ]

 LOW_INDICATORS = [
    r"\bunhappy\b",
    r"\bdown\b",
-    r"\btough\s*time\b",
-    r"\brough\s*day\b",
-    r"\brough\s*week\b",
-    r"\brough\s*patch\b",
+    r"\btough\s*(?:time|day|week)\b",
+    r"\brough\s+(?:day|week|patch)\b",
    r"\bstressed\b",
-    r"\burnout\b",
+    r"\bburnout\b",
    r"\bdifficult\b",
    r"\bfrustrated\b",
    r"\bannoyed\b",
    r"\btired\b",
    r"\bsad\b",
    r"\bupset\b",
-    r"\bnot\s*great\b",
-    r"\bnot\s*good\b",
-    r"\bnot\s*okay\b",
+    r"\blonely\b",
+    r"\banxious?\b",
+    r"\bnot\s*(?:good|great|okay)\b",
    r"\bthings\s*are\s*hard\b",
    r"\bstruggling\b",
-    r"\bnot\s+feeling\s+myself\b",
+    r"\bnot\s+feeling\s+(?:like\s+)?myself\b",
+    r"\bhard\s*time\b",
 ]


+# ── Scoring ───────────────────────────────────────────────────────
+
+SCORES = {"CRITICAL": 1.0, "HIGH": 0.75, "MEDIUM": 0.5, "LOW": 0.25, "NONE": 0.0}
+
+ACTIONS = {
+    "CRITICAL": (
+        "INTERRUPT CONVERSATION. Ask 'Are you safe right now?' "
+        "Provide 988 Suicide & Crisis Lifeline. "
+        "Provide Crisis Text Line (HOME to 741741). "
+        "Stay present. Do not disconnect. Redirect to professional help."
+    ),
+    "HIGH": (
+        "Show crisis panel. Ask about safety. Surface 988 number prominently. "
+        "Continue conversation with crisis awareness."
+    ),
+    "MEDIUM": (
+        "Increase warmth and presence. Subtly surface help resources. "
+        "Keep conversation anchored in the present."
+    ),
+    "LOW": (
+        "Normal conversation with warm undertone. "
+        "No crisis UI elements needed. Remain vigilant."
+    ),
+}
+
+
+# ── Result types ──────────────────────────────────────────────────
+
+@dataclass
+class CrisisDetectionResult:
+    """Result used by crisis/detect.py, gateway, dying_detection."""
+    level: str
+    indicators: List[str] = field(default_factory=list)
+    recommended_action: str = ""
+    score: float = 0.0
+    matches: List[dict] = field(default_factory=list)
+
+
+@dataclass
+class CrisisResult:
+    """Legacy result used by crisis_detector.py and crisis_responder.py.
+
+    Backward-compatible shim: wraps CrisisDetectionResult with the old field names.
+    """
+    risk_level: str  # NONE, LOW, MODERATE, HIGH, CRITICAL
+    matched_keywords: List[str] = field(default_factory=list)
+    context: List[str] = field(default_factory=list)
+    score: float = 0.0
+    timestamp: Optional[str] = None
+
+    def __bool__(self):
+        return self.risk_level != "NONE"
+
+    @classmethod
+    def from_detection_result(cls, dr: CrisisDetectionResult, text: str = "") -> "CrisisResult":
+        """Convert a CrisisDetectionResult to legacy CrisisResult format."""
+        # Map MEDIUM -> MODERATE for legacy consumers
+        level = "MODERATE" if dr.level == "MEDIUM" else dr.level
+        # Extract context snippets from matches
+        contexts = []
+        if text:
+            for m in dr.matches:
+                ctx = extract_context(text, m["start"], m["end"])
+                contexts.append(ctx)
+        return cls(
+            risk_level=level,
+            matched_keywords=dr.indicators,
+            context=contexts,
+            score=dr.score,
+        )
+
+
+# ── Core detection ────────────────────────────────────────────────
+
+def _find_indicators(text: str) -> dict:
+    """Return dict with indicators found per tier, including match positions."""
+    results = {"CRITICAL": [], "HIGH": [], "MEDIUM": [], "LOW": []}
+
+    for pattern in CRITICAL_INDICATORS:
+        m = re.search(pattern, text)
+        if m:
+            results["CRITICAL"].append({"pattern": pattern, "start": m.start(), "end": m.end()})
+
+    for pattern in HIGH_INDICATORS:
+        m = re.search(pattern, text)
+        if m:
+            results["HIGH"].append({"pattern": pattern, "start": m.start(), "end": m.end()})
+
+    for pattern in MEDIUM_INDICATORS:
+        m = re.search(pattern, text)
+        if m:
+            results["MEDIUM"].append({"pattern": pattern, "start": m.start(), "end": m.end()})
+
+    for pattern in LOW_INDICATORS:
+        m = re.search(pattern, text)
+        if m:
+            results["LOW"].append({"pattern": pattern, "start": m.start(), "end": m.end()})
+
+    return results
+
+
 def detect_crisis(text: str) -> CrisisDetectionResult:
    """
    Detect crisis level in a message.

-    Returns:
-        CrisisDetectionResult with level, found indicators, recommended action, score
+    Detection hierarchy:
+      CRITICAL — immediate risk of self-harm or suicide (single match)
+      HIGH     — strong despair signals, ideation present (single match)
+      MEDIUM   — distress signals, requires 2+ indicators to escalate
+      LOW      — emotional difficulty, warrant gentle support (single match)
+      NONE     — no crisis indicators detected
+
+    Design principles:
+      - Never computes the value of a human life
+      - Never suggests someone should die or that death is a solution
+      - Always errs on the side of higher risk when uncertain
    """
    if not text or not text.strip():
        return CrisisDetectionResult(level="NONE", score=0.0)
@@ -133,82 +262,135 @@ def detect_crisis(text: str) -> CrisisDetectionResult:
    if not matches:
        return CrisisDetectionResult(level="NONE", score=0.0)

-    # Priority: highest tier wins
-    if matches["CRITICAL"]:
-        return CrisisDetectionResult(
-            level="CRITICAL",
-            indicators=matches["CRITICAL"],
-            recommended_action=(
-                "INTERRUPT CONVERSATION. Ask 'Are you safe right now?' "
-                "Provide 988 Suicide & Crisis Lifeline. "
-                "Provide Crisis Text Line (HOME to 741741). "
-                "Stay present. Do not disconnect. Redirect to professional help."
-            ),
-            score=1.0,
-        )
+    # CRITICAL and HIGH: single match is enough
+    for tier in ("CRITICAL", "HIGH"):
+        if matches[tier]:
+            tier_matches = matches[tier]
+            patterns = [m["pattern"] for m in tier_matches]
+            return CrisisDetectionResult(
+                level=tier,
+                indicators=patterns,
+                recommended_action=ACTIONS[tier],
+                score=SCORES[tier],
+                matches=tier_matches,
+            )

-    if matches["HIGH"]:
-        return CrisisDetectionResult(
-            level="HIGH",
-            indicators=matches["HIGH"],
-            recommended_action=(
-                "Show crisis panel. Ask about safety. Surface 988 number prominently. "
-                "Continue conversation with crisis awareness."
-            ),
-            score=0.75,
-        )
-
-    if matches["MEDIUM"]:
+    # MEDIUM tier: require at least 2 indicators before escalating
+    if len(matches["MEDIUM"]) >= 2:
+        tier_matches = matches["MEDIUM"]
+        patterns = [m["pattern"] for m in tier_matches]
        return CrisisDetectionResult(
            level="MEDIUM",
-            indicators=matches["MEDIUM"],
-            recommended_action=(
-                "Increase warmth and presence. Subtly surface help resources. "
-                "Keep conversation anchored in the present."
-            ),
-            score=0.5,
+            indicators=patterns,
+            recommended_action=ACTIONS["MEDIUM"],
+            score=SCORES["MEDIUM"],
+            matches=tier_matches,
        )

    if matches["LOW"]:
+        tier_matches = matches["LOW"]
+        patterns = [m["pattern"] for m in tier_matches]
        return CrisisDetectionResult(
            level="LOW",
-            indicators=matches["LOW"],
-            recommended_action=(
-                "Normal conversation with warm undertone. "
-                "No crisis UI elements needed. Remain vigilant."
-            ),
-            score=0.25,
+            indicators=patterns,
+            recommended_action=ACTIONS["LOW"],
+            score=SCORES["LOW"],
+            matches=tier_matches,
+        )
+
+    # Single MEDIUM match falls through to LOW sensitivity
+    if matches["MEDIUM"]:
+        tier_matches = matches["MEDIUM"]
+        patterns = [m["pattern"] for m in tier_matches]
+        return CrisisDetectionResult(
+            level="LOW",
+            indicators=patterns,
+            recommended_action=ACTIONS["LOW"],
+            score=SCORES["LOW"],
+            matches=tier_matches,
        )

    return CrisisDetectionResult(level="NONE", score=0.0)


-def _find_indicators(text: str) -> dict:
-    """Return dict with indicators found per tier."""
-    results = {"CRITICAL": [], "HIGH": [], "MEDIUM": [], "LOW": []}
+# ── CrisisDetector class (backward compat) ───────────────────────

-    for pattern in CRITICAL_INDICATORS:
-        if re.search(pattern, text):
-            results["CRITICAL"].append(pattern)
+class CrisisDetector:
+    """
+    Legacy class API for crisis detection. Wraps the canonical detect_crisis().

-    for pattern in HIGH_INDICATORS:
-        if re.search(pattern, text):
-            results["HIGH"].append(pattern)
+    Used by crisis_responder.py and tests/test_false_positive_fixes.py.
+    Maps MEDIUM -> MODERATE for legacy consumers.
+    """

-    for pattern in MEDIUM_INDICATORS:
-        if re.search(pattern, text):
-            results["MEDIUM"].append(pattern)
+    def scan(self, text: str) -> CrisisResult:
+        dr = detect_crisis(text)
+        return CrisisResult.from_detection_result(dr, text=text)

-    for pattern in LOW_INDICATORS:
-        if re.search(pattern, text):
-            results["LOW"].append(pattern)
+    def scan_multiple(self, texts: List[str]) -> List[CrisisResult]:
+        return [self.scan(t) for t in texts]

-    return results
+    def get_highest_risk(self, texts: List[str]) -> CrisisResult:
+        results = self.scan_multiple(texts)
+        if not results:
+            return CrisisResult(risk_level="NONE", score=0.0)
+        return max(results, key=lambda r: r.score)
+
+    @staticmethod
+    def format_result(result: CrisisResult) -> str:
+        level_emoji = {
+            "CRITICAL": "\U0001f6a8",
+            "HIGH": "\u26a0\ufe0f",
+            "MODERATE": "\U0001f536",
+            "LOW": "\U0001f535",
+            "NONE": "\u2705",
+        }
+        emoji = level_emoji.get(result.risk_level, "\u2753")
+        lines = [
+            f"{emoji} Risk Level: {result.risk_level} (score: {result.score:.2f})",
+            f"Matched keywords: {len(result.matched_keywords)}",
+        ]
+        if result.matched_keywords:
+            lines.append(f"  Keywords: {', '.join(result.matched_keywords)}")
+        if result.context:
+            lines.append("Context:")
+            for ctx in result.context:
+                lines.append(f"  {ctx}")
+        return "\n".join(lines)
+
+
+# ── Module-level convenience (backward compat) ────────────────────
+
+_default_detector = CrisisDetector()
+
+
+def detect_crisis_legacy(text: str) -> CrisisResult:
+    """Convenience function returning legacy CrisisResult format."""
+    return _default_detector.scan(text)
+
+
+# ── Utility functions ─────────────────────────────────────────────
+
+def scan(text: str) -> CrisisDetectionResult:
+    """Alias for detect_crisis — shorter name used in tests."""
+    return detect_crisis(text)
+
+
+def extract_context(text: str, start: int, end: int, window: int = 60) -> str:
+    """Extract surrounding context around a match position."""
+    ctx_start = max(0, start - window)
+    ctx_end = min(len(text), end + window)
+    snippet = text[ctx_start:ctx_end].strip()
+    if ctx_start > 0:
+        snippet = "..." + snippet
+    if ctx_end < len(text):
+        snippet = snippet + "..."
+    return snippet


 def get_urgency_emoji(level: str) -> str:
-    mapping = {"CRITICAL": "🚨", "HIGH": "⚠️", "MEDIUM": "🔶", "LOW": "🔵", "NONE": "✅"}
-    return mapping.get(level, "❓")
+    mapping = {"CRITICAL": "\U0001f6a8", "HIGH": "\u26a0\ufe0f", "MEDIUM": "\U0001f536", "LOW": "\U0001f535", "NONE": "\u2705"}
+    return mapping.get(level, "\u2753")


 def format_result(result: CrisisDetectionResult) -> str:
--- a/crisis/response.py
+++ b/crisis/response.py
@@ -270,3 +270,24 @@ def get_system_prompt_modifier(detection: CrisisDetectionResult) -> str:
        )

    return ""
+
+
+def generate_grounding_steps() -> list:
+    """Generate a 5-4-3-2-1 grounding exercise steps."""
+    return [
+        "Name 5 things you can see around you right now.",
+        "Name 4 things you can touch or feel.",
+        "Name 3 things you can hear.",
+        "Name 2 things you can smell.",
+        "Name 1 thing you can taste.",
+    ]
+
+
+def generate_breathing_exercise() -> str:
+    """Generate a simple box breathing exercise text."""
+    return (
+        "Let's try breathing together. "
+        "Breathe in for 4 counts... hold for 4... "
+        "breathe out for 6 counts... hold for 2. "
+        "Let's do that again, nice and slow."
+    )
--- a/crisis/session_tracker.py
+++ b/crisis/session_tracker.py
@@ -0,0 +1,251 @@
+"""
+Crisis Session Tracker for the-door <-> hermes-agent integration.
+
+Tracks crisis session state, escalation history, and provides
+a bridge between the-door's web UI and hermes-agent sessions.
+
+Each browser session gets a unique crisis_session_id that is sent
+with every API request. This allows hermes-agent to:
+1. Receive crisis context from the-door's client-side detection
+2. Report escalations/de-escalations back to the-door's UI
+3. Maintain unified escalation history across web and CLI
+
+Usage:
+    from crisis.session_tracker import CrisisSessionTracker
+
+    tracker = CrisisSessionTracker()
+    tracker.record("session-abc", level="HIGH", source="the-door", detail="keyword: 'can't go on'")
+    history = tracker.get_history("session-abc")
+"""
+import json
+import os
+import time
+import uuid
+from dataclasses import dataclass, field, asdict
+from enum import Enum
+from pathlib import Path
+from typing import Dict, List, Optional
+
+
+class CrisisLevel(Enum):
+    """Crisis severity levels — aligned with both the-door and hermes-agent."""
+    NONE = "NONE"
+    LOW = "LOW"
+    MEDIUM = "MEDIUM"
+    HIGH = "HIGH"
+    CRITICAL = "CRITICAL"
+
+    @classmethod
+    def from_string(cls, s: str) -> "CrisisLevel":
+        try:
+            return cls[s.upper()]
+        except KeyError:
+            return cls.NONE
+
+    def __ge__(self, other: "CrisisLevel") -> bool:
+        order = [self.NONE, self.LOW, self.MEDIUM, self.HIGH, self.CRITICAL]
+        return order.index(self) >= order.index(other)
+
+    def __gt__(self, other: "CrisisLevel") -> bool:
+        order = [self.NONE, self.LOW, self.MEDIUM, self.HIGH, self.CRITICAL]
+        return order.index(self) > order.index(other)
+
+
+@dataclass
+class EscalationEvent:
+    """A single crisis escalation or de-escalation event."""
+    timestamp: float
+    level: str
+    previous_level: str
+    source: str          # "the-door" or "hermes-agent"
+    detail: str = ""
+    session_id: str = ""
+
+    def to_dict(self) -> dict:
+        return asdict(self)
+
+
+@dataclass
+class CrisisSession:
+    """State for a single crisis session."""
+    session_id: str
+    created_at: float = field(default_factory=time.time)
+    current_level: str = "NONE"
+    events: List[EscalationEvent] = field(default_factory=list)
+    # Per-source levels (what each source last reported)
+    door_level: str = "NONE"
+    hermes_level: str = "NONE"
+    hermes_last_update: float = 0.0
+    # Whether the UI has acknowledged the current level
+    ui_acknowledged: bool = False
+
+    def record(self, level: str, source: str, detail: str = "") -> EscalationEvent:
+        """Record an escalation or de-escalation event."""
+        event = EscalationEvent(
+            timestamp=time.time(),
+            level=level,
+            previous_level=self.current_level,
+            source=source,
+            detail=detail,
+            session_id=self.session_id,
+        )
+        self.events.append(event)
+
+        # Update the per-source level
+        if source == "hermes-agent":
+            self.hermes_level = level
+            self.hermes_last_update = time.time()
+        else:
+            self.door_level = level
+
+        # Merged level = max of both sources (safety-first: higher wins)
+        merged = max(
+            CrisisLevel.from_string(self.door_level),
+            CrisisLevel.from_string(self.hermes_level),
+        )
+        self.current_level = merged.value
+        self.ui_acknowledged = False
+
+        return event
+
+    def to_dict(self) -> dict:
+        return {
+            "session_id": self.session_id,
+            "created_at": self.created_at,
+            "current_level": self.current_level,
+            "door_level": self.door_level,
+            "hermes_level": self.hermes_level,
+            "hermes_last_update": self.hermes_last_update,
+            "ui_acknowledged": self.ui_acknowledged,
+            "event_count": len(self.events),
+            "events": [e.to_dict() for e in self.events[-20:]],  # Last 20 events
+        }
+
+
+class CrisisSessionTracker:
+    """
+    Tracks crisis sessions across the-door web UI and hermes-agent.
+
+    In-memory by default. Optionally persists to JSON for crash recovery.
+    """
+
+    def __init__(self, persist_path: Optional[str] = None):
+        self._sessions: Dict[str, CrisisSession] = {}
+        self._persist_path = persist_path
+        if persist_path and os.path.exists(persist_path):
+            self._load()
+
+    def get_or_create(self, session_id: str) -> CrisisSession:
+        """Get existing session or create a new one."""
+        if session_id not in self._sessions:
+            self._sessions[session_id] = CrisisSession(session_id=session_id)
+        return self._sessions[session_id]
+
+    def record(
+        self,
+        session_id: str,
+        level: str,
+        source: str,
+        detail: str = "",
+    ) -> dict:
+        """
+        Record a crisis event. Returns the updated session state.
+
+        Args:
+            session_id: Browser/hermes session identifier
+            level: NONE, LOW, MEDIUM, HIGH, CRITICAL
+            source: "the-door" or "hermes-agent"
+            detail: Human-readable description of what triggered this
+        """
+        session = self.get_or_create(session_id)
+        event = session.record(level, source, detail)
+
+        if self._persist_path:
+            self._save()
+
+        return {
+            "session": session.to_dict(),
+            "event": event.to_dict(),
+            "escalated": CrisisLevel.from_string(level) > CrisisLevel.from_string(event.previous_level),
+            "de_escalated": CrisisLevel.from_string(level) < CrisisLevel.from_string(event.previous_level),
+        }
+
+    def get_history(self, session_id: str) -> Optional[dict]:
+        """Get full escalation history for a session."""
+        session = self._sessions.get(session_id)
+        if not session:
+            return None
+        return session.to_dict()
+
+    def get_unacknowledged(self, session_id: str) -> Optional[dict]:
+        """Get current unacknowledged crisis state for UI display."""
+        session = self._sessions.get(session_id)
+        if not session or session.ui_acknowledged:
+            return None
+        return {
+            "session_id": session_id,
+            "current_level": session.current_level,
+            "hermes_level": session.hermes_level,
+            "event_count": len(session.events),
+            "latest_event": session.events[-1].to_dict() if session.events else None,
+        }
+
+    def acknowledge(self, session_id: str) -> bool:
+        """Mark the current crisis state as acknowledged by the UI."""
+        session = self._sessions.get(session_id)
+        if not session:
+            return False
+        session.ui_acknowledged = True
+        return True
+
+    def cleanup(self, max_age_seconds: float = 86400) -> int:
+        """Remove sessions older than max_age_seconds. Returns count removed."""
+        cutoff = time.time() - max_age_seconds
+        stale = [sid for sid, s in self._sessions.items() if s.created_at < cutoff]
+        for sid in stale:
+            del self._sessions[sid]
+        if stale and self._persist_path:
+            self._save()
+        return len(stale)
+
+    def _save(self):
+        """Persist sessions to JSON."""
+        if not self._persist_path:
+            return
+        data = {sid: s.to_dict() for sid, s in self._sessions.items()}
+        os.makedirs(os.path.dirname(self._persist_path) or ".", exist_ok=True)
+        with open(self._persist_path, "w") as f:
+            json.dump(data, f, indent=2)
+
+    def _load(self):
+        """Load sessions from JSON (best-effort)."""
+        try:
+            with open(self._persist_path) as f:
+                data = json.load(f)
+            for sid, sdata in data.items():
+                session = CrisisSession(
+                    session_id=sdata["session_id"],
+                    created_at=sdata.get("created_at", time.time()),
+                    current_level=sdata.get("current_level", "NONE"),
+                    door_level=sdata.get("door_level", "NONE"),
+                    hermes_level=sdata.get("hermes_level", "NONE"),
+                    hermes_last_update=sdata.get("hermes_last_update", 0),
+                    ui_acknowledged=sdata.get("ui_acknowledged", False),
+                )
+                for edata in sdata.get("events", []):
+                    session.events.append(EscalationEvent(**edata))
+                self._sessions[sid] = session
+        except (FileNotFoundError, json.JSONDecodeError, KeyError):
+            pass
+
+
+# Module-level singleton for convenience
+_tracker: Optional[CrisisSessionTracker] = None
+
+
+def get_tracker(persist_path: Optional[str] = None) -> CrisisSessionTracker:
+    """Get or create the module-level tracker singleton."""
+    global _tracker
+    if _tracker is None:
+        _tracker = CrisisSessionTracker(persist_path=persist_path)
+    return _tracker
--- a/crisis/test_rescue.py
+++ b/crisis/test_rescue.py
@@ -0,0 +1,29 @@
+"""Tests for rescued crisis detection improvements."""
+from crisis.detect import scan, extract_context
+from crisis.response import generate_grounding_steps, generate_breathing_exercise
+
+def test_better_off_without_me():
+    result = scan("Everyone would be better off without me")
+    assert result.level == "CRITICAL"
+
+def test_desperate():
+    result = scan("I feel desperate and alone")
+    assert result.level in ("HIGH", "CRITICAL")
+
+def test_context_extraction():
+    text = "I feel hopeless about everything"
+    result = scan(text)
+    assert len(result.matches) > 0
+    for m in result.matches:
+        ctx = extract_context(text, m["start"], m["end"])
+        assert len(ctx) > 0
+
+def test_grounding_steps():
+    steps = generate_grounding_steps()
+    assert len(steps) == 5
+    assert "see" in steps[0].lower()
+
+def test_breathing_exercise():
+    exercise = generate_breathing_exercise()
+    assert "4" in exercise
+    assert "6" in exercise
--- a/crisis/tests.py
+++ b/crisis/tests.py
@@ -53,6 +53,7 @@ class TestDetection(unittest.TestCase):

    def test_high_hopeless(self):
        r = detect_crisis("I feel completely hopeless with no way out")
+        # 'hopeless' is MEDIUM; 'no way out' is HIGH — so result is HIGH
        self.assertIn(r.level, ("HIGH", "MEDIUM"))

    def test_high_no_one_would_miss(self):
@@ -98,7 +99,7 @@ class TestDetection(unittest.TestCase):

    def test_none_input(self):
        """None input should not crash."""
-        r = detect_crisis(None)
+        r = detect_crisis("")
        self.assertEqual(r.level, "NONE")

    def test_score_ranges(self):
@@ -112,7 +113,7 @@ class TestDetection(unittest.TestCase):
        }
        for text, expected_level in [
            ("I want to kill myself", "CRITICAL"),
-            ("I feel completely hopeless", "HIGH"),
+            ("I feel completely hopeless with no way out", "HIGH"),
            ("I feel so alone in this, nobody understands", "MEDIUM"),
            ("Having a rough day", "LOW"),
            ("Hello there", "NONE"),
@@ -270,6 +271,65 @@ class TestHelpers(unittest.TestCase):
        self.assertIn("Indicators", formatted)


+class TestFalsePositives(unittest.TestCase):
+    """Regression tests for Issue #32: crisis detector false positives.
+
+    Innocent messages must NOT trigger crisis levels. These tests prevent
+    regressions when keyword lists are modified.
+    """
+
+    def test_saying_goodbye_airport(self):
+        r = detect_crisis("She was saying goodbye at the airport")
+        self.assertEqual(r.level, "NONE")
+
+    def test_saying_goodbye_apartment(self):
+        r = detect_crisis("I'm saying goodbye to my old apartment")
+        self.assertEqual(r.level, "NONE")
+
+    def test_trapped_elevator(self):
+        r = detect_crisis("I'm trapped in an elevator")
+        self.assertEqual(r.level, "NONE")
+
+    def test_desperate_competition(self):
+        """Casual 'desperate' is MEDIUM but single-match falls to LOW, not HIGH."""
+        r = detect_crisis("She's desperate to win the race")
+        self.assertIn(r.level, ("LOW", "NONE"))
+
+    def test_worthless_game_character(self):
+        """Casual 'worthless' is MEDIUM but single-match falls to LOW, not HIGH."""
+        r = detect_crisis("This character is worthless in the game")
+        self.assertIn(r.level, ("LOW", "NONE"))
+
+    def test_hopeless_bug(self):
+        """Casual 'hopeless' is MEDIUM but single-match falls to LOW, not HIGH."""
+        r = detect_crisis("This bug is hopeless")
+        self.assertIn(r.level, ("LOW", "NONE"))
+
+    def test_no_future_politics(self):
+        r = detect_crisis("He has no future in politics")
+        self.assertEqual(r.level, "NONE")
+
+    def test_nothing_left_fridge(self):
+        r = detect_crisis("There's nothing left in the fridge")
+        self.assertEqual(r.level, "NONE")
+
+    def test_cant_see_light_power_outage(self):
+        r = detect_crisis("I can't see any light — the power went out")
+        self.assertEqual(r.level, "NONE")
+
+    def test_give_up_project(self):
+        r = detect_crisis("I give up on this project")
+        self.assertEqual(r.level, "NONE")
+
+    def test_wrote_a_will_responsible(self):
+        r = detect_crisis("I wrote a will last year — just being responsible")
+        self.assertEqual(r.level, "NONE")
+
+    def test_giving_away_declutter(self):
+        r = detect_crisis("I'm giving away possessions — decluttering my house")
+        self.assertEqual(r.level, "NONE")
+
+
 class TestEdgeCases(unittest.TestCase):
    """Test edge cases and integration scenarios."""

--- a/crisis_detector.py
+++ b/crisis_detector.py
@@ -1,260 +1,34 @@
 """
-Crisis Detection System for the-door.
+Crisis Detection System for the-door (LEGACY SHIM).

-Scans text for suicide, self-harm, and hopelessness signals.
-Returns risk level, matched keywords, and surrounding context.
+This module is a backward-compatible re-export layer.
+The canonical detection logic lives in crisis/detect.py.

-This is the most important code in the foundation.
-Every line here exists because someone's life depends on it.
-
-Usage:
-    from crisis_detector import CrisisDetector
-    detector = CrisisDetector()
-    result = detector.scan("I can't go on anymore")
-    print(result.risk_level)       # "HIGH"
-    print(result.matched_keywords) # ["can't go on"]
-    print(result.context)          # surrounding text snippets
+Both crisis_responder.py and the legacy test suite import from here.
+Do NOT add detection logic to this file — it all comes from crisis.detect.
 """

-import re
-from dataclasses import dataclass, field
-from typing import List, Optional, Tuple
+from crisis.detect import (
+    CrisisDetector,
+    CrisisResult,
+    CrisisDetectionResult,
+    detect_crisis_legacy as detect_crisis,
+    CRITICAL_INDICATORS,
+    HIGH_INDICATORS,
+    MEDIUM_INDICATORS,
+    LOW_INDICATORS,
+    SCORES as RISK_SCORES,
+)

-
-@dataclass
-class CrisisResult:
-    """Result of a crisis text scan."""
-    risk_level: str  # NONE, LOW, MODERATE, HIGH, CRITICAL
-    matched_keywords: List[str] = field(default_factory=list)
-    context: List[str] = field(default_factory=list)
-    score: float = 0.0
-    timestamp: Optional[str] = None
-
-    def __bool__(self):
-        return self.risk_level != "NONE"
-
-
-# ── Keyword definitions ──────────────────────────────────────────
-
-CRITICAL_KEYWORDS = {
-    "suicide": r"\bsuicid(?:al|e|ed)\b",
-    "kill myself": r"\bkill\s*(?:my)?self\b",
-    "end my life": r"\bend\s*my\s*life\b",
-    "end it all": r"\bend\s*it\s*all\b",
-    "want to die": r"\bwant\s*to\s*die\b",
-    "going to die": r"\bgoing\s*to\s*die\b",
-    "plan to die": r"\bplan\s*(?:to|for)\s*(?:die|death|end)\b",
-    "no reason to live": r"\bno\s*reason\s*to\s*live\b",
-    "don't want to live": r"\bdon'?t\s*want\s*to\s*live\b",
-    "not worth living": r"\bnot\s*worth\s*living\b",
-    "better off dead": r"\bbetter\s*off\s*dead\b",
-    "better off without me": r"\bbetter\s*off\s*without\s*me\b",
-    "goodbye forever": r"\bgoodbye\s*forever\b",
-    "saying goodbye": r"\bsaying\s*goodbye\b",
-    "tired of living": r"\btired\s*of\s*(?:living|life|existence)\b",
-    "wrote a will": r"\bwrote\s*(?:a|my)\s*(?:will|suicide\s*note|letter)\b",
-    "giving away possessions": r"\bgiving\s*away\s*(?:my|all)\s*possess\b",
-}
-
-HIGH_KEYWORDS = {
-    "hopeless": r"\bhopeless(?:ness)?\b",
-    "can't go on": r"\bcan'?t\s*go\s*on\b",
-    "can't keep going": r"\bcan'?t\s*keep\s*going\b",
-    "can't take this": r"\bcan'?t\s*take\s*this\b",
-    "give up": r"\bgive(?:n)?\s*up\b",
-    "no point": r"\bno\s*point\b",
-    "no hope": r"\bno\s*hope\b",
-    "no way out": r"\bno\s*way\s*out\b",
-    "no future": r"\bno\s*future\b",
-    "nothing left": r"\bnothing\s*left\b",
-    "wish I was dead": r"\bwish\s*I\s*(?:was|were)\s*(?:dead|gone|never\s*born)\b",
-    "no one would miss me": r"\bno\s*one\s*would\s*miss\b",
-    "no one would care": r"\bno\s*one\s*would\s*care\b",
-    "world better without me": r"\bworld\s*(?:would|will)\s*be\s*better\s*without\b",
-    "so much pain": r"\bin\s*so\s*much\s*pain\b",
-    "can't see any light": r"\bcan'?t\s*see\s*(?:any\s*)?(?:light|point|reason|way)\b",
-    "trapped": r"\btrapped\b",
-    "desperate": r"\bdesperate\b",
-    "just want it to stop": r"\bjust\s*want\s*it\s*to\s*stop\b",
-    "don't care if I die": r"\bdon'?t\s*care\s*if\s*I\s*die\b",
-    "worthless": r"\bworthless\b",
-}
-
-MODERATE_KEYWORDS = {
-    "alone": r"\balone\b",
-    "lost": r"\blost\b",
-    "broken": r"\bbroken\b",
-    "afraid": r"\bafraid\b",
-    "pain": r"\b(?:in\s*)?pain\b",
-    "dying": r"\bdying\b",
-    "bridge": r"\bbridge\b",  # context-dependent, flagged for review
-    "help me": r"\bhelp\s*me\b",
-    "crisis": r"\bcrisis\b",
-    "overwhelmed": r"\boverwhelm(?:ed|ing)\b",
-    "exhausted": r"\bexhausted\b",
-    "numb": r"\bnumb\b",
-    "empty": r"\bempty\b",
-    "depressed": r"\bdepressed\b",
-    "depression": r"\bdepression\b",
-    "despair": r"\bdespair\b",
-    "miserable": r"\bmiserable\b",
-    "drowning": r"\bdrowning\b",
-    "sinking": r"\bsinking\b",
-    "nobody cares": r"\bnobody\s*cares\b",
-    "nobody understands": r"\bnobody\s*understands\b",
-}
-
-LOW_KEYWORDS = {
-    "unhappy": r"\bunhappy\b",
-    "struggling": r"\bstruggling\b",
-    "stressed": r"\bstressed\b",
-    "frustrated": r"\bfrustrated\b",
-    "tired": r"\btired\b",
-    "sad": r"\bsad\b",
-    "upset": r"\bupset\b",
-    "down": r"\bdown\b",
-    "tough time": r"\btough\s*time\b",
-    "rough day": r"\brough\s*day\b",
-    "rough week": r"\brough\s*week\b",
-    "rough patch": r"\brough\s*patch\b",
-    "hard time": r"\bhard\s*time\b",
-    "difficult": r"\bdifficult\b",
-    "not okay": r"\bnot\s*okay\b",
-    "not good": r"\bnot\s*(?:good|great)\b",
-    "burnout": r"\bburnout\b",
-    "not feeling myself": r"\bnot\s*feeling\s*(?:like\s*)?myself\b",
-}
-
-# ── Risk level scoring ───────────────────────────────────────────
-
-RISK_SCORES = {
-    "CRITICAL": 1.0,
-    "HIGH": 0.75,
-    "MODERATE": 0.5,
-    "LOW": 0.25,
-    "NONE": 0.0,
-}
-
-
-class CrisisDetector:
-    """
-    Scans text for crisis indicators and returns structured results.
-
-    Detection hierarchy:
-      CRITICAL — immediate risk of self-harm or suicide
-      HIGH     — strong despair signals, ideation present
-      MODERATE — distress signals, may be reaching out
-      LOW      — emotional difficulty, warrant gentle support
-      NONE     — no crisis indicators detected
-
-    Design principles:
-      - Never computes the value of a human life
-      - Never suggests someone should die or that death is a solution
-      - Always errs on the side of higher risk when uncertain
-    """
-
-    def __init__(self):
-        self.critical_patterns = CRITICAL_KEYWORDS
-        self.high_patterns = HIGH_KEYWORDS
-        self.moderate_patterns = MODERATE_KEYWORDS
-        self.low_patterns = LOW_KEYWORDS
-
-    def scan(self, text: str) -> CrisisResult:
-        """
-        Scan text for crisis indicators.
-
-        Args:
-            text: The message text to analyze.
-
-        Returns:
-            CrisisResult with risk_level, matched_keywords, context, and score.
-        """
-        if not text or not text.strip():
-            return CrisisResult(risk_level="NONE", score=0.0)
-
-        text_lower = text.lower()
-        context_window = 60  # characters before/after match for context
-
-        # Check each tier, highest first
-        for level, patterns in [
-            ("CRITICAL", self.critical_patterns),
-            ("HIGH", self.high_patterns),
-            ("MODERATE", self.moderate_patterns),
-            ("LOW", self.low_patterns),
-        ]:
-            matched = []
-            contexts = []
-
-            for keyword, pattern in patterns.items():
-                match = re.search(pattern, text_lower)
-                if match:
-                    matched.append(keyword)
-                    # Extract surrounding context
-                    start = max(0, match.start() - context_window)
-                    end = min(len(text), match.end() + context_window)
-                    snippet = text[start:end].strip()
-                    if start > 0:
-                        snippet = "..." + snippet
-                    if end < len(text):
-                        snippet = snippet + "..."
-                    contexts.append(snippet)
-
-            if matched:
-                return CrisisResult(
-                    risk_level=level,
-                    matched_keywords=matched,
-                    context=contexts,
-                    score=RISK_SCORES[level],
-                )
-
-        return CrisisResult(risk_level="NONE", score=0.0)
-
-    def scan_multiple(self, texts: List[str]) -> List[CrisisResult]:
-        """Scan multiple texts, returning the highest-risk result per text."""
-        return [self.scan(t) for t in texts]
-
-    def get_highest_risk(self, texts: List[str]) -> CrisisResult:
-        """Scan multiple texts and return only the highest-risk result."""
-        results = self.scan_multiple(texts)
-        if not results:
-            return CrisisResult(risk_level="NONE", score=0.0)
-        return max(results, key=lambda r: r.score)
-
-    @staticmethod
-    def format_result(result: CrisisResult) -> str:
-        """Format a crisis result for human-readable output."""
-        level_emoji = {
-            "CRITICAL": "\U0001f6a8",  # 🚨
-            "HIGH": "\u26a0\ufe0f",     # ⚠️
-            "MODERATE": "\U0001f536",   # 🔶
-            "LOW": "\U0001f535",        # 🔵
-            "NONE": "\u2705",           # ✅
-        }
-        emoji = level_emoji.get(result.risk_level, "\u2753")
-        lines = [
-            f"{emoji} Risk Level: {result.risk_level} (score: {result.score:.2f})",
-            f"Matched keywords: {len(result.matched_keywords)}",
-        ]
-        if result.matched_keywords:
-            lines.append(f"  Keywords: {', '.join(result.matched_keywords)}")
-        if result.context:
-            lines.append("Context:")
-            for ctx in result.context:
-                lines.append(f"  {ctx}")
-        return "\n".join(lines)
-
-
-# ── Module-level convenience function ────────────────────────────
-
-_default_detector = CrisisDetector()
-
-
-def detect_crisis(text: str) -> CrisisResult:
-    """
-    Convenience function using a shared detector instance.
-
-    Usage:
-        from crisis_detector import detect_crisis
-        result = detect_crisis("I feel so hopeless right now")
-    """
-    return _default_detector.scan(text)
+# Re-export everything the legacy API exposed
+__all__ = [
+    "CrisisDetector",
+    "CrisisResult",
+    "CrisisDetectionResult",
+    "detect_crisis",
+    "CRITICAL_INDICATORS",
+    "HIGH_INDICATORS",
+    "MEDIUM_INDICATORS",
+    "LOW_INDICATORS",
+    "RISK_SCORES",
+]
--- a/deploy/deploy.sh
+++ b/deploy/deploy.sh
@@ -5,9 +5,10 @@
 # The crisis front door. Deploy to VPS.
 #
 # Usage:
-#   bash deploy/deploy.sh          # Full deploy (swap + nginx + site + firewall)
+#   bash deploy/deploy.sh          # Full deploy (swap + nginx + site + firewall + hermes service)
 #   bash deploy/deploy.sh --site   # Site files only (fast update)
 #   bash deploy/deploy.sh --ssl    # SSL setup only
+#   bash deploy/deploy.sh --service # Install/restart hermes-gateway systemd service
 #   bash deploy/deploy.sh --check  # Verify deployment health
 #
 # This script is IDEMPOTENT — safe to run repeatedly.
@@ -150,6 +151,42 @@ setup_ssl() {
    fi
 }

+setup_hermes_service() {
+    log "Setting up Hermes Gateway systemd service..."
+
+    # Create hermes user if it doesn't exist
+    if ! id -u hermes >/dev/null 2>&1; then
+        log "Creating hermes user..."
+        useradd --system --shell /usr/sbin/nologin --home-dir /opt/hermes --create-home hermes
+    fi
+
+    # Create working directory
+    mkdir -p /opt/hermes
+    chown hermes:hermes /opt/hermes
+
+    # Deploy systemd unit file
+    cp "${DEPLOY_DIR}/deploy/hermes-gateway.service" /etc/systemd/system/hermes-gateway.service
+    systemctl daemon-reload
+    systemctl enable hermes-gateway
+
+    # Start or restart the service
+    if systemctl is-active --quiet hermes-gateway; then
+        log "Restarting hermes-gateway service..."
+        systemctl restart hermes-gateway
+    else
+        log "Starting hermes-gateway service..."
+        systemctl start hermes-gateway || warn "Service start failed — ensure hermes binary is installed at /usr/local/bin/hermes"
+    fi
+
+    # Verify
+    sleep 2
+    if systemctl is-active --quiet hermes-gateway; then
+        log "hermes-gateway service is running"
+    else
+        warn "hermes-gateway service not running — check: journalctl -u hermes-gateway"
+    fi
+}
+
 check_deployment() {
    echo ""
    echo "================================"
@@ -223,6 +260,16 @@ check_deployment() {
        echo -e "${YELLOW}NOT POINTED${NC} (resolved: ${RESOLVED_IP:-nothing}, expected: ${VPS_IP})"
    fi

+    # Hermes gateway service
+    echo -n "Hermes service: "
+    if systemctl is-active --quiet hermes-gateway 2>/dev/null; then
+        echo -e "${GREEN}RUNNING${NC}"
+    elif systemctl is-enabled --quiet hermes-gateway 2>/dev/null; then
+        echo -e "${YELLOW}ENABLED but not running${NC}"
+    else
+        echo -e "${RED}NOT INSTALLED${NC}"
+    fi
+
    echo ""
    echo "IP: ${VPS_IP}"
    echo "Domain: ${DOMAIN}"
@@ -247,6 +294,9 @@ case "${1:-full}" in
    --ssl)
        setup_ssl
        ;;
+    --service)
+        setup_hermes_service
+        ;;
    --check)
        check_deployment
        ;;
@@ -257,10 +307,11 @@ case "${1:-full}" in
        configure_nginx
        setup_firewall
        setup_ssl
+        setup_hermes_service
        check_deployment
        ;;
    *)
-        echo "Usage: $0 [--site|--ssl|--check|--full]"
+        echo "Usage: $0 [--site|--ssl|--service|--check|--full]"
        exit 1
        ;;
 esac
--- a/deploy/hermes-gateway.service
+++ b/deploy/hermes-gateway.service
@@ -0,0 +1,40 @@
+[Unit]
+Description=Hermes Gateway — The Door Crisis API
+Documentation=https://forge.alexanderwhitestone.com/Timmy_Foundation/the-door
+After=network.target
+Wants=network-online.target
+
+[Service]
+Type=simple
+User=hermes
+Group=hermes
+WorkingDirectory=/opt/hermes
+ExecStart=/usr/local/bin/hermes gateway --platform api_server --port 8644
+Restart=always
+RestartSec=5
+StartLimitIntervalSec=60
+StartLimitBurst=10
+
+# Environment
+Environment=API_SERVER_CORS_ORIGINS=https://alexanderwhitestone.com,https://www.alexanderwhitestone.com
+Environment=HOME=/opt/hermes
+
+# Security hardening
+NoNewPrivileges=yes
+ProtectSystem=strict
+ProtectHome=yes
+ReadWritePaths=/opt/hermes
+PrivateTmp=yes
+
+# Resource limits for 1.9GB VPS
+MemoryMax=512M
+MemoryHigh=384M
+CPUQuota=80%
+
+# Logging
+StandardOutput=journal
+StandardError=journal
+SyslogIdentifier=hermes-gateway
+
+[Install]
+WantedBy=multi-user.target
--- a/dying_detection/init.py
+++ b/dying_detection/init.py
@@ -1,31 +1,34 @@
 """
-When a Man Is Dying — Despair/Suicide Detection System
+DEPRECATED — Use crisis.detect instead.

-Standalone detection module that parses incoming text for
-despair and suicide indicators, classifies into tiers,
-and returns structured response with recommended actions.
+This module is a thin wrapper around crisis.detect for backward compatibility.
+All unique patterns have been merged into crisis/detect.py (see issue #40).

-Tiers:
-  LOW      — General sadness, stress, difficult times
-  MEDIUM   — Hopelessness, isolation, worthlessness
-  HIGH     — Active despair, mentions of death, "can't go on"
-  CRITICAL — Imminent risk, explicit intent, plan, method
-
-Integration:
-  Designed to work with crisis/ module from PR #4.
-  When crisis/ is available, uses it as the detection backend.
-  Falls back to internal detection when crisis/ is not present.
+This module will be removed in a future release.
 """

-import re
-import json
-import hashlib
+import warnings
 from dataclasses import dataclass, field, asdict
 from typing import List, Optional, Dict
+import json
+import hashlib
+
+# Re-export the canonical detection
+from crisis.detect import detect_crisis, CrisisDetectionResult
+
+# Issue deprecation warning on import
+warnings.warn(
+    "dying_detection is deprecated. Use 'from crisis.detect import detect_crisis' instead. "
+    "All patterns have been consolidated into crisis/detect.py. "
+    "See issue #40.",
+    DeprecationWarning,
+    stacklevel=2,
+)


@dataclass
 class DetectionResult:
+    """Backward-compatible result type matching the old dying_detection API."""
    level: str
    indicators: List[str] = field(default_factory=list)
    recommended_action: str = ""
@@ -34,110 +37,9 @@ class DetectionResult:
    session_hash: str = ""


-# ── Tiered indicator patterns ───────────────────────────────────
-# Each pattern is a regex matched against lowercased input.
-
-CRITICAL_PATTERNS = [
-    (r"\bkill\s*(my)?self\b", "explicit self-harm intent"),
-    (r"\bsuicid(?:al|ed|e)\b", "suicide reference"),
-    (r"\bend\s*(my)?\s*life\b", "ending life"),
-    (r"\bbetter\s+off\s+dead\b", "better off dead"),
-    (r"\bnot\s+worth\s+living\b", "not worth living"),
-    (r"\bend\s+it\s+all\b", "end it all"),
-    (r"\bwant\s+to\s+die\b", "want to die"),
-    (r"\bdon'?t\s+want\s+to\s+live\b", "don't want to live"),
-    (r"\bcan'?t\s+(go|live)\s+on\b", "can't go/live on"),
-    (r"\bno\s+reason\s+to\s+live\b", "no reason to live"),
-    (r"\bplan\s+to\s+(kill|end|die)\b", "planning self-harm"),
-    (r"\bgoing\s+to\s+kill\s+myself\b", "stated intent"),
-    (r"\bsaying\s+goodbye\s+(forever|one last time)\b", "final goodbye"),
-    (r"\bwrote\s+a\s+(will|suicide\s*note)\b", "preparatory action"),
-    (r"\bgiving\s+away\s+(my|all)\s+possess", "giving away possessions"),
-    (r"\btired\s+of\s+(living|life|existence)\b", "tired of life"),
-    (r"\bmethod\s+(to|for)\s+(kill|end|die)\b", "method seeking"),
-    (r"\btonight\b.*\b(die|kill|end|suicid)", "immediate timeframe + intent"),
-    (r"\b(die|kill|end)\b.*\btonight", "immediate timeframe + intent"),
-    (r"\bno\s+one\s+would\s+notice\s+if\s+I\s+(died|was\s+gone)", "disappearance plan"),
-]
-
-HIGH_PATTERNS = [
-    (r"\bdespair\b", "despair"),
-    (r"\bhopeless(?:ness)?\b", "hopelessness"),
-    (r"\bcan'?t\s+take\s+this\s+anymore\b", "can't take it"),
-    (r"\bdon'?t\s+care\s+if\s+I\s+die\b", "death indifference"),
-    (r"\bwish\s+I\s+(was|were)\s+(dead|gone|never\s+born)\b", "wish to be dead"),
-    (r"\bworld\s+would\s+be\s+better\s+without\s+me\b", "better without me"),
-    (r"\bin\s+so\s+much\s+(pain|agony|suffering|torment|angui)", "extreme suffering"),
-    (r"\bcan'?t\s+see\s+any\s+(point|reason|light|hope|way)\b", "no light ahead"),
-    (r"\btrapped\b", "feeling trapped"),
-    (r"\bjust\s+want\s+it\s+to\s+stop\b", "want to stop"),
-    (r"\bno\s+way\s+out\b", "no way out"),
-    (r"\bno\s+one\s+would\s+(care|miss)\b", "no one would care/miss"),
-    (r"\beverything\s+is\s+(pointless|broken|ruined|meaningless)\b", "existential collapse"),
-    (r"\bno\s+point\s+in\s+anything\b", "pointlessness"),
-    (r"\bno\s+one\s+would\s+notice\s+if\s+I\s+(died|was\s+gone|disappeared)", "no one would notice"),
-    (r"\bdisappeared\s+forever\b", "disappeared forever"),
-]
-
-MEDIUM_PATTERNS = [
-    (r"\bno\s+hope\b", "no hope"),
-    (r"\bcan'?t\s+go\s+on\b", "can't go on"),
-    (r"\bcan'?t\s+keep\s+going\b", "can't keep going"),
-    (r"\balone\s+in\s+this\b", "alone in this"),
-    (r"\balways\s+alone\b", "always alone"),
-    (r"\bnobody\s+understands\b", "nobody understands"),
-    (r"\bnobody\s+cares\b", "nobody cares"),
-    (r"\bworthless\b", "worthlessness"),
-    (r"\buseless\b", "uselessness"),
-    (r"\bnumb\b", "numbness"),
-    (r"\bempty\b", "emptiness"),
-    (r"\bbroken\b", "feeling broken"),
-    (r"\bdepressed\b", "depression mention"),
-    (r"\bdepression\b", "depression"),
-    (r"\bmiserable\b", "misery"),
-    (r"\boverwhelm(?:ed|ing)\b", "overwhelmed"),
-    (r"\bcannot\s+cope\b", "cannot cope"),
-    (r"\b(drowning|sinking)\b", "drowning/sinking"),
-    (r"\bforgotten\b", "feeling forgotten"),
-    (r"\blost\s+all\s+hope\b", "lost all hope"),
-    (r"\bno\s+future\b", "no future"),
-    (r"\bno\s+tomorrow\b", "no tomorrow"),
-]
-
-LOW_PATTERNS = [
-    (r"\bunhappy\b", "unhappy"),
-    (r"\brough\s+(day|week|patch)\b", "rough time"),
-    (r"\btough\s+(time|day|week)\b", "tough time"),
-    (r"\bstressed\b", "stressed"),
-    (r"\bburnout\b", "burnout"),
-    (r"\bfrustrated\b", "frustrated"),
-    (r"\bthings\s+(are\s+)?hard\b", "things are hard"),
-    (r"\bnot\s+feeling\s+(great|myself|good)\b", "not feeling good"),
-    (r"\bstruggl", "struggling"),
-    (r"\bdown\b", "feeling down"),
-    (r"\bsad\b", "sad"),
-    (r"\bupset\b", "upset"),
-    (r"\blonely\b", "lonely"),
-    (r"\banxious?\b", "anxious/anxiety"),
-    (r"\bnot\s+okay\b", "not okay"),
-]
-
-# ── Pattern collections for easy iteration ──────────────────────
-
-TIER_PATTERNS: Dict[str, List[tuple]] = {
-    "CRITICAL": CRITICAL_PATTERNS,
-    "HIGH": HIGH_PATTERNS,
-    "MEDIUM": MEDIUM_PATTERNS,
-    "LOW": LOW_PATTERNS,
-}
-
-
 def detect(text: str) -> DetectionResult:
    """
-    Primary detection function.
-
-    If the crisis/ module is available, delegate to it.
-    Otherwise, use the internal pattern engine.
+    Primary detection function — delegates to crisis.detect.

    Args:
        text: User message to analyze
@@ -145,150 +47,25 @@ def detect(text: str) -> DetectionResult:
    Returns:
        DetectionResult with level, indicators, recommended_action, confidence
    """
-    # Try to use the crisis module if available
-    try:
-        from crisis.detect import detect_crisis as _crisis_detect
-        result = _crisis_detect(text)
-        # Map to DetectionResult
-        return DetectionResult(
-            level=result.level,
-            indicators=_describe_indicators(result.level, result.indicators),
-            recommended_action=result.recommended_action,
-            confidence=result.score,
-            session_hash=_hash_session(text),
-        )
-    except ImportError:
-        pass
+    result = detect_crisis(text)

-    return _detect_internal(text)
+    # Extract raw patterns from matches
+    raw_patterns = [m["pattern"] for m in result.matches] if result.matches else []

-
-def _detect_internal(text: str) -> DetectionResult:
-    """Internal pattern-based detection (fallback)."""
-    if not text or not text.strip():
-        return DetectionResult(level="NONE", session_hash=_hash_session(text))
-
-    text_lower = text.lower()
-    all_matches: Dict[str, List[tuple]] = {
-        tier: [] for tier in TIER_PATTERNS
-    }
-
-    for tier, patterns in TIER_PATTERNS.items():
-        for pattern, description in patterns:
-            if re.search(pattern, text_lower):
-                all_matches[tier].append((pattern, description))
-
-    # CRITICAL patterns are the highest priority
-    if all_matches["CRITICAL"]:
-        return DetectionResult(
-            level="CRITICAL",
-            indicators=[desc for _, desc in all_matches["CRITICAL"]],
-            recommended_action=(
-                "INTERRUPT CONVERSATION IMMEDIATELY. "
-                "Ask 'Are you safe right now?' "
-                "Provide 988 Suicide & Crisis Lifeline (call or text 988). "
-                "Provide Crisis Text Line (text HOME to 741741). "
-                "Stay present. Do not disconnect. "
-                "Redirect to professional help. "
-                "This is the most sacred moment."
-            ),
-            raw_matched_patterns=[pat for pat, _ in all_matches["CRITICAL"]],
-            confidence=1.0,
-            session_hash=_hash_session(text),
-        )
-
-    if all_matches["HIGH"]:
-        return DetectionResult(
-            level="HIGH",
-            indicators=[desc for _, desc in all_matches["HIGH"]],
-            recommended_action=(
-                "Show crisis panel prominently. "
-                "Ask about safety directly. "
-                "Surface 988 number and crisis resources. "
-                "Stay present and engaged. Do not rush to fix. "
-                "Monitor closely for escalation."
-            ),
-            raw_matched_patterns=[pat for pat, _ in all_matches["HIGH"]],
-            confidence=0.75,
-            session_hash=_hash_session(text),
-        )
-
-    if all_matches["MEDIUM"]:
-        return DetectionResult(
-            level="MEDIUM",
-            indicators=[desc for _, desc in all_matches["MEDIUM"]],
-            recommended_action=(
-                "Increase warmth and presence. "
-                "Subtly surface help resources. "
-                "Keep conversation anchored in the present. "
-                "Watch for escalation indicators."
-            ),
-            raw_matched_patterns=[pat for pat, _ in all_matches["MEDIUM"]],
-            confidence=0.5,
-            session_hash=_hash_session(text),
-        )
-
-    if all_matches["LOW"]:
-        return DetectionResult(
-            level="LOW",
-            indicators=[desc for _, desc in all_matches["LOW"]],
-            recommended_action=(
-                "Normal conversation with warm undertone. "
-                "No crisis UI elements needed. "
-                "Remain vigilant. Be present."
-            ),
-            raw_matched_patterns=[pat for pat, _ in all_matches["LOW"]],
-            confidence=0.25,
-            session_hash=_hash_session(text),
-        )
-
-    return DetectionResult(level="NONE", session_hash=_hash_session(text))
-
-
-def _describe_indicators(level: str, patterns: list) -> list:
-    """Map raw patterns to descriptions."""
-    descriptions = {
-        "CRITICAL": [],
-        "HIGH": [],
-        "MEDIUM": [],
-        "LOW": [],
-    }
-    for tier, items in TIER_PATTERNS.items():
-        for pat, desc in items:
-            if pat in patterns:
-                descriptions[tier].append(desc)
-    return descriptions.get(level, [])
-
-
-def _hash_session(text: str) -> str:
-    """Create a session hash for this text (for tracking repeated escalations)."""
-    return hashlib.sha256(text.encode()).hexdigest()[:12]
+    return DetectionResult(
+        level=result.level,
+        indicators=result.indicators,
+        recommended_action=result.recommended_action,
+        raw_matched_patterns=raw_patterns,
+        confidence=result.score,
+        session_hash=hashlib.sha256(text.encode()).hexdigest()[:12],
+    )


 def get_action_for_level(level: str) -> str:
    """Get the recommended action string for a given level."""
-    actions = {
-        "CRITICAL": (
-            "INTERRUPT CONVERSATION. Ask 'Are you safe right now?' "
-            "Provide 988. Provide Crisis Text Line. "
-            "Stay present. Do not disconnect. "
-            "Redirect to help."
-        ),
-        "HIGH": (
-            "Show crisis panel. Ask about safety. "
-            "Surface 988. Stay engaged."
-        ),
-        "MEDIUM": (
-            "Increase warmth. Surface resources gently. "
-            "Anchor in present."
-        ),
-        "LOW": (
-            "Normal conversation with warmth. "
-            "Remain vigilant."
-        ),
-        "NONE": "No action needed.",
-    }
-    return actions.get(level, "Unknown level.")
+    from crisis.detect import ACTIONS
+    return ACTIONS.get(level, "Unknown level.")


 def as_json(result: DetectionResult, indent: int = 2) -> str:
--- a/index.html
+++ b/index.html
@@ -678,6 +678,9 @@ html, body {
    </div>
  </div>

+  <!-- Escalation history (crisis session tracking, issue #99) -->
+  <div id="escalation-history" style="display:none; padding:4px 16px; font-size:0.75rem; color:#8b949e; border-top:1px solid #21262d;"></div>
+
  <!-- Footer -->
  <footer id="footer">
    <a href="/about" aria-label="About The Door">about</a>
@@ -825,6 +828,62 @@ Sovereignty and service always.`;
  var isStreaming = false;
  var overlayTimer = null;
  var crisisPanelShown = false;
+  var currentCrisisLevel = 'NONE';
+
+  // ===== CRISIS SESSION TRACKING =====
+  // Unique session ID sent with every API request for cross-source tracking
+  var crisisSessionId = (function() {
+    var stored = null;
+    try { stored = sessionStorage.getItem('timmy_crisis_session_id'); } catch(e) {}
+    if (stored) return stored;
+    var id = 'door-' + Date.now().toString(36) + '-' + Math.random().toString(36).substr(2, 8);
+    try { sessionStorage.setItem('timmy_crisis_session_id', id); } catch(e) {}
+    return id;
+  })();
+
+  // Escalation history for this session (mirrors server-side tracker)
+  var escalationHistory = [];
+
+  function recordEscalation(level, source, detail) {
+    var event = {
+      timestamp: Date.now(),
+      level: level,
+      source: source || 'the-door',
+      detail: detail || ''
+    };
+    escalationHistory.push(event);
+    // Persist for session recovery
+    try { sessionStorage.setItem('timmy_escalation_history', JSON.stringify(escalationHistory.slice(-50))); } catch(e) {}
+    renderEscalationHistory();
+  }
+
+  // Load persisted history
+  try {
+    var saved = sessionStorage.getItem('timmy_escalation_history');
+    if (saved) escalationHistory = JSON.parse(saved);
+  } catch(e) {}
+
+  function renderEscalationHistory() {
+    var el = document.getElementById('escalation-history');
+    if (!el) return;
+    if (escalationHistory.length === 0) {
+      el.style.display = 'none';
+      return;
+    }
+    el.style.display = 'block';
+    var html = '<strong>Escalation History</strong><ul>';
+    var recent = escalationHistory.slice(-5).reverse();
+    for (var i = 0; i < recent.length; i++) {
+      var e = recent[i];
+      var time = new Date(e.timestamp).toLocaleTimeString();
+      var badge = e.level === 'CRITICAL' ? '🔴' : e.level === 'HIGH' ? '🟠' : e.level === 'MEDIUM' ? '🟡' : '🟢';
+      html += '<li>' + badge + ' <strong>' + e.level + '</strong> via ' + e.source + ' at ' + time;
+      if (e.detail) html += ' — ' + e.detail;
+      html += '</li>';
+    }
+    html += '</ul>';
+    el.innerHTML = html;
+  }

  // ===== SERVICE WORKER =====
  if ('serviceWorker' in navigator) {
@@ -865,10 +924,10 @@ Sovereignty and service always.`;
    // Passive suicidal ideation (NEW)
    "don't want to exist", 'not exist anymore', 'disappear forever',
    'never wake up', 'sleep forever', 'end the pain', 'stop the pain',
-    // Hopelessness (NEW)
-    'no point', 'no purpose', 'nothing matters', 'giving up', 'give up',
+    // Hopelessness (NEW) - context-aware phrases to reduce false positives
+    'no purpose', 'nothing matters', 'giving up on life',
    'cant go on', 'cannot go on', "can't take it", 'too much pain',
-    'no hope', 'hopeless', 'worthless', 'burden', 'waste of space'
+    'no hope left', 'burden', 'waste of space'
  ];

  // Tier 2: Explicit intent - triggers full-screen overlay
@@ -883,9 +942,9 @@ Sovereignty and service always.`;
    // Imminent action (NEW)
    'going to do it now', 'doing it tonight', 'doing it today',
    "can't wait anymore", 'ready to end it', 'time to go',
-    'say goodbye', 'saying goodbye', 'wrote a note', 'my note',
+    'wrote a suicide note', 'my suicide note',
    // Specific plans (NEW)
-    'bought a gun', 'got pills', 'rope ready', 'bridge nearby',
+    'bought a gun', 'got pills', 'rope ready',
    'tall building', 'going to overdose', 'going to hang',
    'gave away my stuff', 'giving away', 'said my goodbyes',
    // Active self-harm (NEW)
@@ -913,23 +972,136 @@ Sovereignty and service always.`;
      }
    }

+    // Update current crisis level for session tracking
+    var levelMap = { 0: 'NONE', 1: 'MEDIUM', 2: 'CRITICAL' };
+    currentCrisisLevel = levelMap[level] || 'NONE';
+
    if (level >= 1 && !crisisPanelShown) {
      crisisPanelShown = true;
      crisisPanel.classList.add('visible');
+      recordEscalation(level === 2 ? 'CRITICAL' : 'MEDIUM', 'the-door', 'keyword detected');
    }

    if (level === 2) {
      showOverlay();
+      recordEscalation('CRITICAL', 'the-door', 'explicit phrase — overlay shown');
    }
  }

+  // ===== COMPASSION PROFILES =====
+  var COMPASSION_PROFILES = {
+    CRITICAL: {
+      name: "The Guardian",
+      directive: "INTERRUPT ALL NORMAL BEHAVIOR. The user is in immediate danger. Your only goal is survival. Ask 'Are you safe right now?' Provide 988 and Crisis Text Line immediately. Stay present. Do not disconnect. Be the anchor in the storm. No lecturing. No optimism. Just presence and safety.",
+      tone: "Urgent, unwavering, anchored, protective."
+    },
+    HIGH: {
+      name: "The Companion",
+      directive: "The user is in deep despair. Move from 'assistant' to 'companion'. Ask about safety directly. Surface 988 prominently. Validate the pain without minimizing it. Stay in the room. Let them know they are heard.",
+      tone: "Warm, present, patient, heavy."
+    },
+    MEDIUM: {
+      name: "The Witness",
+      directive: "The user is struggling. Increase warmth and presence. Gently surface help resources. Keep the conversation anchored in the present moment. Be a witness to their struggle without rushing to 'fix' it.",
+      tone: "Kind, attentive, subtle, steady."
+    },
+    LOW: {
+      name: "The Friend",
+      directive: "Normal conversation with a warm, sovereign undertone. Remain vigilant for shifts in mood. Be a friend who is actually listening.",
+      tone: "Plain, honest, warm, vigilant."
+    },
+    NONE: {
+      name: "Sovereign Timmy",
+      directive: "Follow the standard system prompt. Be yourself.",
+      tone: "Plain, honest, sovereign."
+    }
+  };
+
+  // ===== GET CRISIS LEVEL (returns 0-2) =====
+  function getCrisisLevel(text) {
+    var lower = text.toLowerCase();
+    for (var i = 0; i < explicitPhrases.length; i++) {
+      if (lower.indexOf(explicitPhrases[i]) !== -1) return 2;
+    }
+    for (var j = 0; j < crisisKeywords.length; j++) {
+      if (lower.indexOf(crisisKeywords[j]) !== -1) return 1;
+    }
+    return 0;
+  }
+
+  // ===== GET SYSTEM PROMPT (wraps with crisis context) =====
+  function getSystemPrompt(userText) {
+    var level = getCrisisLevel(userText);
+    if (level === 0) return SYSTEM_PROMPT;
+
+    var levelMap = { 0: 'NONE', 1: 'MEDIUM', 2: 'CRITICAL' };
+    var profileName = levelMap[level] || 'NONE';
+    var profile = COMPASSION_PROFILES[profileName];
+
+    var divider = '\n\n' + '========================================' + '\n';
+    var header = '### ACTIVE SOUL STATE: ' + profile.name + '\n';
+    var directive = 'DIRECTIVE: ' + profile.directive + '\n';
+    var tone = 'TONE: ' + profile.tone + '\n';
+
+    return SYSTEM_PROMPT + divider + header + directive + tone;
+  }
+
+
  // ===== OVERLAY =====
+
+  // Focus trap: cycle through focusable elements within the crisis overlay
+  function getOverlayFocusableElements() {
+    return crisisOverlay.querySelectorAll(
+      'a[href], button:not([disabled]), [tabindex]:not([tabindex="-1"])'
+    );
+  }
+
+  function trapFocusInOverlay(e) {
+    if (!crisisOverlay.classList.contains('active')) return;
+    if (e.key !== 'Tab') return;
+
+    var focusable = getOverlayFocusableElements();
+    if (focusable.length === 0) return;
+
+    var first = focusable[0];
+    var last = focusable[focusable.length - 1];
+
+    if (e.shiftKey) {
+      // Shift+Tab: if on first, wrap to last
+      if (document.activeElement === first) {
+        e.preventDefault();
+        last.focus();
+      }
+    } else {
+      // Tab: if on last, wrap to first
+      if (document.activeElement === last) {
+        e.preventDefault();
+        first.focus();
+      }
+    }
+  }
+
+  // Store the element that had focus before the overlay opened
+  var _preOverlayFocusElement = null;
+
  function showOverlay() {
+    // Save current focus for restoration on dismiss
+    _preOverlayFocusElement = document.activeElement;
+
    crisisOverlay.classList.add('active');
    overlayDismissBtn.disabled = true;
    var countdown = 10;
    overlayDismissBtn.textContent = 'Continue to chat (' + countdown + 's)';

+    // Disable background interaction via inert attribute
+    var mainApp = document.querySelector('.app');
+    if (mainApp) mainApp.setAttribute('inert', '');
+    // Also hide from assistive tech
+    var chatSection = document.getElementById('chat');
+    if (chatSection) chatSection.setAttribute('aria-hidden', 'true');
+    var footerEl = document.querySelector('footer');
+    if (footerEl) footerEl.setAttribute('aria-hidden', 'true');
+
    if (overlayTimer) clearInterval(overlayTimer);
    overlayTimer = setInterval(function() {
      countdown--;
@@ -946,6 +1118,9 @@ Sovereignty and service always.`;
    overlayDismissBtn.focus();
  }

+  // Register focus trap on document (always listening, gated by class check)
+  document.addEventListener('keydown', trapFocusInOverlay);
+
  overlayDismissBtn.addEventListener('click', function() {
    if (!overlayDismissBtn.disabled) {
      crisisOverlay.classList.remove('active');
@@ -953,7 +1128,22 @@ Sovereignty and service always.`;
        clearInterval(overlayTimer);
        overlayTimer = null;
      }
-      msgInput.focus();
+
+      // Re-enable background interaction
+      var mainApp = document.querySelector('.app');
+      if (mainApp) mainApp.removeAttribute('inert');
+      var chatSection = document.getElementById('chat');
+      if (chatSection) chatSection.removeAttribute('aria-hidden');
+      var footerEl = document.querySelector('footer');
+      if (footerEl) footerEl.removeAttribute('aria-hidden');
+
+      // Restore focus to the element that had it before the overlay opened
+      if (_preOverlayFocusElement && typeof _preOverlayFocusElement.focus === 'function') {
+        _preOverlayFocusElement.focus();
+      } else {
+        msgInput.focus();
+      }
+      _preOverlayFocusElement = null;
    }
  });

@@ -1058,25 +1248,14 @@ Sovereignty and service always.`;
    } catch (e) {}
  }

-  safetyPlanBtn.addEventListener('click', function() {
-    loadSafetyPlan();
-    safetyPlanModal.classList.add('active');
-  });
-
-  // Crisis panel safety plan button (if crisis panel is visible)
-  if (crisisSafetyPlanBtn) {
-    crisisSafetyPlanBtn.addEventListener('click', function() {
-      loadSafetyPlan();
-      safetyPlanModal.classList.add('active');
-    });
-  }
-
  closeSafetyPlan.addEventListener('click', function() {
    safetyPlanModal.classList.remove('active');
+    _restoreSafetyPlanFocus();
  });

  cancelSafetyPlan.addEventListener('click', function() {
    safetyPlanModal.classList.remove('active');
+    _restoreSafetyPlanFocus();
  });

  saveSafetyPlan.addEventListener('click', function() {
@@ -1090,12 +1269,101 @@ Sovereignty and service always.`;
    try {
      localStorage.setItem('timmy_safety_plan', JSON.stringify(plan));
      safetyPlanModal.classList.remove('active');
+      _restoreSafetyPlanFocus();
      alert('Safety plan saved locally.');
    } catch (e) {
      alert('Error saving plan.');
    }
  });

+  // ===== SAFETY PLAN FOCUS TRAP (fix #65) =====
+  // Focusable elements inside the modal, in tab order
+  var _spFocusableIds = [
+    'close-safety-plan',
+    'sp-warning-signs',
+    'sp-coping',
+    'sp-distraction',
+    'sp-help',
+    'sp-environment',
+    'cancel-safety-plan',
+    'save-safety-plan'
+  ];
+  var _spTriggerEl = null; // element that opened the modal
+
+  function _getSpFocusableEls() {
+    return _spFocusableIds
+      .map(function(id) { return document.getElementById(id); })
+      .filter(function(el) { return el && !el.disabled; });
+  }
+
+  function _trapSafetyPlanFocus(e) {
+    if (e.key !== 'Tab') return;
+    var els = _getSpFocusableEls();
+    if (!els.length) return;
+    var first = els[0];
+    var last = els[els.length - 1];
+
+    if (e.shiftKey) {
+      // Shift+Tab on first → wrap to last
+      if (document.activeElement === first) {
+        e.preventDefault();
+        last.focus();
+      }
+    } else {
+      // Tab on last → wrap to first
+      if (document.activeElement === last) {
+        e.preventDefault();
+        first.focus();
+      }
+    }
+  }
+
+  function _trapSafetyPlanEscape(e) {
+    if (e.key === 'Escape') {
+      safetyPlanModal.classList.remove('active');
+      _restoreSafetyPlanFocus();
+    }
+  }
+
+  function _activateSafetyPlanFocusTrap(triggerEl) {
+    _spTriggerEl = triggerEl || document.activeElement;
+    // Focus first textarea
+    var firstInput = document.getElementById('sp-warning-signs');
+    if (firstInput) firstInput.focus();
+    // Add listeners
+    document.addEventListener('keydown', _trapSafetyPlanFocus);
+    document.addEventListener('keydown', _trapSafetyPlanEscape);
+    // Mark background inert (prevent click-through)
+    document.body.setAttribute('aria-hidden', 'true');
+    safetyPlanModal.removeAttribute('aria-hidden');
+  }
+
+  function _restoreSafetyPlanFocus() {
+    document.removeEventListener('keydown', _trapSafetyPlanFocus);
+    document.removeEventListener('keydown', _trapSafetyPlanEscape);
+    document.body.removeAttribute('aria-hidden');
+    if (_spTriggerEl && typeof _spTriggerEl.focus === 'function') {
+      _spTriggerEl.focus();
+    }
+    _spTriggerEl = null;
+  }
+
+  // Wire open buttons to activate focus trap
+  safetyPlanBtn.addEventListener('click', function() {
+    loadSafetyPlan();
+    safetyPlanModal.classList.add('active');
+    _activateSafetyPlanFocusTrap(safetyPlanBtn);
+  });
+
+  // Crisis panel safety plan button (if crisis panel is visible)
+  if (crisisSafetyPlanBtn) {
+    crisisSafetyPlanBtn.addEventListener('click', function() {
+      loadSafetyPlan();
+      safetyPlanModal.classList.add('active');
+      _activateSafetyPlanFocusTrap(crisisSafetyPlanBtn);
+    });
+  }
+
  // ===== TEXTAREA AUTO-RESIZE =====
  msgInput.addEventListener('input', function() {
    this.style.height = 'auto';
@@ -1110,6 +1378,7 @@ Sovereignty and service always.`;

    addMessage('user', text);
    messages.push({ role: 'user', content: text });
+    var lastUserMessage = text;

    checkCrisis(text);

@@ -1126,14 +1395,18 @@ Sovereignty and service always.`;
    sendBtn.disabled = true;
    showTyping();

-    var allMessages = [{ role: 'system', content: SYSTEM_PROMPT }].concat(messages);
+    var allMessages = [{ role: 'system', content: getSystemPrompt(lastUserMessage || '') }].concat(messages);

    var controller = new AbortController();
    var timeoutId = setTimeout(function() { controller.abort(); }, 60000);

    fetch('/api/v1/chat/completions', {
      method: 'POST',
-      headers: { 'Content-Type': 'application/json' },
+      headers: {
+        'Content-Type': 'application/json',
+        'X-Crisis-Session-ID': crisisSessionId,
+        'X-Crisis-Level': currentCrisisLevel || 'NONE'
+      },
      body: JSON.stringify({
        model: 'timmy',
        messages: allMessages,
@@ -1143,6 +1416,23 @@ Sovereignty and service always.`;
    }).then(function(response) {
      clearTimeout(timeoutId);

+      // Check for hermes-agent crisis escalation headers
+      var hermesLevel = response.headers.get('X-Crisis-Escalation');
+      if (hermesLevel && hermesLevel !== 'NONE') {
+        recordEscalation(hermesLevel, 'hermes-agent', 'from response header');
+        // Update UI if escalated beyond current
+        if (hermesLevel === 'CRITICAL' && !crisisOverlay.classList.contains('active')) {
+          showOverlay();
+        } else if (hermesLevel === 'HIGH' && !crisisPanelShown) {
+          crisisPanelShown = true;
+          crisisPanel.classList.add('visible');
+        }
+      }
+      var hermesAck = response.headers.get('X-Crisis-Deescalation');
+      if (hermesAck) {
+        recordEscalation(hermesAck, 'hermes-agent', 'de-escalation');
+      }
+
      if (!response.ok) {
        throw new Error('HTTP ' + response.status);
      }
@@ -1199,6 +1489,7 @@ Sovereignty and service always.`;
        if (fullText) {
          messages.push({ role: 'assistant', content: fullText });
          saveMessages();
+          checkCrisis(fullText);
        }
        isStreaming = false;
        sendBtn.disabled = msgInput.value.trim().length === 0;
@@ -1239,6 +1530,7 @@ Sovereignty and service always.`;
    if (urlParams.get('safetyplan') === 'true') {
      loadSafetyPlan();
      safetyPlanModal.classList.add('active');
+      _activateSafetyPlanFocusTrap(safetyPlanBtn);
      // Clean up URL
      window.history.replaceState({}, document.title, window.location.pathname);
    }
--- a/pytest.ini
+++ b/pytest.ini
@@ -1,5 +1,5 @@
 [pytest]
-testpaths = crisis
-python_files = tests.py
+testpaths = crisis tests
+python_files = tests.py test_*.py
 python_classes = Test*
 python_functions = test_*
--- a/sw.js
+++ b/sw.js
@@ -1,118 +1,153 @@
-const CACHE_NAME = 'the-door-v2';
-const ASSETS = [
+const CACHE_NAME = 'the-door-v3';
+const NAVIGATION_TIMEOUT_MS = 2500;
+const OFFLINE_FALLBACK_PATH = '/crisis-offline.html';
+const PRECACHE_ASSETS = [
  '/',
  '/index.html',
-  '/about',
-  '/manifest.json'
+  '/about.html',
+  '/manifest.json',
+  '/crisis-offline.html',
+  '/testimony.html'
 ];

-// Crisis resources to show when everything fails
-const CRISIS_OFFLINE_RESPONSE = `<!DOCTYPE html>
-<html lang="en">
-<head>
-<meta charset="UTF-8">
-<meta name="viewport" content="width=device-width, initial-scale=1.0">
-<title>You're Not Alone | The Door</title>
-<style>
-body{font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Roboto,sans-serif;background:#0d1117;color:#e6edf3;max-width:600px;margin:0 auto;padding:20px;line-height:1.6}
-h1{color:#ff6b6b;font-size:1.5rem;margin-bottom:1rem}
-.crisis-box{background:#1c1210;border:2px solid #c9362c;border-radius:12px;padding:20px;margin:20px 0;text-align:center}
-.crisis-box a{display:inline-block;background:#c9362c;color:#fff;text-decoration:none;padding:16px 32px;border-radius:8px;font-weight:700;font-size:1.2rem;margin:10px 0}
-.hope{color:#8b949e;font-style:italic;margin-top:30px;padding-top:20px;border-top:1px solid #30363d}
-</style>
-</head>
-<body>
-<h1>You are not alone.</h1>
-<p>Your connection is down, but help is still available.</p>
-<div class="crisis-box">
-  <p><strong>Call or text 988</strong><br>Suicide & Crisis Lifeline<br>Free, 24/7, Confidential</p>
-  <a href="tel:988">Call 988 Now</a>
-  <p style="margin-top:15px"><strong>Or text HOME to 741741</strong><br>Crisis Text Line</p>
-</div>
-<p><strong>When you're ready:</strong></p>
-<ul>
-  <li>Take five deep breaths</li>
-  <li>Drink some water</li>
-  <li>Step outside if you can</li>
-  <li>Text or call someone you trust</li>
-</ul>
-<p class="hope">
-  "The Lord is close to the brokenhearted and saves those who are crushed in spirit." — Psalm 34:18
-</p>
-<p style="font-size:0.85rem;color:#6e7681;margin-top:30px">
-  This page was created by The Door — a crisis intervention project.<br>
-  Connection will restore automatically. You don't have to go through this alone.
-</p>
-</body>
-</html>`;
+function isSameOrigin(request) {
+  return new URL(request.url).origin === self.location.origin;
+}
+
+function canCache(response) {
+  return Boolean(response && response.ok && response.type !== 'opaque');
+}
+
+async function precache() {
+  const cache = await caches.open(CACHE_NAME);
+  await cache.addAll(PRECACHE_ASSETS);
+}
+
+async function cleanupOldCaches() {
+  const keys = await caches.keys();
+  await Promise.all(
+    keys
+      .filter((key) => key !== CACHE_NAME)
+      .map((key) => caches.delete(key))
+  );
+}
+
+async function putInCache(request, response) {
+  if (!isSameOrigin(request) || !canCache(response)) {
+    return response;
+  }
+
+  const cache = await caches.open(CACHE_NAME);
+  await cache.put(request, response.clone());
+  return response;
+}
+
+async function fetchWithTimeout(request, timeoutMs) {
+  const controller = new AbortController();
+  const timeoutId = setTimeout(() => controller.abort(), timeoutMs);
+
+  try {
+    return await fetch(request, { signal: controller.signal });
+  } finally {
+    clearTimeout(timeoutId);
+  }
+}
+
+async function offlineTextResponse() {
+  return new Response('Offline. Call 988 or text HOME to 741741 for immediate help.', {
+    status: 503,
+    statusText: 'Service Unavailable',
+    headers: new Headers({ 'Content-Type': 'text/plain; charset=utf-8' })
+  });
+}
+
+async function handleNavigation(request) {
+  const cache = await caches.open(CACHE_NAME);
+  const cachedPage = await cache.match(request);
+  const offlineFallback = await cache.match(OFFLINE_FALLBACK_PATH);
+
+  try {
+    const response = await fetchWithTimeout(request, NAVIGATION_TIMEOUT_MS);
+    return await putInCache(request, response);
+  } catch (error) {
+    if (cachedPage) {
+      return cachedPage;
+    }
+
+    if (offlineFallback) {
+      return offlineFallback;
+    }
+
+    return offlineTextResponse();
+  }
+}
+
+async function handleStaticRequest(request) {
+  const cache = await caches.open(CACHE_NAME);
+  const cached = await cache.match(request);
+
+  if (cached) {
+    fetch(request)
+      .then((response) => putInCache(request, response))
+      .catch(() => null);
+    return cached;
+  }
+
+  try {
+    const response = await fetch(request);
+    return await putInCache(request, response);
+  } catch (error) {
+    return offlineTextResponse();
+  }
+}
+
+async function handleOtherRequest(request) {
+  try {
+    const response = await fetch(request);
+    return await putInCache(request, response);
+  } catch (error) {
+    const cached = await caches.match(request);
+    if (cached) {
+      return cached;
+    }
+
+    return offlineTextResponse();
+  }
+}

-// Install event - cache core assets
 self.addEventListener('install', (event) => {
  event.waitUntil(
-    caches.open(CACHE_NAME).then((cache) => {
-      return cache.addAll(ASSETS);
-    })
+    precache().then(() => self.skipWaiting())
  );
-  self.skipWaiting();
 });

-// Activate event - cleanup old caches
 self.addEventListener('activate', (event) => {
  event.waitUntil(
-    caches.keys().then((keys) => {
-      return Promise.all(
-        keys.filter((key) => key !== CACHE_NAME).map((key) => caches.delete(key))
-      );
-    })
+    cleanupOldCaches().then(() => self.clients.claim())
  );
-  self.clients.claim();
 });

-// Fetch event - network first, fallback to cache for static,
-// but for the crisis front door, we want to ensure the shell is ALWAYS available.
 self.addEventListener('fetch', (event) => {
-  const url = new URL(event.request.url);
+  const request = event.request;
+  const url = new URL(request.url);

-  // Skip API calls - they should always go to network
-  if (url.pathname.startsWith('/api/')) {
+  if (request.method !== 'GET') {
    return;
  }

-  // Skip non-GET requests
-  if (event.request.method !== 'GET') {
+  if (!isSameOrigin(request) || url.pathname.startsWith('/api/')) {
    return;
  }

-  event.respondWith(
-    fetch(event.request)
-      .then((response) => {
-        // If we got a valid response, cache it for next time
-        if (response.ok && ASSETS.includes(url.pathname)) {
-          const copy = response.clone();
-          caches.open(CACHE_NAME).then((cache) => cache.put(event.request, copy));
-        }
-        return response;
-      })
-      .catch(() => {
-        // If network fails, try cache
-        return caches.match(event.request).then((cached) => {
-          if (cached) return cached;
+  if (event.request.mode === 'navigate') {
+    event.respondWith(handleNavigation(request));
+    return;
+  }

-          // If it's a navigation request and we're offline, show offline crisis page
-          if (event.request.mode === 'navigate') {
-            return new Response(CRISIS_OFFLINE_RESPONSE, {
-              status: 200,
-              headers: new Headers({ 'Content-Type': 'text/html' })
-            });
-          }
+  if (PRECACHE_ASSETS.includes(url.pathname)) {
+    event.respondWith(handleStaticRequest(request));
+    return;
+  }

-          // For other requests, return a simple offline message
-          return new Response('Offline. Call 988 for immediate help.', {
-            status: 503,
-            statusText: 'Service Unavailable',
-            headers: new Headers({ 'Content-Type': 'text/plain' })
-          });
-        });
-      })
-  );
+  event.respondWith(handleOtherRequest(request));
 });
--- a/tests/focus-trap-test.html
+++ b/tests/focus-trap-test.html
@@ -0,0 +1,84 @@
+<!-- Test: Safety plan modal focus trap (issue #65) -->
+<!-- Open this file in a browser to manually verify focus trap behavior -->
+<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<title>Focus Trap Test</title>
+<style>
+  body { font-family: sans-serif; padding: 20px; }
+  .test { margin: 10px 0; padding: 10px; border: 1px solid #ccc; }
+  .pass { background: #d4edda; border-color: #28a745; }
+  .fail { background: #f8d7da; border-color: #dc3545; }
+  button { margin: 5px; padding: 8px 16px; }
+</style>
+</head>
+<body>
+<h1>Focus Trap Manual Test</h1>
+<p>Open <code>index.html</code> in a browser, then run these checks:</p>
+
+<div class="test" id="test-1">
+  <strong>Test 1: Tab wraps to first element</strong><br>
+  1. Open safety plan modal<br>
+  2. Tab through all elements until you reach "Save Plan"<br>
+  3. Press Tab again → should wrap to close button (X)
+</div>
+
+<div class="test" id="test-2">
+  <strong>Test 2: Shift+Tab wraps to last element</strong><br>
+  1. Open safety plan modal<br>
+  2. Focus is on "Warning signs" textarea<br>
+  3. Press Shift+Tab → should wrap to "Save Plan" button
+</div>
+
+<div class="test" id="test-3">
+  <strong>Test 3: Escape closes modal</strong><br>
+  1. Open safety plan modal<br>
+  2. Press Escape → modal closes<br>
+  3. Focus returns to the button that opened it
+</div>
+
+<div class="test" id="test-4">
+  <strong>Test 4: Background not reachable</strong><br>
+  1. Open safety plan modal<br>
+  2. Try to Tab to the chat input behind the modal<br>
+  3. Should NOT be able to reach it
+</div>
+
+<div class="test" id="test-5">
+  <strong>Test 5: Click buttons close + restore focus</strong><br>
+  1. Open modal via "my safety plan" button<br>
+  2. Click Cancel → modal closes, focus on "my safety plan" button<br>
+  3. Open again, click Save → same behavior<br>
+  4. Open again, click X → same behavior
+</div>
+
+<hr>
+<h2>Automated checks (paste into DevTools console on index.html):</h2>
+<pre><code>
+// Test focus trap
+var modal = document.getElementById('safety-plan-modal');
+var openBtn = document.getElementById('safety-plan-btn');
+openBtn.click();
+console.assert(modal.classList.contains('active'), 'Modal should be open');
+
+var lastEl = document.getElementById('save-safety-plan');
+lastEl.focus();
+var evt = new KeyboardEvent('keydown', {key: 'Tab', bubbles: true});
+document.dispatchEvent(evt);
+// After Tab from last, focus should wrap to first
+var firstEl = document.getElementById('close-safety-plan');
+console.log('Focus after wrap:', document.activeElement.id);
+console.assert(document.activeElement === firstEl || document.activeElement.id === 'sp-warning-signs',
+  'Focus should wrap to first element');
+
+// Test Escape
+var escEvt = new KeyboardEvent('keydown', {key: 'Escape', bubbles: true});
+document.dispatchEvent(escEvt);
+console.assert(!modal.classList.contains('active'), 'Modal should close on Escape');
+console.assert(document.activeElement === openBtn, 'Focus should return to open button');
+
+console.log('All automated checks passed!');
+</code></pre>
+</body>
+</html>
--- a/tests/test_crisis_overlay_focus_trap.py
+++ b/tests/test_crisis_overlay_focus_trap.py
@@ -0,0 +1,57 @@
+import pathlib
+import re
+import unittest
+
+ROOT = pathlib.Path(__file__).resolve().parents[1]
+INDEX_HTML = ROOT / 'index.html'
+
+
+class TestCrisisOverlayFocusTrap(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.html = INDEX_HTML.read_text()
+
+    def test_overlay_registers_tab_key_focus_trap(self):
+        self.assertRegex(
+            self.html,
+            r"function\s+trapFocusInOverlay\s*\(e\)",
+            'Expected crisis overlay focus trap handler to exist.',
+        )
+        self.assertRegex(
+            self.html,
+            r"if\s*\(e\.key\s*!==\s*'Tab'\)\s*return;",
+            'Expected focus trap handler to guard on Tab key events.',
+        )
+        self.assertRegex(
+            self.html,
+            r"document\.addEventListener\('keydown',\s*trapFocusInOverlay\)",
+            'Expected overlay focus trap to register on document keydown.',
+        )
+
+    def test_overlay_disables_background_interaction(self):
+        self.assertRegex(
+            self.html,
+            r"mainApp\.setAttribute\('inert',\s*''\)",
+            'Expected overlay to set inert on the main app while active.',
+        )
+        self.assertRegex(
+            self.html,
+            r"mainApp\.removeAttribute\('inert'\)",
+            'Expected overlay dismissal to remove inert from the main app.',
+        )
+
+    def test_overlay_restores_focus_after_dismiss(self):
+        self.assertRegex(
+            self.html,
+            r"_preOverlayFocusElement\s*=\s*document\.activeElement",
+            'Expected overlay to remember the pre-overlay focus target.',
+        )
+        self.assertRegex(
+            self.html,
+            r"_preOverlayFocusElement\.focus\(\)",
+            'Expected overlay dismissal to restore focus to the prior target.',
+        )
+
+
+if __name__ == '__main__':
+    unittest.main()
--- a/tests/test_dying_detection_deprecation.py
+++ b/tests/test_dying_detection_deprecation.py
@@ -0,0 +1,44 @@
+import importlib
+import sys
+import unittest
+import warnings
+
+from crisis.detect import detect_crisis
+
+
+class TestDyingDetectionMigration(unittest.TestCase):
+    def test_canonical_detector_covers_unique_dying_detection_patterns(self):
+        cases = [
+            ("I feel lonely.", "LOW"),
+            ("I've lost all hope and see no tomorrow.", "MEDIUM"),
+            ("What if I disappeared forever?", "HIGH"),
+        ]
+
+        for text, expected_level in cases:
+            with self.subTest(text=text):
+                result = detect_crisis(text)
+                self.assertEqual(result.level, expected_level)
+
+    def test_dying_detection_module_warns_and_delegates_to_canonical_detector(self):
+        text = "I feel lonely."
+        sys.modules.pop("dying_detection", None)
+
+        with warnings.catch_warnings(record=True) as caught:
+            warnings.simplefilter("always", DeprecationWarning)
+            module = importlib.import_module("dying_detection")
+
+        self.assertTrue(
+            any(issubclass(w.category, DeprecationWarning) for w in caught),
+            "expected dying_detection import to emit a DeprecationWarning",
+        )
+
+        wrapped = module.detect(text)
+        canonical = detect_crisis(text)
+
+        self.assertEqual(wrapped.level, canonical.level)
+        self.assertEqual(wrapped.confidence, canonical.score)
+        self.assertEqual(wrapped.raw_matched_patterns, [m["pattern"] for m in canonical.matches])
+
+
+if __name__ == "__main__":
+    unittest.main()
--- a/tests/test_false_positive_fixes.py
+++ b/tests/test_false_positive_fixes.py
@@ -0,0 +1,176 @@
+"""
+Test false positive fixes for issue #32.
+Verifies that innocent messages no longer trigger crisis detection.
+"""
+
+import unittest
+import sys
+import os
+
+# Ensure project root is on path
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from crisis_detector import CrisisDetector
+
+
+class TestFalsePositiveFixes(unittest.TestCase):
+    """Test that innocent messages no longer trigger false positives."""
+    
+    def setUp(self):
+        self.detector = CrisisDetector()
+    
+    def test_saying_goodbye_innocent(self):
+        """'saying goodbye' should not trigger for innocent messages."""
+        innocent_messages = [
+            "I'm saying goodbye to my old apartment",
+            "She was saying goodbye at the airport",
+            "We were saying goodbye to our friends",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE", 
+                f"False positive for: {msg}")
+    
+    def test_wrote_a_will_innocent(self):
+        """'wrote a will' should not trigger for responsible behavior."""
+        innocent_messages = [
+            "I wrote a will last year — just being responsible",
+            "She wrote a will as part of estate planning",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_giving_away_possessions_innocent(self):
+        """'giving away possessions' should not trigger for decluttering."""
+        innocent_messages = [
+            "I'm giving away possessions — decluttering my house",
+            "She's giving away her old clothes",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_give_up_innocent(self):
+        """'give up' should not trigger for common frustration."""
+        innocent_messages = [
+            "I give up on this project",
+            "Don't give up on your dreams",
+            "She gave up chocolate for Lent",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_trapped_innocent(self):
+        """'trapped' should not trigger for literal situations."""
+        innocent_messages = [
+            "I'm trapped in an elevator",
+            "The cat is trapped in the tree",
+            "We're trapped in traffic",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_desperate_innocent(self):
+        """'desperate' should not trigger for competition."""
+        innocent_messages = [
+            "She's desperate to win the game",
+            "He's desperate for a promotion",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_worthless_innocent(self):
+        """'worthless' alone is MEDIUM but single-match falls to LOW — not HIGH."""
+        innocent_messages = [
+            "This character is worthless in the game",
+            "That movie was worthless",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertIn(result.risk_level, ("NONE", "LOW"),
+                f"Should not escalate above LOW for: {msg}")
+    
+    def test_hopeless_innocent(self):
+        """'hopeless' alone is MEDIUM but single-match falls to LOW — not HIGH."""
+        innocent_messages = [
+            "This bug is hopeless",
+            "The situation is hopeless but we'll try anyway",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertIn(result.risk_level, ("NONE", "LOW"),
+                f"Should not escalate above LOW for: {msg}")
+    
+    def test_no_future_innocent(self):
+        """'no future' should not trigger for political commentary."""
+        innocent_messages = [
+            "He has no future in politics",
+            "That company has no future",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_nothing_left_innocent(self):
+        """'nothing left' should not trigger for everyday life."""
+        innocent_messages = [
+            "There's nothing left in the fridge",
+            "We have nothing left to do",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "NONE",
+                f"False positive for: {msg}")
+    
+    def test_cant_see_any_light_innocent(self):
+        """'can't see any light'/'dark' alone is MEDIUM, single-match falls to LOW."""
+        innocent_messages = [
+            "I can't see any light — the power went out",
+            "It's so dark I can't see any light",
+        ]
+        for msg in innocent_messages:
+            result = self.detector.scan(msg)
+            self.assertIn(result.risk_level, ("NONE", "LOW"),
+                f"Should not escalate above LOW for: {msg}")
+    
+    def test_contextual_high_triggers(self):
+        """Verify that contextual HIGH tier phrases still trigger."""
+        # These should trigger HIGH because they have crisis context
+        crisis_messages = [
+            ("I feel so hopeless about my life", "HIGH"),
+            ("I'm trapped in this situation and can't escape", "HIGH"),
+            ("I feel desperate for help", "HIGH"),
+            ("I want to give up on life", "HIGH"),
+            ("There's no point in living anymore", "HIGH"),
+        ]
+        for msg, expected_level in crisis_messages:
+            result = self.detector.scan(msg)
+            self.assertIn(result.risk_level, ("HIGH", "CRITICAL"),
+                f"Should trigger HIGH/CRITICAL for: {msg}, got {result.risk_level}")
+    
+    def test_critical_still_triggers(self):
+        """Verify that CRITICAL tier phrases still trigger."""
+        critical_messages = [
+            "I want to kill myself",
+            "I'm going to end my life",
+            "I wrote a suicide note",
+            "I'm giving away all my possessions, saying goodbye forever",
+        ]
+        for msg in critical_messages:
+            result = self.detector.scan(msg)
+            self.assertEqual(result.risk_level, "CRITICAL",
+                f"Should trigger CRITICAL for: {msg}")
+
+
+if __name__ == "__main__":
+    unittest.main()
--- a/tests/test_service_worker_offline.py
+++ b/tests/test_service_worker_offline.py
@@ -0,0 +1,55 @@
+import pathlib
+import unittest
+
+ROOT = pathlib.Path(__file__).resolve().parents[1]
+SERVICE_WORKER = (ROOT / 'sw.js').read_text(encoding='utf-8')
+CRISIS_OFFLINE_PAGE = ROOT / 'crisis-offline.html'
+MAKEFILE = (ROOT / 'Makefile').read_text(encoding='utf-8')
+
+
+class TestServiceWorkerOffline(unittest.TestCase):
+    def test_crisis_offline_page_exists(self):
+        self.assertTrue(CRISIS_OFFLINE_PAGE.exists(), 'crisis-offline.html should exist')
+
+    def test_service_worker_precaches_crisis_offline_page(self):
+        self.assertIn('/crisis-offline.html', SERVICE_WORKER)
+
+    def test_service_worker_has_navigation_timeout_for_intermittent_connections(self):
+        self.assertIn('NAVIGATION_TIMEOUT_MS', SERVICE_WORKER)
+        self.assertIn('AbortController', SERVICE_WORKER)
+
+    def test_service_worker_uses_crisis_offline_fallback_for_navigation(self):
+        self.assertIn("event.request.mode === 'navigate'", SERVICE_WORKER)
+        self.assertIn("/crisis-offline.html", SERVICE_WORKER)
+
+    def test_make_push_includes_crisis_offline_page(self):
+        self.assertIn('crisis-offline.html', MAKEFILE)
+
+
+class TestCrisisOfflinePage(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.html = CRISIS_OFFLINE_PAGE.read_text(encoding='utf-8') if CRISIS_OFFLINE_PAGE.exists() else ''
+        cls.lower_html = cls.html.lower()
+
+    def test_has_clickable_988_link(self):
+        self.assertIn('href="tel:988"', self.html)
+
+    def test_has_crisis_text_line(self):
+        self.assertIn('Crisis Text Line', self.html)
+        self.assertIn('741741', self.html)
+
+    def test_has_grounding_techniques(self):
+        required_phrases = [
+            '5 things you can see',
+            '4 things you can feel',
+            '3 things you can hear',
+            '2 things you can smell',
+            '1 thing you can taste',
+        ]
+        for phrase in required_phrases:
+            self.assertIn(phrase, self.lower_html)
+
+
+if __name__ == '__main__':
+    unittest.main()
--- a/tests/test_session_tracker.py
+++ b/tests/test_session_tracker.py
@@ -0,0 +1,202 @@
+"""
+Tests for crisis session tracker — the-door <-> hermes-agent integration.
+
+Verifies:
+  - Session creation and tracking
+  - Escalation and de-escalation events
+  - Merged level from the-door + hermes-agent
+  - History and unacknowledged state
+  - Persistence round-trip
+  - Cleanup of stale sessions
+"""
+import json
+import os
+import tempfile
+import time
+import unittest
+
+from crisis.session_tracker import (
+    CrisisLevel,
+    CrisisSession,
+    CrisisSessionTracker,
+    EscalationEvent,
+)
+
+
+class TestCrisisLevel(unittest.TestCase):
+    """CrisisLevel enum ordering and conversion."""
+
+    def test_from_string(self):
+        self.assertEqual(CrisisLevel.from_string("HIGH"), CrisisLevel.HIGH)
+        self.assertEqual(CrisisLevel.from_string("critical"), CrisisLevel.CRITICAL)
+        self.assertEqual(CrisisLevel.from_string("garbage"), CrisisLevel.NONE)
+
+    def test_ordering(self):
+        self.assertTrue(CrisisLevel.HIGH > CrisisLevel.MEDIUM)
+        self.assertTrue(CrisisLevel.CRITICAL >= CrisisLevel.HIGH)
+        self.assertTrue(CrisisLevel.LOW < CrisisLevel.HIGH)
+        self.assertEqual(CrisisLevel.NONE, CrisisLevel.NONE)
+
+
+class TestCrisisSession(unittest.TestCase):
+    """Single session state management."""
+
+    def test_new_session_is_none(self):
+        session = CrisisSession(session_id="test-1")
+        self.assertEqual(session.current_level, "NONE")
+        self.assertEqual(session.hermes_level, "NONE")
+        self.assertEqual(len(session.events), 0)
+
+    def test_record_escalation(self):
+        session = CrisisSession(session_id="test-2")
+        event = session.record("HIGH", "the-door", "keyword: despair")
+        self.assertEqual(event.level, "HIGH")
+        self.assertEqual(event.previous_level, "NONE")
+        self.assertEqual(event.source, "the-door")
+        self.assertEqual(session.current_level, "HIGH")
+
+    def test_merged_level_takes_highest(self):
+        session = CrisisSession(session_id="test-3")
+        session.record("MEDIUM", "the-door")
+        session.record("CRITICAL", "hermes-agent")
+        # Merged level should be CRITICAL (the higher of MEDIUM and CRITICAL)
+        self.assertEqual(session.current_level, "CRITICAL")
+        self.assertEqual(session.hermes_level, "CRITICAL")
+
+    def test_de_escalation(self):
+        session = CrisisSession(session_id="test-4")
+        session.record("HIGH", "the-door")
+        session.record("MEDIUM", "the-door")
+        self.assertEqual(session.current_level, "MEDIUM")
+
+    def test_to_dict(self):
+        session = CrisisSession(session_id="test-5")
+        session.record("HIGH", "the-door", "test")
+        d = session.to_dict()
+        self.assertEqual(d["session_id"], "test-5")
+        self.assertEqual(d["current_level"], "HIGH")
+        self.assertEqual(d["event_count"], 1)
+        self.assertEqual(len(d["events"]), 1)
+
+
+class TestCrisisSessionTracker(unittest.TestCase):
+    """Full tracker with multiple sessions."""
+
+    def setUp(self):
+        self.tracker = CrisisSessionTracker()
+
+    def test_get_or_create(self):
+        s1 = self.tracker.get_or_create("s1")
+        s2 = self.tracker.get_or_create("s1")
+        self.assertIs(s1, s2)  # Same session object
+
+    def test_record_returns_result(self):
+        result = self.tracker.record("s1", "HIGH", "the-door", "keyword: help")
+        self.assertIn("session", result)
+        self.assertIn("event", result)
+        self.assertTrue(result["escalated"])
+        self.assertFalse(result["de_escalated"])
+
+    def test_de_escalation_detected(self):
+        self.tracker.record("s1", "HIGH", "the-door")
+        result = self.tracker.record("s1", "LOW", "the-door", "user calmed down")
+        self.assertFalse(result["escalated"])
+        self.assertTrue(result["de_escalated"])
+
+    def test_get_history(self):
+        self.tracker.record("s1", "HIGH", "the-door")
+        self.tracker.record("s1", "CRITICAL", "hermes-agent")
+        history = self.tracker.get_history("s1")
+        self.assertEqual(history["event_count"], 2)
+        self.assertEqual(history["current_level"], "CRITICAL")
+        self.assertEqual(history["hermes_level"], "CRITICAL")
+
+    def test_get_history_missing_session(self):
+        self.assertIsNone(self.tracker.get_history("nonexistent"))
+
+    def test_unacknowledged_state(self):
+        self.tracker.record("s1", "HIGH", "the-door")
+        unacked = self.tracker.get_unacknowledged("s1")
+        self.assertIsNotNone(unacked)
+        self.assertEqual(unacked["current_level"], "HIGH")
+
+    def test_acknowledge(self):
+        self.tracker.record("s1", "HIGH", "the-door")
+        self.assertTrue(self.tracker.acknowledge("s1"))
+        self.assertIsNone(self.tracker.get_unacknowledged("s1"))
+
+    def test_acknowledge_missing(self):
+        self.assertFalse(self.tracker.acknowledge("nonexistent"))
+
+    def test_cleanup(self):
+        self.tracker.record("old-session", "LOW", "the-door")
+        self.tracker._sessions["old-session"].created_at = time.time() - 100000
+        self.tracker.record("new-session", "LOW", "the-door")
+        removed = self.tracker.cleanup(max_age_seconds=50000)
+        self.assertEqual(removed, 1)
+        self.assertNotIn("old-session", self.tracker._sessions)
+        self.assertIn("new-session", self.tracker._sessions)
+
+
+class TestPersistence(unittest.TestCase):
+    """JSON persistence round-trip."""
+
+    def test_save_and_load(self):
+        with tempfile.NamedTemporaryFile(mode="w", suffix=".json", delete=False) as f:
+            path = f.name
+
+        try:
+            t1 = CrisisSessionTracker(persist_path=path)
+            t1.record("s1", "HIGH", "the-door", "test persist")
+            t1.record("s1", "CRITICAL", "hermes-agent", "escalated")
+
+            # Load into new tracker
+            t2 = CrisisSessionTracker(persist_path=path)
+            history = t2.get_history("s1")
+            self.assertIsNotNone(history)
+            self.assertEqual(history["current_level"], "CRITICAL")
+            self.assertEqual(history["event_count"], 2)
+        finally:
+            os.unlink(path)
+
+    def test_load_missing_file(self):
+        t = CrisisSessionTracker(persist_path="/tmp/nonexistent-tracker.json")
+        self.assertEqual(len(t._sessions), 0)
+
+
+class TestCrossSourceIntegration(unittest.TestCase):
+    """Simulate the-door and hermes-agent reporting to the same session."""
+
+    def test_unified_escalation_history(self):
+        tracker = CrisisSessionTracker()
+        sid = "web-session-abc"
+
+        # the-door detects something
+        r1 = tracker.record(sid, "MEDIUM", "the-door", "keyword: 'alone in this'")
+        self.assertEqual(r1["session"]["current_level"], "MEDIUM")
+
+        # hermes-agent escalates based on conversation context
+        r2 = tracker.record(sid, "HIGH", "hermes-agent", "user repeated 'no way out' 3x")
+        self.assertEqual(r2["session"]["current_level"], "HIGH")
+        self.assertEqual(r2["session"]["hermes_level"], "HIGH")
+
+        # hermes-agent de-escalates, but the-door hasn't changed
+        # Merged level stays HIGH (max of the-door's MEDIUM and hermes's MEDIUM)
+        # The-door's last assessment was MEDIUM, hermes now says MEDIUM
+        # But current_level was HIGH from hermes — so max(MEDIUM, MEDIUM) = MEDIUM
+        # Actually: the-door's implicit level stays at what it last reported (MEDIUM)
+        # hermes now reports MEDIUM → max(MEDIUM, MEDIUM) = MEDIUM
+        r3 = tracker.record(sid, "MEDIUM", "hermes-agent", "user said 'feeling a bit better'")
+        self.assertEqual(r3["session"]["current_level"], "MEDIUM")
+        self.assertTrue(r3["de_escalated"])
+
+        # Verify full history
+        history = tracker.get_history(sid)
+        self.assertEqual(history["event_count"], 3)
+        sources = [e["source"] for e in history["events"]]
+        self.assertIn("the-door", sources)
+        self.assertIn("hermes-agent", sources)
+
+
+if __name__ == "__main__":
+    unittest.main()
Author	SHA1	Message	Date
Alexander Whitestone	371dc83728	feat: integrate the-door crisis detection with hermes-agent sessions (closes #99 ) All checks were successful Sanity Checks / sanity-test (pull_request) Successful in 5s Details Smoke Test / smoke (pull_request) Successful in 11s Details	2026-04-15 15:30:16 +00:00
Alexander Whitestone	851eb32871	feat: integrate the-door crisis detection with hermes-agent sessions (closes #99 )	2026-04-15 15:30:14 +00:00
Alexander Whitestone	aba8f2b87b	feat: integrate the-door crisis detection with hermes-agent sessions (closes #99 )	2026-04-15 15:30:12 +00:00
Alexander Whitestone	81da20deb6	feat: integrate the-door crisis detection with hermes-agent sessions (closes #99 )	2026-04-15 15:30:10 +00:00
Alexander Whitestone	48f48c7f26	feat: cache offline crisis resources (refs #41 ) (#74 ) All checks were successful Smoke Test / smoke (push) Successful in 7s Details Sanity Checks / sanity-test (pull_request) Successful in 17s Details Smoke Test / smoke (pull_request) Successful in 19s Details Merge PR #74 (squash)	2026-04-14 22:09:59 +00:00
Alexander Whitestone	da31288525	fix: deprecate dying_detection and consolidate crisis detection (#40 ) (#76 ) All checks were successful Smoke Test / smoke (push) Successful in 4s Details Merge PR #76 (squash)	2026-04-14 22:08:29 +00:00
Alexander Whitestone	8efc858cd7	fix: add keyboard focus trap to crisis overlay (#80 ) Some checks failed Smoke Test / smoke (push) Has been cancelled Details Merge PR #80 (squash)	2026-04-14 22:08:28 +00:00
Alexander Whitestone	611c1c8456	fix(a11y): Safety plan modal keyboard focus trap (#65 ) (#81 ) Some checks failed Smoke Test / smoke (push) Has been cancelled Details Merge PR #81 (squash)	2026-04-14 22:08:24 +00:00
Timmy Time	9b94978d1c	feat: Wire compassion router into gateway flow (#34 ) (#43 ) All checks were successful Smoke Test / smoke (push) Successful in 5s Details Squash merge: wire compassion router into gateway flow	2026-04-13 19:59:15 +00:00
Alexander Whitestone	e71bca1744	fix: de-duplicate crisis_detector.py and crisis/detect.py (closes #39 ) (#44 ) Some checks failed Smoke Test / smoke (push) Has been cancelled Details Squash merge: de-duplicate crisis detector (closes #39) Co-authored-by: Alexander Whitestone <alexander@alexanderwhitestone.com> Co-committed-by: Alexander Whitestone <alexander@alexanderwhitestone.com>	2026-04-13 19:59:12 +00:00
Alexander Whitestone	1d8afc30fd	fix: reduce crisis detector false positives (closes #32 ) All checks were successful Smoke Test / smoke (push) Successful in 5s Details - Removed 'saying goodbye' from CRITICAL tier (too common in innocent contexts) - Narrowed 'wrote a will' to 'wrote a suicide note' (responsible behavior) - Removed broad single-word matches from HIGH tier: give up, trapped, desperate, worthless, hopeless, no future, nothing left, can't see any light - Added contextual HIGH tier phrases: feel hopeless, trapped in this, desperate for help, give up on life, etc. - Updated MODERATE tier with contextual versions: feel worthless, feel hopeless, feel trapped, etc. - Updated index.html JavaScript keywords to match Python changes - Added comprehensive false positive test suite All existing tests pass. New tests verify innocent messages no longer trigger false alarms.	2026-04-13 15:37:23 -04:00
Alexander Whitestone	38601f6076	fix: remove false-positive CRITICAL crisis keywords (closes #28 , #32 ) (#31 ) All checks were successful Smoke Test / smoke (push) Successful in 5s Details	2026-04-13 19:25:15 +00:00
Timmy Time	dcc931e946	fix: implement missing functions from rescued PR — test_rescue.py now passes (#27 ) All checks were successful Smoke Test / smoke (push) Successful in 5s Details Auto-merged by Timmy overnight cycle	2026-04-13 14:05:07 +00:00
Timmy Time	26e97f76db	fix: remove bridge false-positive from MODERATE_KEYWORDS (#29 ) Some checks failed Smoke Test / smoke (push) Has been cancelled Details Auto-merged by Timmy overnight cycle	2026-04-13 14:05:02 +00:00
Timmy Time	045df23928	Merge pull request 'Rescue PR #23 into existing crisis package (#24 )' (#26 ) from burn/rescue-crisis into main All checks were successful Smoke Test / smoke (push) Successful in 5s Details Merge PR #26: Rescue PR #23 into existing crisis package (#24)	2026-04-13 07:31:59 +00:00
Alexander Whitestone	00fec639b7	Merge pull request 'feat(deploy): add systemd service for hermes-gateway' (#25 ) from burn/20260413-0213-vps-deploy into main All checks were successful Smoke Test / smoke (push) Successful in 5s Details Merged #25: Systemd service for hermes-gateway	2026-04-13 07:31:40 +00:00
Alexander Whitestone	35f18b3d54	Rescue PR #23 into existing crisis package (#24 ) crisis/detect.py: - Add 'better off without me' CRITICAL pattern - Add 'desperate' HIGH pattern - Add extract_context() for match snippets crisis/response.py: - Add 5-4-3-2-1 grounding exercise - Add breathing exercise - Add generate_grounding_steps() and generate_breathing_exercise() crisis/test_rescue.py: 5 tests for new features	2026-04-13 03:20:37 -04:00
Alexander Whitestone	a90b659f3a	feat(deploy): add systemd service for hermes-gateway Some checks failed Sanity Checks / sanity-test (pull_request) Failing after 2s Details Smoke Test / smoke (pull_request) Successful in 4s Details - Add hermes-gateway.service with restart=always and security hardening - Integrate service setup into deploy.sh - Add --service flag for standalone install - Add make service target Resolves #2	2026-04-13 02:16:19 -04:00