Add multi-factor conviction gate to activation

Make "qualified" mean an edge candidate, not just R:R + confidence.
The gate now also requires (all admin-configurable, defaults on):
- high conviction: recommended_action LONG_HIGH / SHORT_HIGH only
- clean read: risk_level Low (no contradicting signals)
- probable primary target: best target probability >= min (default 60)

- Shared predicate: app/services/qualification.py +
  frontend/src/lib/qualification.ts (mirrored)
- Activation config extended (min_target_probability,
  require_high_conviction, exclude_conflicts) with bool-aware
  get/update + validation
- /trades/performance switched to ?qualified_only=true, applying
  the full gate server-side; confidence breakdown stays unfiltered
- Dashboard "Qualified", Signals "Qualified only" toggle, and
  Track Record all use the one gate; Admin gains the new controls

Sentiment provider runtime config (prior change) included.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

frontend/src/components/admin/ActivationSettings.tsx

@@ -6,6 +6,9 @@ import { SkeletonTable } from '../ui/Skeleton';
 const DEFAULTS: ActivationConfig = {
   min_rr: 2,
   min_confidence: 70,
+  min_target_probability: 60,
+  require_high_conviction: true,
+  exclude_conflicts: true,
 };
 
 export function ActivationSettings() {
@@ -19,12 +22,12 @@ export function ActivationSettings() {
   }, [data]);
 
   const onSave = () => {
-    update.mutate(form as unknown as Record<string, number>);
+    update.mutate(form);
   };
 
   const onReset = () => {
     setForm(DEFAULTS);
-    update.mutate(DEFAULTS as unknown as Record<string, number>);
+    update.mutate(DEFAULTS);
   };
 
   if (isLoading) return <SkeletonTable rows={2} cols={2} />;
@@ -33,16 +36,16 @@ export function ActivationSettings() {
   return (
     <div className="glass p-5 space-y-4">
       <div>
-        <h3 className="text-sm font-semibold text-gray-200">Activation Thresholds</h3>
+        <h3 className="text-sm font-semibold text-gray-200">Activation Gate</h3>
         <p className="mt-1 text-xs text-gray-500">
-          What counts as a signal worth acting on. Used as the default Signals filters, the
-          Dashboard's qualified-setup metrics, and the Track Record's "qualified only" view.
-          All setups are still evaluated regardless, so these thresholds can be validated
-          against the confidence breakdown.
+          What counts as a signal worth acting on. Drives the Dashboard's "Qualified" metric, the
+          Signals "Qualified only" view, and the Track Record's qualified stats. All setups are
+          still evaluated regardless — tighten the gate, then watch qualified expectancy in the
+          Track Record to find what actually wins.
         </p>
       </div>
 
-      <div className="grid gap-4 md:grid-cols-2">
+      <div className="grid gap-4 md:grid-cols-3">
         <label className="block space-y-1">
           <span className="text-xs text-gray-400">Min Risk:Reward (1 : x)</span>
           <input
@@ -53,6 +56,7 @@ export function ActivationSettings() {
             onChange={(e) => setForm((prev) => ({ ...prev, min_rr: Number(e.target.value) }))}
             className="w-full input-glass px-3 py-2 text-sm"
           />
+          <span className="text-[11px] text-gray-600">Set above your scanner floor or it does nothing.</span>
         </label>
         <label className="block space-y-1">
           <span className="text-xs text-gray-400">Min Confidence (%)</span>
@@ -66,6 +70,50 @@ export function ActivationSettings() {
             className="w-full input-glass px-3 py-2 text-sm"
           />
         </label>
+        <label className="block space-y-1">
+          <span className="text-xs text-gray-400">Min Target Probability (%)</span>
+          <input
+            type="number"
+            min={0}
+            max={100}
+            step={1}
+            value={form.min_target_probability}
+            onChange={(e) => setForm((prev) => ({ ...prev, min_target_probability: Number(e.target.value) }))}
+            className="w-full input-glass px-3 py-2 text-sm"
+          />
+          <span className="text-[11px] text-gray-600">Best target's probability must clear this. 0 disables.</span>
+        </label>
+      </div>
+
+      <div className="grid gap-3 md:grid-cols-2">
+        <label className="flex cursor-pointer items-start gap-2.5 text-sm text-gray-300">
+          <input
+            type="checkbox"
+            checked={form.require_high_conviction}
+            onChange={(e) => setForm((prev) => ({ ...prev, require_high_conviction: e.target.checked }))}
+            className="mt-0.5 h-4 w-4 cursor-pointer accent-blue-400"
+          />
+          <span>
+            Require high conviction
+            <span className="mt-0.5 block text-[11px] text-gray-500">
+              Only LONG (High) / SHORT (High) — the signals must clearly pick a side.
+            </span>
+          </span>
+        </label>
+        <label className="flex cursor-pointer items-start gap-2.5 text-sm text-gray-300">
+          <input
+            type="checkbox"
+            checked={form.exclude_conflicts}
+            onChange={(e) => setForm((prev) => ({ ...prev, exclude_conflicts: e.target.checked }))}
+            className="mt-0.5 h-4 w-4 cursor-pointer accent-blue-400"
+          />
+          <span>
+            Exclude conflicted setups
+            <span className="mt-0.5 block text-[11px] text-gray-500">
+              Risk level must be Low — drops setups with contradicting signals.
+            </span>
+          </span>
+        </label>
       </div>
 
       <div className="flex items-center gap-2">

frontend/src/components/admin/SentimentProviderSettings.tsx

@@ -0,0 +1,154 @@
+import { useEffect, useState } from 'react';
+import {
+  useSentimentSettings,
+  useUpdateSentimentSettings,
+  useTestSentimentProvider,
+} from '../../hooks/useAdmin';
+import { Select } from '../ui/Field';
+import { SkeletonTable } from '../ui/Skeleton';
+import type { SentimentTestResult } from '../../lib/types';
+
+const SOURCE_LABEL: Record<string, string> = {
+  database: 'configured here',
+  environment: 'from environment (.env)',
+  none: 'not configured',
+};
+
+export function SentimentProviderSettings() {
+  const { data, isLoading, isError, error } = useSentimentSettings();
+  const update = useUpdateSentimentSettings();
+  const test = useTestSentimentProvider();
+
+  const [provider, setProvider] = useState('openai');
+  const [model, setModel] = useState('');
+  const [apiKey, setApiKey] = useState('');
+  const [testResult, setTestResult] = useState<SentimentTestResult | null>(null);
+
+  useEffect(() => {
+    if (data) {
+      setProvider(data.provider);
+      setModel(data.model);
+    }
+  }, [data]);
+
+  if (isLoading) return <SkeletonTable rows={3} cols={2} />;
+  if (isError) return <p className="text-sm text-red-400">{(error as Error)?.message || 'Failed to load sentiment settings'}</p>;
+  if (!data) return null;
+
+  const onProviderChange = (next: string) => {
+    setProvider(next);
+    // Auto-fill the model with the new provider's default unless the user has a
+    // custom value that isn't the previous provider's default.
+    const defaults = data.default_models;
+    if (!model || Object.values(defaults).includes(model)) {
+      setModel(defaults[next] ?? '');
+    }
+  };
+
+  const onSave = () => {
+    setTestResult(null);
+    update.mutate({
+      provider,
+      model,
+      ...(apiKey ? { api_key: apiKey } : {}),
+    });
+    setApiKey('');
+  };
+
+  const onTest = () => {
+    test.mutate('AAPL', { onSuccess: (res) => setTestResult(res) });
+  };
+
+  const keyConfigured = data.api_key_configured;
+
+  return (
+    <div className="glass p-5 space-y-4">
+      <div>
+        <h3 className="text-sm font-semibold text-gray-200">Sentiment LLM Provider</h3>
+        <p className="mt-1 text-xs text-gray-500">
+          Switch the model that powers sentiment analysis without redeploying. Applies to the
+          scheduled sentiment job and manual fetches on the next run.
+        </p>
+      </div>
+
+      <div className="grid gap-4 md:grid-cols-2">
+        <label className="block space-y-1">
+          <span className="text-xs text-gray-400">Provider</span>
+          <Select value={provider} onChange={(e) => onProviderChange(e.target.value)} className="w-full !py-2">
+            {data.valid_providers.map((p) => (
+              <option key={p} value={p}>{p}</option>
+            ))}
+          </Select>
+        </label>
+
+        <label className="block space-y-1">
+          <span className="text-xs text-gray-400">Model</span>
+          <input
+            type="text"
+            value={model}
+            onChange={(e) => setModel(e.target.value)}
+            placeholder={data.default_models[provider] ?? ''}
+            className="w-full input-glass px-3 py-2 text-sm"
+          />
+        </label>
+      </div>
+
+      <label className="block space-y-1">
+        <span className="text-xs text-gray-400">API Key</span>
+        <input
+          type="password"
+          value={apiKey}
+          onChange={(e) => setApiKey(e.target.value)}
+          autoComplete="new-password"
+          placeholder={keyConfigured ? '•••••••••• (leave blank to keep current)' : 'Paste API key…'}
+          className="w-full input-glass px-3 py-2 text-sm"
+        />
+        <span className="text-[11px] text-gray-500">
+          Key status:{' '}
+          <span className={keyConfigured ? 'text-emerald-400' : 'text-amber-400'}>
+            {SOURCE_LABEL[data.api_key_source] ?? data.api_key_source}
+          </span>
+          {' '}· write-only, never displayed
+        </span>
+      </label>
+
+      <div className="flex flex-wrap items-center gap-2">
+        <button className="btn-primary px-4 py-2 text-sm" onClick={onSave} disabled={update.isPending}>
+          {update.isPending ? 'Saving…' : 'Save Provider'}
+        </button>
+        <button
+          className="px-4 py-2 text-sm rounded border border-white/[0.1] text-gray-300 hover:text-white disabled:opacity-50"
+          onClick={onTest}
+          disabled={test.isPending}
+        >
+          {test.isPending ? 'Testing (AAPL)…' : 'Test Connection'}
+        </button>
+        <span className="text-[11px] text-gray-500">Test fetches live sentiment for AAPL with the saved config.</span>
+      </div>
+
+      {testResult && (
+        <div
+          className={`rounded-lg border px-4 py-3 text-sm ${
+            testResult.ok
+              ? 'border-emerald-500/20 bg-emerald-500/10 text-emerald-300'
+              : 'border-red-500/20 bg-red-500/10 text-red-400'
+          }`}
+        >
+          {testResult.ok ? (
+            <>
+              <span className="font-medium">✓ {testResult.provider} / {testResult.model}</span> — {testResult.ticker}:{' '}
+              <span className="font-semibold">{testResult.classification}</span>{' '}
+              <span className="num">({testResult.confidence}%)</span>
+              {testResult.reasoning && <p className="mt-1 text-xs text-emerald-300/80">{testResult.reasoning}</p>}
+            </>
+          ) : (
+            <>
+              <span className="font-medium">✗ Test failed</span>
+              <p className="mt-1 text-xs">{testResult.error}</p>
+            </>
+          )}
+        </div>
+      )}
+    </div>
+  );
+}