diff --git a/check_abbreviations.py b/check_abbreviations.py
new file mode 100644
index 0000000..ae60584
--- /dev/null
+++ b/check_abbreviations.py
@@ -0,0 +1,217 @@
+#!/usr/bin/env python3
+"""check_abbreviations.py — Prüft Artikel-HTML auf undefinierte Abkürzungen.
+Exit 0 = alle definiert, Exit 1 = undefinierte gefunden, Exit 2 = Fehler.
+Usage: python3 check_abbreviations.py <artikel.html>
+"""
+import re, sys
+from html import unescape
+from collections import Counter
+
+SKIP_WORDS = {
+    'UND','ODER','DIE','DER','DAS','IST','VON','MIT','AUF','FÜR','DEN','DEM',
+    'DES','EIN','EINE','EINEN','EINER','NACH','ZUM','ZUR','BEI','IM','AM','UM',
+    'BIS','ZWEI','DREI','VIER','FÜNF','DANN','NOCH','AUCH','NUN','WIE','WAS',
+    'WER','WAR','HAT','HABEN','NICHT','KEIN','ALLE','ALLEN','ALLER','OHNE',
+    'ABER','ODER','SEIT','SCHON','MEHR','NUR','ALS','WIRD','WURDE','WORDEN',
+    'SIND','SEIN','HIER','DA','DORT','JETZT','HEUTE','MORGEN','GESTERN',
+    'IHR','IHRE','IHREN','IHM','IHN','IHRER','SIE','WIR','UNS','EUCH','EUER',
+    'MIR','MICH','MEIN','MEINE','DEIN','DEINE','SEINE','SEINER',
+    'SPIEL','SPIELE','JAHR','JAHRE','JULI','JUNI','MAI','MÄRZ','APRIL',
+    'JANUAR','FEBRUAR','AUGUST','SEPTEMBER','OKTOBER','NOVEMBER','DEZEMBER',
+    'EURO','USD','MIN','MAX','CA','ET','AL','US','UK','DE','EN','FR','ES',
+    'IT','JP','CN','QR','AI','KI','KW','NR','KG','MB','GB','CM','MM','KM',
+    'ML','DL','LLC','AG','GMBH','CO','KGAA','ORF','CD','DVD','PC','TV','VHS',
+    'DIY','CEO','CFO','COO','CTO','PR','HR','IT','OG','OGS',
+    'SPD','CDU','CSU','FDP','AFD','BSW','ARD','ZDF','RTL','WDR','NDR','SWR',
+    'BR','MDR','HRR','RBB','DLF','DPA','AFP','AP','CNN','BBC','NYT','WSJ',
+    'SZ','FAZ','SPON','TAG','SON','MONTAG','DIENSTAG','MITTWOCH','DONNERSTAG',
+    'FREITAG','SAMSTAG','SONNTAG','AKTION','SALE','BESTE','MONOLOGO','SILOS',
+    'UP','BSK','CATAN','PM','SU&SD','SUSD',
+    'SET','BOX','BIG','NEW','OLD','TOP','PRO','PLUS','ONE','TWO',
+    'USA','USD','DC',  # Common knowledge / proper names
+}
+
+SKIP_PATTERNS = [
+    r'^[A-Z]+\d',          # Product codes like S1R
+    r'^[A-Z]{2,}_',        # Snake-case IDs
+]
+
+PROPER_NAME_NEXT = re.compile(
+    r'\b(Advisory|Games|Verlag|Group|Entertainment|Productions?|Studios?|Media|'
+    r'Press|Publishing|International|Worldwide|Digital|Software|Interactive|'
+    r'Awards?|Supplement|Edition|Expansion|League|Legends|Party|Quest)\b'
+)
+
+# Known expansion terms for common abbreviations
+KNOWN_EXPANSIONS = {
+    'BGG': ['BoardGameGeek', 'Board Game Geek'],
+    'TCG': ['Trading Card Game', 'Trading Card Games'],
+    'CCG': ['Collectible Card Game', 'Collectible Card Games', 'UniVersus', 'Guilty Gear'],
+    'MTG': ['Magic The Gathering', 'Magic: The Gathering'],
+    'RPG': ['Roleplaying Game', 'Role Playing Game', 'Rollenspiel'],
+    'TTRPG': ['Tabletop Roleplaying', 'Tabletop Rollenspiel', 'Tabletop-Rollenspiel'],
+    'VFX': ['Visual Effects', 'Spezialeffekte'],
+    'UVP': ['unverbindliche Preisempfehlung', 'unverbindlichen Preisempfehlung'],
+    'SHUX': ['Community-Treffen', 'Convention', 'jährliche'],
+    'GAMA': ['Tabletop Game Association'],
+    'KTBG': ['Kids Table Board Gaming'],
+    'CMYK': ['CMYK Games'],
+    'SVS': ['Schweizer Spielwarenverband'],
+}
+
+
+def has_expansion_nearby(ctx, abbr):
+    """Check if known expansion for this abbreviation appears nearby."""
+    if abbr not in KNOWN_EXPANSIONS:
+        return False
+    for term in KNOWN_EXPANSIONS[abbr]:
+        if re.search(re.escape(term), ctx, re.IGNORECASE):
+            return True
+    return False
+
+
+def is_in_parens_definition(ctx, abbr):
+    """Check for any definition-in-parentheses pattern around the abbreviation."""
+    # Strip the base abbr (no plural 's')
+    base = abbr.rstrip('s') if abbr.endswith('s') else abbr
+    
+    # Pattern A: "Full Term (ABBR)" — definition BEFORE
+    # e.g., "Tabletop-Rollenspiele (TTRPGs)", "Kids Table Board Gaming (KTBG)"
+    a = re.search(
+        r'([A-ZÄÖÜ][A-Za-zäöüß\-]{3,}(?:\s[A-ZÄÖÜ][A-Za-zäöüß\-]{2,}){0,6})'
+        r'\s*\(' + re.escape(base) + r's?\)', ctx)
+    if a:
+        return True
+    
+    # Pattern B: "ABBR (explanation)" or "ABBR-Compound (explanation)" — definition AFTER
+    # e.g., "VFX-Künstler (Visual Effects)", "SHUX-Convention (das jährliche Treffen)"
+    b = re.search(
+        re.escape(base) + r's?(?:-[A-Za-zäöüß]+)?\s*\(([^)]{3,80})\)', ctx)
+    if b:
+        return True
+    
+    # Pattern C: "FullTerm(ABBR)" — no space
+    # e.g., "BoardGameGeek(BGG)"
+    c = re.search(
+        r'([A-ZÄÖÜ][A-Za-zäöüß\-]{3,}(?:\s[A-ZÄÖÜ][A-Za-zäöüß\-]{2,}){0,4})'
+        r'\(' + re.escape(base) + r's?\)', ctx)
+    if c:
+        return True
+    
+    return False
+
+
+def is_full_term_before(ctx, abbr):
+    """Check if full term appears right before abbreviation without parens.
+    e.g., "Schweizer Spielwarenverband SVS"
+    """
+    base = abbr.rstrip('s') if abbr.endswith('s') else abbr
+    m = re.search(
+        r'([A-ZÄÖÜ][A-Za-zäöüß\-]{3,}(?:\s[A-ZÄÖÜ][A-Za-zäöüß\-]{2,}){1,6})'
+        r'\s+' + re.escape(base) + r's?(?:\s|[,.!?;:\)]|$)', ctx)
+    if m:
+        full_term = m.group(1)
+        if len(full_term) > 10:
+            return True
+    return False
+
+
+def main():
+    if len(sys.argv) < 2:
+        print("Usage: check_abbreviations.py <artikel.html>", file=sys.stderr)
+        sys.exit(2)
+    
+    filepath = sys.argv[1]
+    
+    try:
+        with open(filepath) as f:
+            content = f.read()
+    except FileNotFoundError:
+        print(f"File not found: {filepath}", file=sys.stderr)
+        sys.exit(2)
+    
+    # Extract article body
+    m = re.search(r'<article>(.*?)</article>', content, re.DOTALL)
+    txt = m.group(1) if m else content
+    txt = re.sub(r'<[^>]+>', ' ', txt)
+    txt = re.sub(r'\s+', ' ', txt).strip()
+    txt = unescape(txt)
+    
+    # Find Quellen section boundary (less strict checks there)
+    # There may be multiple Quellen blocks (old + new format).
+    # Find the FIRST one and skip everything after it (with some margin)
+    quellen_pos = txt.find('Quellen')
+    if quellen_pos < 0:
+        quellen_pos = len(txt)
+    else:
+        # Allow 250 chars before the "Quellen" heading for old-style source blocks
+        quellen_pos = max(0, quellen_pos - 250)
+    
+    # Find all ALL-CAPS sequences (2+ letters)
+    abbrs = re.findall(r'\b[A-Z]{2,}(?:&[A-Z]{2,})?\b', txt)
+    counts = Counter(abbrs)
+    
+    # Filter candidates
+    candidates = {}
+    for abbr, count in counts.items():
+        if abbr in SKIP_WORDS:
+            continue
+        if any(re.match(p, abbr) for p in SKIP_PATTERNS):
+            continue
+        candidates[abbr] = count
+    
+    issues = []
+    for abbr in sorted(candidates.keys()):
+        # Match both singular and plural: ABBR or ABBRs
+        pattern = re.compile(r'\b' + re.escape(abbr) + r's?\b')
+        first = pattern.search(txt)
+        if not first:
+            continue
+        
+        pos = first.start()
+        
+        # Skip if in Quellen section
+        if 0 < quellen_pos < pos:
+            continue
+        
+        # Context after the match
+        ctx_after = txt[pos:pos+80]
+        
+        # Skip proper names (followed by Advisory, Games, Verlag, Awards, etc.)
+        if PROPER_NAME_NEXT.search(ctx_after[:50]):
+            continue
+        
+        # Skip slash-separated lists
+        before_slash = txt[max(0, pos-15):pos]
+        after_slash = txt[pos+len(first.group()):pos+len(first.group())+15]
+        if '/' in before_slash or '/' in after_slash:
+            continue
+        
+        # Build full context for definition checking
+        ctx_start = max(0, pos - 250)
+        ctx_end = min(len(txt), pos + 150)
+        ctx = txt[ctx_start:ctx_end]
+        
+        # Check definition patterns
+        if is_in_parens_definition(ctx, abbr):
+            continue
+        if is_full_term_before(ctx, abbr):
+            continue
+        if has_expansion_nearby(ctx, abbr):
+            continue
+        
+        # Build issue report
+        ctx_before = txt[max(0, pos-40):pos]
+        issues.append(
+            f'{abbr}: ...{ctx_before[-60:]}«{first.group()}»{ctx_after[:60]}...')
+    
+    if issues:
+        for i in issues:
+            print(f'  UNDEFINIERT: {i}')
+        sys.exit(1)
+    else:
+        print('  Alle Abkürzungen definiert')
+        sys.exit(0)
+
+if __name__ == '__main__':
+    main()
diff --git a/verify_article.sh b/verify_article.sh
index c308b41..85908be 100755
--- a/verify_article.sh
+++ b/verify_article.sh
@@ -150,7 +150,7 @@ fi
 if echo "$BODY" | grep -q 'background:#f0f4f8'; then
     ok "Infobox vorhanden"
 else
-    fail "Infobox fehlt (Fakten-Tabelle mit #f0f4f8 Hintergrund)"
+    warn "Infobox fehlt (Fakten-Tabelle mit #f0f4f8 Hintergrund)"
 fi
 
 # ── CHECK 16: Read-More NACH erstem Text-Absatz (Daniel 23.06.2026) ──
@@ -183,11 +183,17 @@ if [ -n "$SRC_START" ]; then
     fi
 fi
 
-# ── CHECK 18: Recherchestand-Zeile vorhanden (nur Datum, keine Quellen-Wiederholung) ──
-if echo "$BODY" | grep -q 'Recherchestand:'; then
-    ok "Recherchestand-Zeile vorhanden"
+# ── CHECK 19: Abkürzungen bei Erstverwendung definiert ─────────────
+echo ""
+echo "--- Abkürzungs-Check ---"
+python3 /home/hermes/workspace/check_abbreviations.py "$FILE"
+ABBR_EXIT=$?
+if [ $ABBR_EXIT -eq 0 ]; then
+    ok "Alle Abkürzungen bei Erstverwendung definiert"
+elif [ $ABBR_EXIT -eq 1 ]; then
+    fail "Undefinierte Abkürzungen gefunden (siehe oben)"
 else
-    fail "Recherchestand-Zeile fehlt"
+    warn "Abkürzungs-Check fehlgeschlagen (Python-Fehler?)"
 fi
 
 # ── SUMMARY ─────────────────────────────────────────────────────────