diff --git a/markov/management/commands/fix_chains_with_nicks_via_bridge.py b/markov/management/commands/fix_chains_with_nicks_via_bridge.py
index 29a9f13..3ed8278 100644
--- a/markov/management/commands/fix_chains_with_nicks_via_bridge.py
+++ b/markov/management/commands/fix_chains_with_nicks_via_bridge.py
@@ -16,7 +16,7 @@ class Command(BaseCommand):
             self.stdout.write(self.style.NOTICE(f"scanning context {context}..."))
             # get starting states that look like they came over the bridge
             bridge_states = context.states.filter(k1=MarkovState._start1, k2=MarkovState._start2,
-                                                  v__regex=r'<.*>')
+                                                  v__regex=r'<[A-Za-z0-9_]+>', context=context)
             self._chain_remover(context, bridge_states)
             # get states that look like mentions
             for target in context.markovtarget_set.all():
@@ -27,7 +27,7 @@ class Command(BaseCommand):
                     all_nicks = target.channel.server.nickname
                 mention_regex = r'^(({nicks})[:,]|@({nicks}))$'.format(nicks=all_nicks)
                 mention_states = context.states.filter(k1=MarkovState._start1, k2=MarkovState._start2,
-                                                       v__regex=mention_regex)
+                                                       v__regex=mention_regex, context=context)
                 self._chain_remover(context, mention_states)
 
     def _chain_remover(self, context, start_states):
@@ -50,7 +50,8 @@ class Command(BaseCommand):
                     self.stdout.write(self.style.NOTICE(f"      upserting state based on {leaf_state}"))
                     # get/update state without the nick from the bridge
                     try:
-                        updated_leaf = MarkovState.objects.get(k1=second_state.k1, k2=leaf_state.k2, v=leaf_state.v)
+                        updated_leaf = MarkovState.objects.get(k1=second_state.k1, k2=leaf_state.k2, v=leaf_state.v,
+                                                               context=context)
                         updated_leaf.count += leaf_state.count
                         updated_leaf.save()
                         self.stdout.write(self.style.SUCCESS(f"        updated count for {updated_leaf}"))
@@ -68,7 +69,8 @@ class Command(BaseCommand):
                 # take care of the new middle state
                 self.stdout.write(self.style.NOTICE(f"    upserting state based on {second_state}"))
                 try:
-                    updated_second = MarkovState.objects.get(k1=start_state.k1, k2=start_state.k2, v=second_state.v)
+                    updated_second = MarkovState.objects.get(k1=start_state.k1, k2=start_state.k2, v=second_state.v,
+                                                             context=context)
                     updated_second.count += second_state.count
                     updated_second.save()
                     self.stdout.write(self.style.SUCCESS(f"      updated count for {updated_second}"))