diff --git a/.github/workflows/nrn-modeldb-ci.yaml b/.github/workflows/nrn-modeldb-ci.yaml
index 2653f35..0619fab 100644
--- a/.github/workflows/nrn-modeldb-ci.yaml
+++ b/.github/workflows/nrn-modeldb-ci.yaml
@@ -159,7 +159,6 @@ jobs:
           python -m pip install $NEURON_V1
         fi
         nrn_ver=`python -c "from neuron import __version__ as nrn_ver; print(nrn_ver)"`
-        ps uxf # debug
         runmodels --gout --workdir=$nrn_ver $MODELS_TO_RUN
         # Filter out the gout data before generating HTML reports. The HTML
         # diff uses the original gout files on disk anyway. Compress the large
@@ -186,7 +185,6 @@ jobs:
           python -m pip install $NEURON_V2
         fi
         nrn_ver=`python -c "from neuron import __version__ as nrn_ver; print(nrn_ver)"`
-        ps uxf # debug
         runmodels --gout --workdir=$nrn_ver $MODELS_TO_RUN
         # Filter out the gout data before generating HTML reports. The HTML
         # diff uses the original gout files on disk anyway. Compress the large
diff --git a/modeldb/modeldb.py b/modeldb/modeldb.py
index 53bc402..3927d01 100644
--- a/modeldb/modeldb.py
+++ b/modeldb/modeldb.py
@@ -133,6 +133,13 @@ def query(field):
         }
         # If we were passed a non-None `model_list`, restrict those models now.
         if model_list is not None:
+            missing_ids = set(model_list) - set(metadata.keys())
+            if len(missing_ids):
+                raise Exception(
+                    "Model IDs {} were explicitly requested, but are not known NEURON models.".format(
+                        missing_ids
+                    )
+                )
             metadata = {model_id: metadata[model_id] for model_id in model_list}
         # For each model in `metadata`, check if a cached entry exists and is
         # up to date. If not, download it.