diff --git a/atd-etl/cris_import/README.md b/atd-etl/cris_import/README.md
index b6af7ee3e..58567c065 100644
--- a/atd-etl/cris_import/README.md
+++ b/atd-etl/cris_import/README.md
@@ -12,3 +12,12 @@ The contents of this directory define a Docker image that can be used in an Airf
 ### Zip file acquisition
 
 The CRIS import usually works by pulling down a zip archive from the SFTP endpoint. However, during development, it's much easier if you can have it use a zip file that you have locally on your machine instead. This can be accomplished by putting a zip file (still encrypted and using "CRIS extract" password) in a folder named `atd-etl/cris_import/development_extracts/`. Create the directory if needed. If there are no zip files in that folder, the program will automatically revert to inspecting the SFTP endpoint.
+
+### Local testing
+
+Make a copy of `env-template` and name it `env`. Fill in the values using the 1Password Connect Server secrets (see entries titled `Endpoint for 1Password Connect Server API` and `Vault ID of API Accessible Secrets vault`) and your personal access token.
+
+Drop a CRIS extract zip file into your development folder as described above, and run the import script:
+```bash
+docker compose run cris-import
+```
diff --git a/atd-etl/cris_import/cris_import.py b/atd-etl/cris_import/cris_import.py
index 9141ceee7..e5bed7c74 100755
--- a/atd-etl/cris_import/cris_import.py
+++ b/atd-etl/cris_import/cris_import.py
@@ -612,7 +612,23 @@ def align_records(map_state):
             input_column_names = util.get_input_column_names(pg, map_state["import_schema"], table, target_columns)
 
             # iterate over each imported record and determine correct action
+            should_skip_update = False
+
             for source in imported_records:
+                # Check unique key columns to make sure they all have a value
+                for key_column in key_columns:
+                    key_column_value = source[key_column]
+                    if key_column_value is None:
+                        print("\nSkipping because unique key column is missing")    
+                        print(f"Table: {table}")
+                        print(f"Missing key column: {key_column}")
+                        should_skip_update = True
+
+                # If we are missing a column that uniquely identifies the record, we should skip the update
+                if should_skip_update:
+                    for key_column in key_columns:
+                        print(f"{key_column}: {source[key_column]}")
+                    continue
 
                 # generate some record specific SQL fragments to identify the record in larger queries
                 record_key_sql, import_key_sql = util.get_key_clauses(table_keys, output_map, table, source, map_state["import_schema"])
diff --git a/atd-etl/cris_import/env-template b/atd-etl/cris_import/env-template
new file mode 100644
index 000000000..7650c06f4
--- /dev/null
+++ b/atd-etl/cris_import/env-template
@@ -0,0 +1,3 @@
+OP_API_TOKEN=
+OP_CONNECT=
+OP_VAULT_ID=