fix: Change log-step to a jinja template rather than a simpleeval expression

agents-api/agents_api/activities/task_steps/get_value_step.py

@@ -1,4 +1,3 @@
-
 from beartype import beartype
 from temporalio import activity
 

agents-api/agents_api/activities/task_steps/log_step.py

@@ -6,8 +6,8 @@
     StepContext,
     StepOutcome,
 )
+from ...common.utils.template import render_template
 from ...env import testing
-from .base_evaluate import base_evaluate
 
 
 @beartype
@@ -17,8 +17,10 @@ async def log_step(context: StepContext) -> StepOutcome:
     try:
         assert isinstance(context.current_step, LogStep)
 
-        expr: str = context.current_step.log
-        output = await base_evaluate(expr, context.model_dump())
+        template: str = context.current_step.log
+        output = await render_template(
+            template, context.model_dump(), skip_vars=["developer_id"]
+        )
 
         result = StepOutcome(output=output)
         return result

agents-api/tests/test_execution_workflow.py

@@ -407,7 +407,9 @@ async def _(
                 "other_workflow": [
                     # Testing that we can access the input
                     {"evaluate": {"hello": '_["test"]'}},
-                    {"log": '_["hell"]'},  # <--- The "hell" key does not exist
+                    {
+                        "log": '{{_["hell"].strip()}}'
+                    },  # <--- The "hell" key does not exist
                 ],
                 "main": [
                     # Testing that we can access the input