CUA feature - extraction data combined assisstant's message (#2157)

skyvern/forge/agent.py

@@ -1165,7 +1165,9 @@ class ForgeAgent:
             ):
                 working_page = await browser_state.must_get_working_page()
                 # refresh task in case the extracted information is updated previously
-                task = await app.DATABASE.get_task(task.task_id, task.organization_id)
+                refreshed_task = await app.DATABASE.get_task(task.task_id, task.organization_id)
+                assert refreshed_task is not None
+                task = refreshed_task
                 extract_action = await self.create_extract_action(task, step, scraped_page)
                 extract_results = await ActionHandler.handle_action(
                     scraped_page, task, step, working_page, extract_action

skyvern/forge/prompts/skyvern/extract-information.j2

@@ -14,7 +14,7 @@ If you are trying to extract the href links which are using the jinja style like
 User Data Extraction Goal: {{ data_extraction_goal }}
 
 {% if previous_extracted_information %}
-Previous contexts or thoughts: {{ previous_extracted_information }}
+Previous contexts or thoughts: ```{{ previous_extracted_information }}```
 {% endif %}
 
 {% if error_code_mapping_str %}

skyvern/forge/sdk/schemas/tasks.py

@@ -232,9 +232,6 @@ class Task(TaskBase):
         if status.cant_have_extracted_info() and extracted_information is not None:
             raise ValueError(f"status_cant_have_extracted_information({self.task_id})")
 
-        if self.extracted_information is not None and extracted_information is not None:
-            raise ValueError(f"cant_override_extracted_information({self.task_id})")
-
         if self.failure_reason is not None and failure_reason is not None:
             raise ValueError(f"cant_override_failure_reason({self.task_id})")
 

skyvern/webeye/actions/parse_actions.py

@@ -338,20 +338,19 @@ async def parse_cua_actions(
             intention=reasoning,
         )
         if skyvern_action_type == "complete":
-            if not task.data_extraction_goal and useful_information:
-                LOG.info(
-                    "Updating task with useful information",
-                    task_id=task.task_id,
-                    organization_id=task.organization_id,
-                    useful_information=useful_information,
-                    assistant_message=assistant_message,
-                    reasoning=reasoning,
-                )
-                await app.DATABASE.update_task(
-                    task.task_id,
-                    organization_id=task.organization_id,
-                    extracted_information=assistant_message,
-                )
+            LOG.info(
+                "Updating task with useful information",
+                task_id=task.task_id,
+                organization_id=task.organization_id,
+                useful_information=useful_information,
+                assistant_message=assistant_message,
+                reasoning=reasoning,
+            )
+            await app.DATABASE.update_task(
+                task.task_id,
+                organization_id=task.organization_id,
+                extracted_information=assistant_message,
+            )
             action = CompleteAction(
                 reasoning=reasoning,
                 intention=reasoning,