Skip to content

Automate browser based workflows with AI

fix: remove default login complete_criterion from script path (#5650)

Shuchang Zheng1092818d9e2c
diff --git a/skyvern/services/script_service.py b/skyvern/services/script_service.py
index c21b9b262..18fe2883a 100644
--- a/skyvern/services/script_service.py
+++ b/skyvern/services/script_service.py
@@ -20,7 +20,6 @@ from jinja2.sandbox import SandboxedEnvironment
from skyvern.config import settings
from skyvern.constants import (
BROWSER_DOWNLOADING_SUFFIX,
- DEFAULT_LOGIN_COMPLETE_CRITERION,
GET_DOWNLOADED_FILES_TIMEOUT,
SAVE_DOWNLOADED_FILES_TIMEOUT,
)
@@ -481,17 +480,12 @@ async def _create_workflow_block_run_and_task(
# Include script parameters as navigation_payload so handlers
# (e.g. file upload) can find URLs like resume_link in the payload.
nav_payload = context.script_run_parameters or None
- # Apply default complete_criterion for login blocks so cached scripts
- # get the same rigorous LLM verification as the agent path (SKY-8540).
- # Without this, the LLM only sees a generic navigation_goal and can
- # falsely mark login as complete when credentials were never entered.
- task_complete_criterion = DEFAULT_LOGIN_COMPLETE_CRITERION if block_type == BlockType.LOGIN else None
task = await app.DATABASE.tasks.create_task(
# fix HACK: changed the type of url to str | None to support None url. url is not used in the script right now.
url=url or "",
title=f"Script {block_type.value} task",
navigation_goal=prompt,
- complete_criterion=task_complete_criterion,
+ complete_criterion=None,
data_extraction_goal=prompt if block_type == BlockType.EXTRACTION else None,
extracted_information_schema=schema,
navigation_payload=nav_payload,