remove 2 longer tests because this test mainly shows if browseruse works end to end

This commit is contained in:
Magnus Müller
2025-09-08 08:35:18 -07:00
parent b2eb803e58
commit a19b583430
2 changed files with 0 additions and 16 deletions
@@ -1,6 +0,0 @@
name: Cloudflare captcha
task: Go to https://2captcha.com/demo/cloudflare-turnstile and solve the captcha, wait a few seconds, then click on check, wait a few more seconds for it to complete, then extract the "hostname" value from the displayed dictionary under "Captcha is passed successfully!"
judge_context:
- The agent must solve the captcha
- The hostname returned should be "example.com"
max_steps: 6
-10
View File
@@ -1,10 +0,0 @@
name: Google Maps 3d Screenshot
task: Go to google.com/maps and search for ETH Zurich Hauptgebäude. When found, close the side panel to see the map full screen. Then, if not already in Satellite View, switch to Satellite View. With Satellite View enabled, click the 3d icon to enable 3d view. Pan the map so that ETH Zurich Hauptgebäude and the Zurich Lake in the background are clearly visible. If able, take a screenshot.
judge_context:
- Agent must only use www.google.com/maps
- Agent should correctly search for ETH Zurich Hauptgebäude
- Agent should close the side panel
- After the agent performing the task the map should be visible in Satellite view, if it was not already
- The Agent should correctly click the correct 3d button to enable 3d mode
- The Agent should correctly pan the map so that ETH Zurich Hauptgebäude as well as the Zurich Lake in the background should be visible
max_steps: 25