Hello @rabbit, Hello @everyone,
I am no more able to use Google services and Amazon website with LAM Playground.
Sometimes it happened to trigger security measures, but this thing is happening on a daily base this week.
LAM perpetrates to trigger identity verification on google services on R1, sometimes I don’t notice it and it tries to reset my password, yesterday I manually logged in from LAM on the web and today it triggered the identity test again.
On Amazon, it finds a login page every single time I try to make a purchase, making it frustrating as it was when trying to use the old playwright based integration system you had for Uber and Doordash.
The thing is I am using R1 on a daily base, probably using it as it was intended to be used, with a Data SIM inserted and making it perform every task I usually perform on my phone. I can’t always be there with my laptop babysitting the LAM when it struggles to login.
As an app developer, I perfectly know that AI bots could sometimes trigger security measures on websites, but it seamed like R1 reached a balance till this week. Maybe is necessary to find some compensations, for example a Wrapper card of the interface LAM is controlling, showing a minimized version of the web page in real time, at least I can ask someone to help me logging in from R1 directly when prompted to.
FYI: Just today, Google raised a “Critical security alert” due to a “Windows device with suspicious activity” from LAM Playground r1.
I created a draft email, LAM paused (
), then I asked it to send it and it did. Then Google logged me out on the LAM device and sent me to a login screen. I logged in using the LAM playground and then it sent an alert directly to my phone (not via SMS) asking me if it was really me trying to sign in, I said “Yes” and had to choose the correct number on my phone that matched the number from the LAM Playground screen.
Did LAM really send an e-mail on your behalf? I’m having it strougling a lot with GMail…
For the security alert: it happens me once every 5 to 10 days and it’s terrible. I hope Rabbit and other partners can get out of it by building relationship… Is sufficient that google recognizes the pattern of the LAM to avoid blocking it.
Yeah, LAM Playground r1:
Go to gmail.com and draft an email to
<address>stating that I want to make a reservation for tomorrow at 5 PM. Only draft this email, do not send this email.
It then started a new email, added in the email address, created a subject, wrote a polite email, and signed it with my first name based off of the email account. Then it saved it as a Draft. I later said
Go to gmail.com, find the draft titled Reservation Request and send it.
And it sent it!
[Edited to put <angle brackets> visibly in post.]
That’s great! When I asked it to write or draft e-mails, it remained stuck on the message body field and it won’t write nothing.