Wonderful idea!
Attempted first run. Hooked up Google and Modal successfully.
Gave prompt, looked like it created a good plan, created code and it ran for a while on Modal, used about $2.50 on Modal via Modal dashboard. Looked like it completed several subtasks on Modal, but then
hung with no indications of why. I assume something in the created code hung, but no visibility on what/why.
It would be nice to see some of the artifacts on the researcher end and on the Modal end, perhaps directory listing, navigating, file browsing, opening and downloading?
How can one recover from a hung remote task? Via better prompting? Agents that watch for hanging and kill tasks and resume?
Wonderful idea!
Attempted first run. Hooked up Google and Modal successfully.
Gave prompt, looked like it created a good plan, created code and it ran for a while on Modal, used about $2.50 on Modal via Modal dashboard. Looked like it completed several subtasks on Modal, but then
hung with no indications of why. I assume something in the created code hung, but no visibility on what/why.
It would be nice to see some of the artifacts on the researcher end and on the Modal end, perhaps directory listing, navigating, file browsing, opening and downloading?
How can one recover from a hung remote task? Via better prompting? Agents that watch for hanging and kill tasks and resume?