The worst part is: it had already spent 31 minutes doing deep code analysis, reading files, and running skills — all that progress is just gone. No partial results, no way to pick up where it left off. I wasted all those tokens and time for nothing.
It would be a huge improvement if the tool could:
- Save intermediate results/checkpoints as it goes (analysis summaries, parsed configs, partial content)
- Let me retry the final output step without redoing all the expensive pre-processing
- Resume from the last successful point when hitting token limits, instead of starting over completely
Right now, if a big task hits this limit, you lose everything — that's really frustrating for long workflows.
The worst part is: it had already spent 31 minutes doing deep code analysis, reading files, and running skills — all that progress is just gone. No partial results, no way to pick up where it left off. I wasted all those tokens and time for nothing.
It would be a huge improvement if the tool could:
Right now, if a big task hits this limit, you lose everything — that's really frustrating for long workflows.