Checkpoints and take over

Checkpoints record where a task stands while a tool works. They make an interrupted task easier to continue in the same tool or another one.

What a checkpoint contains

A connected tool saves a checkpoint after a meaningful step or commit. It records what is done, what remains and the next step. It may also name changed files, the current plan and the last error. Saving one renews the task’s claim.

While Seggo is open, or seggo watch is running, Seggo can also notice changed work. It records branch and commit subjects, plus file names and line counts from uncommitted changes. It does not save file contents in these automatic checkpoints.

Open a task and choose Progress to see checkpoints, commits and status changes together. Each task keeps its newest 50 checkpoints. Older entries are folded into an earlier-work summary.

When a tool stops

Seggo can mark a held task Stalled when imported history shows a usage or rate limit, or when the task has been quiet with uncommitted changes. The task returns to In progress if the tool saves a new checkpoint or Seggo sees it working again.

If a claim lease expires, the task returns to To do. Its checkpoints, branch, folder and files stay in place. Nothing is deleted.

Take over in another tool

Open the task and choose Hand to… or Continue in…. Pick a tool found on your computer and review the context it will receive. The new tool should first read the task and latest checkpoint, then continue in the recorded task folder.

Take over is most useful when the first tool hit a limit, crashed or needs a different strength. The new tool should not redo completed steps unless the checkpoint says they need verification.