Run history
What every run records — outcome, why it fired, and the full transcript of what an agent actually did.
Work that happens while you are away is the work you have the least reason to trust. So it is the work that has to be the most legible.
Every fire of every task is recorded. You can read one task's runs, or All runs across every task at once.
What a run tells you#
Whether it worked. Succeeded, or failed with the exit code it failed on.
When it started, and how long it took.
Why it fired — and this is the one people do not expect. Every run records its own cause: a schedule, an editor event, or you pressing the button. A task that fired at a strange hour can be asked what started me, and it answers from the record rather than from inference.
What it printed. The tail of the output is always there, for every kind of task.
Note
Some runs cannot know their own duration — a run interrupted by a restart has a last-seen time, not a finish time. Those show as unknown rather than as a number that looks precise and is invented. A history that guesses is worse than one that admits the gap.
The transcript#
For agent and flow tasks there is more than output: the run's whole conversation, replayed.
Not a log file — the same renderers that draw a live agent session. Each turn, each tool call, what it read and what it wrote, in the order it happened. One surface whether you are watching it live or opening it a week later.
This is the promise from About applied to unattended work: you can see what the model actually did, not a summary of what it claims to have done.
Note
A transcript is an enrichment; the printed output is the guarantee. Shell tasks have no conversation to show, and a transcript can also be unavailable for other reasons. When there is none, the page simply shows the output. Absence here is normal, never an error.
Nothing, versus could not ask#
A small thing worth naming, because it is the difference between a tool you can reason with and one you cannot.
An empty run list and a run list that failed to load look identical if you are careless about it — both render as "no runs". Semantix treats them as different questions. If the history cannot be fetched, it says so. It will not show you a confident, empty, wrong history and let you conclude that your task has never fired.
The same instinct runs through the rest of the Task Manager: a missed run is recorded as missed rather than quietly skipped, and a duration that cannot be known is not invented.