Review → Agents
The Agents tab is the history of every AI agent execution in the workspace — both the ones you trigger by hand (Quick Assist on a resource or a config diff) and the ones that run in the background (kicked off by automation, reports, or a schedule). Use it to see what the AI was asked, what it produced, and what it cost.
This is distinct from Automate → Runs, which tracks playbook executions. A playbook may invoke an agent; that agent’s work shows up here, while the playbook’s overall run shows up under Automate → Runs.
Runs associated with a resource or location also appear in its History tab, with the same entry and output dialog. Location History includes its sub-locations and their resources. Report-generated runs are shown here only.
What you can do
- Browse agent runs on a day-grouped timeline.
- Filter to the runs you care about.
- Open a run to read its output.
Filters
| Filter | Notes |
|---|---|
| Kind | Quick Assist (you triggered it) or Background (the platform did). |
| Trigger | What started it — User, Automation, Report, or Schedule. |
| Status | Running, Completed, or Failed. |
| From / To | Limit by run date. |
Reading the timeline
Each entry shows the run time, a view (eye) button, the subject — the resource (linked while it exists, or shown as a deleted resource) or location the agent ran against — and a detail line with the kind, the trigger, the run duration, and the token count. A colored chip names the specific agent (for example Config Analysis or Session Analysis): blue for Quick Assist runs, purple for Background runs.
Opening a run
Click the view button to open a run’s output — formatted text, raw structured output when that is what the agent returned, or No output recorded. Timing and token usage stay on the timeline entry’s detail line.
Permissions
- View agent runs:
app.resources.read(you see runs for the resources you can access).