Agent Health Monitor
Separate activity from progress.
Inspect actual task activity, waiting periods and known blockers.
Separate activity from progress.
During a build, especially when you are unsure what a task is waiting for.
A spinning indicator does not explain whether a task is working, waiting or repeating failures. The monitor shows recorded task signals together.
Inspect state, recent activity, requests and errors. Use the findings to decide whether to wait, provide input or investigate recovery.
How to get started.
From your first choice to the result. Here is how to use this feature in your own project.
- 01
Check the most recently recorded activity.
- 02
Distinguish permission waits, provider responses, errors and intentional pauses.
- 03
Open the relevant task or recovery check for closer investigation.
How it could work for you.
Check a task that appears quiet and discover it is waiting for a permission decision.
How to assess the result
The monitor observes reported signals, not the model’s private reasoning or an independent correctness measure.
Where to find it.
Conversation → Tools → Agent Health Monitor
Available actions depend on your plan, mode and project state. Features that require a task become available within an appropriate conversation.What is saved?
Observed events remain in the task history.
What you need to know
Silence alone does not prove a model is stuck. Internal provider activity is not always visible to Linqbase.