# Run diagnostics

Runs presents structured execution details rather than an unrestricted raw-log viewer.

![Run diagnostics](https://www.foxora.ai/docs/images/v5.0.8/reference/runs.webp)

*Studio interface in English. Sample data.*

## Collect useful evidence

1. Search for the affected task or run ID.
2. Expand `Details` and record its source, state, duration, plan, and failed child row.
3. Copy the session ID and compare it with the task transcript.
4. Inspect command or tool output in the transcript or relevant `Stage` surface.
5. Check `Settings` → `System` for endpoint, engine, workspace, or Runtime health when several runs fail together.

## Expected result

You have enough evidence to separate task logic, tool input, plugin authentication, local permission, and service health failures. Share exact IDs and error text with support; a screenshot of the top-level Failed badge alone is usually insufficient.