> ## Documentation Index
> Fetch the complete documentation index at: https://docs.idemeum.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Run details
> The full record of a single run.
## Header
In the idemeum admin portal, go to **Runs** and click any run to open its full record: what was asked, what Assist did, what the employee approved and what it cost.
The header shows the goal in the employee's words, who ran it, on which device and when.
| Card | What it shows |
| ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Run status | **Ticket:** the run's result (Resolved, Escalated or Failed). **Tools:** how the steps went (resolved, partial, unresolved, failed or cancelled). **User:** whether the employee said the problem was fixed. |
| Duration | How long the run took. |
| LLM calls | How many AI model calls the run made. |
| Cost | The AI cost of the run, based on actual token usage. |
## Overview tab
A summary of the run.
* **Skill matched:** the skill that handled the request, linked to its page.
* **AI agent summary:** what Assist found, what the employee chose and what changed. It's the same summary the employee saw.
* **Tool counts:** how many steps ran and how each ended:
* **Approved:** the step ran.
* **Denied:** the employee declined it, a guardrail blocked it, or it failed.
* **Skipped:** Assist showed a preview but didn't make the change.
## Tool activity tab
Every step, in order. Use it to check exactly what Assist did, or to see why a run ended as partial.
Each step shows:
* **Tool and status:** Approved or Denied.
* **Rationale:** what the step was for, such as **"Scanning your home directory to find the largest folders"**.
* **Denial reason:** who stopped the step and why. In the example, a guardrail blocked `disk_scan` because the path was outside the employee's home folder, and the run continued with other checks.
* **Tool output:** the raw result, such as paths, sizes and counts. This is the evidence behind the summary.
## Context tab
Where and how the run happened.
| Section | What it shows |
| -------------- | ------------------------------------------------------------------------------------------------------------------ |
| Identifiers | Session ID and Agent ID. Copy them when checking logs or contacting support. |
| Device context | Hostname, OS version, platform, whether the employee is an admin, and MDM enrollment, as they were during the run. |
| LLM and source | How the run started, the AI provider and model, and tokens used (input, output and cached). |