# Run Step Trace block

> Agent Run Detail with Inline Step Trace and Editable Settings Panel

Single-page detail for one agent run: header with live actions, step trace with per-step tool calls and latency, plus a row-editable settings panel. ReUI Frame, Timeline, Badge, Alert; shadcn Avatar, Button. Use: run debugging, ops monitoring.

- Type: Pro block
- Category: [Agents](https://web.dev.reui.io/blocks/solutions/agents.md)
- Page: https://web.dev.reui.io/blocks/solutions/agents/solution-agents-3
- MCP: `get_block("solution-agents-3")`
- Built with: Alert, Avatar, Badge, Button, Collapsible, Combobox, Dropdown Menu, Field, Frame, Input Group, Item, Select, Separator, Spinner, Switch, Timeline, Tooltip
- Installs with a ReUI Pro or Ultimate license; setup: https://web.dev.reui.io/docs/license-setup

## Best for

A single failed run: a step trace with tool call latency, a Retry Step alert on the failing step and settings edited row by row

## Agent notes

- The page shows one run, replace `RUN_IDENTITY`, `RUN_STEPS` and `DEFAULT_RUN_SETTINGS` in data.tsx with the run you load
- `saveRowEditing` and `saveAllEditing` wait `SAVE_DELAY_MS` on a timer, then set local state and a `toast`, so wire an API call there
- `handleApprove`, `handlePause` and both `handleRetryStep` handlers only fire a `toast`, so no run changes state

## When not to use

- A live run feed watched by the requester, with Summary, Detailed and Raw log views: use [Live Run Feed](https://web.dev.reui.io/blocks/ai-agents/agent-activity/agent-activity-2.md) (`get_block("agent-activity-2")`)
- Many failed runs listed for triage, each opened in a side sheet: use [Failed Run Triage](https://web.dev.reui.io/blocks/solutions/agents/solution-agents-7.md) (`get_block("solution-agents-7")`)

## Third-party libraries

- Sonner v2: [API reference](https://sonner.emilkowal.ski/toast), [docs](https://sonner.emilkowal.ski/getting-started)

## Install

```bash
npx shadcn@latest add @reui/solution-agents-3
```

## About Agents blocks

### When to use

- Run monitoring for a fleet of agents: a live overview, a run queue and failed run triage
- Human review of agent work: an approval inbox, one run's step trace and a nightly evaluation report
- Control pages for agent behavior: tool access, playbooks, run settings and stored memory

### When not to use

- One run narrated live to the person who started it, inside a chat or task view: use the [Agent Activity blocks](https://web.dev.reui.io/blocks/ai-agents/agent-activity.md)
- Model provider routing, prompt reviews, token spend and safety policies for LLM features: use the [AI Ops blocks](https://web.dev.reui.io/blocks/solutions/ai-ops.md)
- An account, workspace, notification or integration page of labeled switches and selects: use the [Settings blocks](https://web.dev.reui.io/blocks/application/settings.md)

### Agent notes

- Every block sets a `min-h-svh` wrapper with its own padding and max width in page.tsx, its root component or both; drop them inside an app shell
- No block mounts the `sonner` toaster, so every `toast` call stays invisible until your layout renders one
- No handler reaches a backend: actions change local `useState` and fire a `toast`, so wire each one to your agent API

Every ReUI item has a twin like this one; the index is https://web.dev.reui.io/llms.txt. Agents with the ReUI MCP (https://web.dev.reui.io/docs/mcp) get the same guidance from the getters named above.