> ## Documentation Index
> Fetch the complete documentation index at: https://docs.alignr.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Your daily review

> Find your next task, read the dashboard and follow work through issues and activity.

Start at **Overview** to decide what needs attention. Use **Workspace health** for collection and coverage problems, **Issues** for observed findings, and **Reviews** for human assessment work.

## Read the dashboard together

Overview requires `dashboard.read`. Its **Next step** points towards unfinished setup or assessment work.

| Card                 | What it answers                                                   |
| -------------------- | ----------------------------------------------------------------- |
| **Live alignment**   | What proportion of checks with a current live result are passing? |
| **Checks completed** | What proportion of applicable checks have current live results?   |
| **Open issues**      | Which observed findings need investigation?                       |
| **Fix history**      | What has been proposed or attempted, and with what outcome?       |

Read alignment alongside coverage. In an illustrative example, ten passing results and ten checks without usable evidence do not establish that all twenty expectations are met. **Not assessed yet** means there are no current verified results to score; it is not a score of zero.

Open a client in **Client checks** to inspect individual results. The **Needs attention** queue links to issue evidence. **Tool health** is workspace-wide; open **Workspace health** for detailed collection and coverage checks.

## Refresh evidence deliberately

Select **Refresh evidence** when you need a discovery sweep. This requires both `sweep.execute` and `sweep.read`. It requests workspace collection; the global client selector does not turn this action into a single-client sync.

Review the sweep outcome and each source's result. A finished attempt with no facts is not proof of a healthy environment: a source may have failed or supplied no usable observations.

Evidence collection and evaluation are distinct milestones. After confirming the intended evidence is available, use the relevant **Run checks** workflow and inspect its evaluation time. Refreshing the dashboard alone does not establish a fresh assessment.

## Choose your next task

1. **Missing or stale evidence:** open [Workspace health](/guides/workspace-health), fix the source or mapping problem and confirm fresh observations.
2. **An observed failure:** open [Issues](/guides/triage-issues), inspect evidence and decide the follow-up.
3. **A question requiring human judgement:** open **Reviews** or the client's manual checks and record the review evidence.
4. **A proposed or attempted fix:** open **Activity** for approvals and fix history.

**Checkpoint:** the next action has a client, a reason and an owner. A quiet issue queue is useful only alongside an understood coverage picture.

## Find your way around

The main navigation separates daily work—**Overview**, **Clients**, **Issues**, **Reviews**, **Activity**—from workspace configuration such as **Standards**, **Integrations**, **MCP** and **Settings**. On a narrow screen, open **Main navigation** for the same permitted destinations.

Use the client selector and the page's own scope labels together. A page reached from a specific client or control may carry an explicit filter. Check that scope before interpreting its counts or asking the assistant to explain it.

Press **⌘K** on macOS or **Ctrl+K** to open search. It offers page shortcuts immediately and searches clients and issues after at least two characters. Use the arrow keys and Enter to open a result. This app search is separate from the documentation site's search.

The header also provides **Documentation and help**, **Ask Alignr**, theme switching and the **Account menu**, including **Account settings** and **Sign out**. Permission-dependent pages and actions may be hidden.

## Inspect Activity and export audit history

Open **Activity** to choose:

| Destination     | Purpose                                                | Permission                                                           |
| --------------- | ------------------------------------------------------ | -------------------------------------------------------------------- |
| **Approvals**   | Review proposed actions before execution.              | `approval.read` to view; `remediation.approve` to decide.            |
| **Fix history** | Review runs, verification and supported rollback.      | `remediation.read`; `remediation.rollback` to reverse eligible runs. |
| **Audit trail** | Inspect recorded actors, actions and affected records. | `audit.read`.                                                        |

The Audit trail destination opens **Logs**. Filter by **Actor kind**, **Action**, **Record type**, **Record id**, **From** and **To**. A closed issue's **view audit trail** action opens its record-specific history.

Select **Export filtered CSV** to export the matching records, or **Export CSV** without filters. Export uses the whole matching result set rather than only the displayed page, up to **50,000 rows**. If the export message says the row limit was reached, narrow the date range and export the remaining periods. Do not treat a truncated file as the complete trail.

Filters are reflected in the page URL, so an authorised colleague can open the same filtered view. Exported files should follow your organisation's handling rules for operational records.

## Distinguish alerts from completed work

The workspace-health banner points to underlying operational warnings. Read success and error messages after an action, then confirm the saved state or resulting record. Dismissing a message does not resolve its underlying issue.

If a page shows last-known data after a refresh error, restore access and refresh before describing it as current. Follow [troubleshooting](/guides/troubleshooting) when an action or result does not match your expectation.