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

# Action and site write tools

> Update the status of an Action Center recommendation, add pages to Site Diagnostics and start page audits through MCP. Page audits use 5 credits per page.

These tools update the Action Center and Site Diagnostics. Read [MCP write access](https://docs.qwairy.co/mcp/write/overview.md) first for plans, scopes, roles, `dryRun`, `expectedVersion` and confirmation tokens. Examples use synthetic data.

| Tool | Scope | Minimum role | Confirmation | REST equivalent |
| - | - | - | - | - |
| `update_action_status` | `write:actions` | Manager | `never` | [`PATCH /brands/{brandId}/actions/{actionId}`](https://docs.qwairy.co/developers/endpoints/actions/update-status) |
| `add_site_pages` | `write:site` | Manager | `never` | [`POST /brands/{brandId}/site-diagnostics/pages`](https://docs.qwairy.co/developers/endpoints/site-diagnostics/add-pages) |
| `start_site_audit` | `write:site` | Manager | `always` | [`POST /brands/{brandId}/site-diagnostics/audits`](https://docs.qwairy.co/developers/endpoints/site-diagnostics/start-audit) |

## update\_action\_status

Move an Action Center recommendation to another status.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `brandId` | string | Yes | Workspace ID |
| `actionId` | string | Yes | Action ID from `get_actions` |
| `status` | string | Yes | `SUGGESTED`, `PENDING`, `IN_PROGRESS`, `COMPLETED`, `DISMISSED` or `RESOLVED` |
| `expectedVersion` | integer | Yes | Action `version` you last read |
| `dryRun` | boolean | No | Preview without changing |
| `confirmationToken` | string | No | Token returned by the preview |

### Example

```text theme={null}
You: Mark the pricing page recommendation as in progress.

Assistant: reads get_actions, then calls update_action_status with
  {"brandId": "brand_01", "actionId": "action_01", "status": "IN_PROGRESS", "expectedVersion": 2}

Qwairy: updates the status and returns the action with its new version.
```

***

## add\_site\_pages

Add up to 50 URLs of the workspace domain to Site Diagnostics. Adding a page does not audit it.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `brandId` | string | Yes | Workspace ID |
| `urls` | string\[] | Yes | 1 to 50 URLs on the workspace domain, each up to 2048 characters |
| `dryRun` | boolean | No | Preview without adding |
| `confirmationToken` | string | No | Token returned by the preview |

URLs that are already monitored are reported as such rather than duplicated.

### Example

```text theme={null}
You: Monitor our pricing and features pages.

Assistant: calls add_site_pages with
  {"brandId": "brand_01", "urls": ["https://acme.example/pricing", "https://acme.example/features"]}

Qwairy: adds both pages and returns their page IDs. No audit has started.
```

***

## start\_site\_audit

Audit up to 50 selected pages, or all monitored pages with `all` (at most 10,000). The MCP call is free. As in the dashboard, each audited page uses 5 credits: the balance is checked when the audit is accepted, each page reserves its credits when its audit starts, and spends them when results are saved. `creditsReserved` is a budget, not an immediate hold or the final amount spent. Every execution needs a confirmation token, bound to the exact pages it previewed.

### Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `brandId` | string | Yes | Workspace ID |
| `pageIds` | string\[] | No | Up to 50 monitored page IDs. Cannot be combined with `all` |
| `all` | boolean | No | Audit every monitored page. Cannot be combined with `pageIds` |
| `dryRun` | boolean | No | Preview the pages and credits without starting |
| `confirmationToken` | string | No | Token returned by the preview |

Pages that are already being audited are skipped. An audit of more than 100 pages to audit is returned at once with its pages `QUEUED`; a background job running every five minutes queues them. Queueing is retried within 24 hours of acceptance, with at most 12 attempts, and pages the queue has not taken by then fail. Follow that same audit with [`get_site_audit`](https://docs.qwairy.co/mcp/tools/strategic#get_site_audit). To be notified by your own systems, subscribe to [`site_audit.completed`](https://docs.qwairy.co/developers/integration-webhooks/events#site_audit-completed) with the REST API.

An insufficient balance detected before acceptance prevents the audit from starting. If the balance is spent after acceptance, a credit refusal is returned only when no page started and the call queued the audit itself (100 pages to audit or fewer); it includes `details.auditId`, so read that retained audit. In every other case the audit is returned: the pages that could not start fail and the pages already started keep running. An accepted audit remains available if queueing is interrupted and resumes automatically with the same ID. `COMPLETED` means all pages have a final result, including failures and skipped pages; inspect the counts. A later audit cannot change these results.

### Example

```text theme={null}
You: Audit the two pages I just added.

Assistant: calls start_site_audit with
  {"brandId": "brand_01", "pageIds": ["page_01", "page_02"]}

Qwairy: returns a preview. Two pages would be audited, for 10 credits. It
returns a confirmationToken and starts nothing.

You: Confirm.

Assistant: repeats the call with the token, then polls get_site_audit until
the audit is COMPLETED.
```

## Related

* [Audit pages until site\_audit.completed](https://docs.qwairy.co/developers/guides/page-audit-site-diagnostics.md) for the REST workflow.
* [Strategic tools](https://docs.qwairy.co/mcp/tools/strategic.md) for the Action Center and Site Diagnostics read tools.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.