Approvals

How AI-generated artifacts move through a human review step before being committed to your test suite.

What requires approval?

VigilQA keeps a human in the loop for any AI-generated artifact that will be executed or committed. The following items flow through the Approvals queue:

• Test case documents — generated by DocGen, reviewed by DocReview. Items that fail DocReview's criteria land in Approvals as rejected items with the rejection reason.
• Test scripts — generated by ScriptGen, validated by ScriptReview. All passing scripts land in Approvals awaiting sign-off before being added to the active test suite.
• LocatorHealer patches — selector replacements proposed after locator failures. Always require approval before being applied.
• Regeneration suggestions — when CoverageGap or ResultAnalyzer recommends regenerating scripts for specific scenarios.

Approvals workflow

Reviewing an item

Navigate to Approvals in the dashboard. Each pending item shows:

• The artifact type (script, document, or patch)
• The scenario ID and domain it belongs to
• A diff view (for patches) or a full preview (for new scripts and documents)
• Any agent-generated notes (e.g. confidence score for LocatorHealer patches, rejection reason for failed DocReview items)

Approving

Click Approve to commit the artifact. For test scripts, the script is immediately added to the active test suite and will be included in the next test run. For LocatorHealer patches, the selector replacement is applied to the existing script.

You can edit a script or document directly in the Approvals view before approving. Your edits are preserved; the item is marked as "manually reviewed and modified".

Rejecting

Click Reject and optionally add a comment explaining what needs to change. Rejected items are returned to the agent's queue for regeneration. For LocatorHealer patches, a rejection dismisses the patch and re-queues the failing test for manual investigation.

Bulk actions

Select multiple items using the checkboxes and use Bulk Approve or Bulk Reject. Bulk approvals are logged with the approving team member's identity for audit purposes.

Auto-approve settings

For mature projects where generated quality is consistently high, you can configure auto-approve rules in Settings → Approvals:

Setting	Description
Auto-approve scripts	Automatically approve ScriptReview-passing scripts without human review. Scripts that fail review still require manual approval.
Auto-approve documents	Automatically approve DocReview-passing documents, skipping the queue entirely.
Auto-approve LocatorHealer patches	Not recommended. LocatorHealer patches always default to requiring manual approval.
Minimum reviewer role	Set the minimum role required to approve items (Engineer or Admin).

Audit trail

Every approval and rejection is recorded in the audit log with:

• Timestamp
• Team member who took the action
• Whether the item was modified before approval
• The artifact version (hash) that was approved

The audit log is accessible from Settings → Audit Log and can be exported as CSV.

Notification settings

You can configure email or Slack notifications for Approvals events. See Notifications. Team members with the Engineer or Admin role can receive notifications when new items enter the queue or when items they submitted are rejected.