Visual Element Audit

Run a visual element audit that inventories every table, figure, and diagram, then checks labeling, narrative references, source documentation, accuracy against source data, and accessibility. The output is a structured YAML report with per-element issues and remediation recommendations.

It checks whether each element has alt text for screen readers, uses colorblind-safe encoding rather than color-only distinctions, maintains adequate color contrast, and keeps text at a readable size. Each failure is flagged with severity and a remediation step.

The scripts/validate_output.py script parses the YAML output and checks for required sections: visual_inventory, labeling_audit, reference_audit, source_verification, and accessibility_evaluation. It reports structural errors such as missing element_id fields or wrong types.

Orphaned figures are visual elements present in a document but never referenced or explained in the narrative text. The audit flags them as high-severity issues, recommending either adding a narrative reference or removing the element.

The audit requires the complete integrated document with all embedded visuals, source data for tables and figures, captions and labels, narrative references, visual specifications such as axes and legends, and the applicable design style guide.