What is internal editor-state data?
Some editing workflows attach state, identifiers, or other technical information to HTML elements. When that data remains in a saved description, it can be separate from the text and structure the merchant intended to publish. This check treats that case as editor-state metadata and requires review.
MDN documents data-* as a generic mechanism for proprietary information exchanged between HTML and the DOM. That does not make every data-* attribute wrong, and it does not let PROVEN identify an editor from an attribute. The detector stays within its documented classification of editorial state and preserves context for review.
Examples of metadata that should not be assumed useless
These fictional examples use generic names and do not represent real merchant data.
<span data-editor-state="…">Product text</span>Keep the text; review only attached informationThe attribute name is illustrative. A real attribute can serve another purpose, so removal follows context review.
An internal identifier remains on a wrapperCompare HTML before and after a controlled copyIf storefront code or an integration depends on it, removal is not safe merely because shoppers do not see the value.
Clean product copy remains inside an element carrying unused editor state[candidate for cleanup]Cleanup should target metadata rather than shopper-facing text or useful structure.
What does PROVEN check?
Checks for internal editor-state information left inside the description.
Why does it matter?
This data is not part of the product content and can retain unnecessary editorial information.
Why does it matter for a catalog?
Editorial metadata can widen the gap between the content a merchant thinks they maintain and the HTML actually stored. Over time, those fragments can make debugging, migration, or description comparison more difficult.
PROVEN does not assume internal editor data is dangerous and does not attribute it to an application without separate evidence. The maintenance goal is to distinguish product content from state that appears to belong to the editing process.
What can PROVEN do?
PROVEN can flag the metadata and its context; removal requires your review.
How PROVEN reviews editor state
Metadata is treated as contextual evidence, not as product copy and not as sufficient reason for automatic deletion.
- Flag the metadataPROVEN shows the fragment classified by this check as editorial-state information and keeps the HTML context where it occurs.
- Check dependenciesThe merchant checks whether a theme, app, or editorial routine uses the information before accepting removal.
- Prepare reviewed changeWhen metadata is confirmed as residue, cleanup can be prepared without altering product copy; the result remains review-gated.
Why is the attribute name not enough?
HTML is extensible and similar attribute forms can serve different purposes. A conclusion based only on a name would confuse a technical mechanism with an integration intent.
Review must check the concrete effect and preserve useful information. PROVEN does not promise that all flagged metadata can be removed with the same operation.
Interpretation limits
- Not every data-* or custom attribute is internal editor state.
- The detector does not automatically identify the application or editor that introduced metadata.
- Metadata presence is not a security or compromise verdict.
- Removal must not damage useful product copy, structure, or integration behavior.
- This page does not promise SEO or commercial effects from cleanup.
Frequently asked questions
How is metadata different from product copy?
Product copy and useful structure are shopper-facing content; editorial metadata describes technical state or editing-workflow information that must be evaluated separately.
Is every data-* attribute a problem?
No. MDN defines data-* as a legitimate custom-data mechanism. This detector must not be generalized to every such attribute.
Can PROVEN tell which application inserted the value?
Not from metadata presence alone. Source attribution would require separate evidence outside this finding.
Why is cleanup review-gated?
Because invisible metadata can still be used by an integration. Review checks dependencies before removal.
Sources and independent context
These sources explain HTML editing in Shopify and the general role of custom attributes and provide context for review.
- Shopify — Using the rich text editorShopify documents HTML editing and troubleshooting transferred formatting in rich-text content.
- MDN — data-* attributesMDN describes data-* as custom attributes for proprietary information accessible through the DOM.