Confirm the inline [verify] markers against live standards before first real-collection use

Confirm the inline [verify] markers against live standards

The scaffold carries 40 inline [verify] markers across 8 files. They flag standards and API details that were not confirmable from the grounding sources when the toolkit was scaffolded. Each needs confirming against the live standard (id.loc.gov, the Getty vocabularies, FAST, VIAF, the PCC AI/MARC-588 guidance, and the relevant schema and profile docs) before Openlight describes real records in a client engagement.

This is not a blocker for the repository or for development. It is a gate before the toolkit is run on a real client collection (see the pilot plan).

Markers by file

CONTRIBUTING.md

• L146: confirmed, write the general shape and mark the endpoint [verify] for a maintainer to confirm

docs/authorities.md

• L72: autosuggest, and Wikidata's reconciliation interface each have their own contract. [verify] the
• L94: unambiguous authoritative match as exact and to send everything else to the cataloguer. [verify] the

docs/provenance-and-disclosure.md

• L79: adapts the exact wording to its own cataloging policy. [verify] Confirm current PCC practice for

prompts/fields/date.md

• L57: normalized, and add a [verify] note in basis so the cataloguer confirms the EDTF syntax. Do not
• L69: benchmark; BCE encoding left to [verify] rather than pinned.

skills/export-for-repository/SKILL.md

• L80: See ../../docs/collection-context.md. [verify] Confirm the MODS and
• L97: | Hyku and Samvera | Bulkrax CSV, or the Bulkrax API | Bulkrax is Samvera's ingest tool; the CSV column headers map to the work-type's properties. Disclosure...
• L98: | DSpace | Simple Archive Format (SAF), or a metadata CSV | SAF packages each item as a directory with a dublin_core.xml and a contents manifest; the CSV p...
• L99: | Islandora and Fedora | MODS XML, or RDF | Islandora ingests MODS; Fedora-native and linked-data deployments take RDF. Disclosure note maps to a MODS note...
• L100: | ArchivesSpace | EAD XML, or the ArchivesSpace API | The arrangement rides the finding-aid hierarchy EAD already models, so collection/series/`is_part_o...
• L101: | General | Dublin Core XML, MODS, or a IIIF Presentation manifest | The platform-neutral default. Dublin Core XML is the safe baseline; MODS for hierarchica...

skills/preprocess-derivatives/SKILL.md

• L65: statement of how the pages relate. [verify] the exact manifest fields read against the
• L98: named PBCore technical fields requires a profile or schema extension. [verify] against the

skills/resolve-authorities/SKILL.md

• L57: linked-data interfaces) has its own contract. Before relying on any of them, [verify] the current
• L76: ranked or fuzzy match above a threshold). [verify] the specific scoring rule and any similarity
• L131: endpoint [verify]ed against its published documentation), adding the vocabulary to

vocab/README.md

• L18: it is marked [verify]. Do not reconstruct an endpoint, an element name, or a rights URI from memory;
• L44: | VIAF | Aggregated international name authority; a cross-reference and a fallback for names not yet in LCNAF | https://viaf.org/ [verify] (grounded in a...
• L47: | GeoNames | Geographic authority | https://www.geonames.org/ [verify] (named in the standards-reference prose, not in its endpoints table) |
• L52: and Wikidata linked-data interfaces, the GeoNames API) has its own contract. [verify] the current
• L85: profile are marked [verify]; a clean-looking but subtly wrong cell is worse than a flagged one.
• L94: | title | dc:title / dcterms:title | <titleInfo><title> | work > title | pbcoreTitle [verify] |
• L95: | description | dc:description / dcterms:description | <abstract> or <note> [verify] | work > description | pbcoreDescription [verify] |
• L96: | creator | dc:creator / dcterms:creator | <name> with <role> creator [verify] | work > agent (role: creator) [verify] | pbcoreCreator `[ve...
• L97: | date | dc:date / dcterms:created [verify] | <originInfo><dateCreated> [verify] | work > date (type: creation) [verify] | pbcoreAssetDate ...
• L98: | type | dc:type / dcterms:type | <typeOfResource> | work > worktype [verify] | pbcoreAssetType [verify] |
• L99: | genre_form | dc:type (refined via the genre/form vocabulary) | <genre> | work > worktype or work > stylePeriod [verify] | pbcoreGenre `[verif...
• L100: | subjects | dc:subject / dcterms:subject | <subject><topic> | work > subject | pbcoreSubject [verify] |
• L101: | depicted_entities | dc:subject (a depicted person, place, object, or event is a subject) | <subject> with <name>, <geographic>, or <topic> by e...
• L102: | spatial | dcterms:spatial | <subject><geographic> or <originInfo><place> [verify] | work > location [verify] | pbcoreCoverage (type: spatia...
• L103: | medium_technique | dcterms:medium | <physicalDescription><form> or <note> [verify] | work > material and work > technique [verify] | N/A `[...
• L104: | language | dc:language / dcterms:language | <language><languageTerm> | work > language [verify] | pbcoreLanguage [verify] (often on an inst...
• L114: <topic>. [verify] the exact MODS element and attribute names against the target MODS profile.
• L118: the Image record and are out of scope for the descriptive crosswalk. [verify] VRA Core element
• L122: equivalent (PBCore models format and instantiation separately) and is marked N/A. [verify] every
• L141: export-stage concern; [verify] the destination field per platform against the
• L146: URIs carry version suffixes that change and are not in the grounding sources; [verify] each against
• L161: [verify] the current statement list and each statement's versioned URI at https://rightsstatements.org/.
• L169: [verify] the license identifiers, version numbers, and canonical URIs at https://creativecommons.org/.

Start by reading the 40 inline markers across CONTRIBUTING.md, docs/, prompts/fields/date.md, skills/, and vocab/README.md, then consult the live standards named in the issue. Confirm each standards, API, schema, profile, and URI detail against its source and update the affected documentation or marker. Done means every marker has been reviewed and the repository no longer relies on unconfirmed claims before pilot use.