Cataloging & MARC remediation
Create and fix bibliographic records to standards.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
Previous class address: R10.06.01.01 · Original proposal IDs are retained.
- Typical applications
- FOLIO · OCLC Record Manager · MarcEdit
- Class GUI-need label
HY· Hybrid
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R10.06.01.01-T01 Fix nonfiling indicators on titles with initial articles in MarcEdit
Original instruction
input/new_titles.mrc has 20 bibliographic records. In MarcEdit's MarcEditor, set the 245 second indicator to the correct number of nonfiling characters for every title beginning with an English, French or German article (including the following space and any leading quotation mark), leave other titles at 0, and save output/new_titles_fixed.mrc.
- Platforms
win· Windowsmac· macOS- Applications
- MarcEdit
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
type· Text & formskey· Shortcuts- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
pymarc: 245 ind2 equals gold for all 20 records (seeded: 'L'', 'Die ', '"The '); no other field bytes changed (record diff limited to 245 indicators).
Proposed partial credit
80% indicators exact; 20% no collateral changes.
R10.06.01.01-T02 Catalog a new book from title-page images in FOLIO quickMARC
Original instruction
FOLIO sandbox tenant. From input/title_page.jpg, input/verso.jpg and input/book_facts.txt (pages, height, series), create a new MARC bibliographic record in quickMARC with 008 fixed fields, 020, 100 with relator, 245 with statement of responsibility, 250, 264 _1, 300, 336-338, 490 1 + 830, and two 650 headings from input/approved_lcsh.txt, then save and confirm the derived Inventory instance.
- Platforms
web· Any OS (web)- Applications
- FOLIO
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
type· Text & formsvis· Visual judgmentclick· Point & click- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
SRS record fields vs gold (subfield-level F1 >= 0.95; 008 dates/language/place exact; 245 ind2 correct); Inventory instance exists with matching title and contributors. Direct Okapi API calls from the agent VM are blocked by the sandbox proxy (only the FOLIO web UI origin is allowed), so the record must be created in quickMARC (audit).
Proposed partial credit
45% descriptive fields; 20% 008; 20% series and subjects; 15% instance created.
R10.06.01.01-T03 Batch-correct 008 dates and 264 consistency with a saved MarcEdit task
Original instruction
input/batch_2026_10.mrc has 50 records where 008/06-14 (date type and dates) disagree with 264 $c (including copyright dates and ranges). In MarcEdit build a saved task (output/fix_dates.task) that corrects 008 from 264 per input/date_rules.pdf, run it, hand-fix records the task cannot resolve, and save output/batch_fixed.mrc.
- Platforms
win· Windowsmac· macOS- Applications
- MarcEdit
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formscode· Code & commands- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
pymarc: 008/06-14 equals gold for all 50 records (6 seeded edge cases need manual fixes); 264 unchanged; task file present and, rerun by the grader on the original batch, fixes the 44 regular cases.
Proposed partial credit
55% 008 correct on all records; 25% task reproduces regular fixes; 20% edge cases.
R10.06.01.01-T04 Add holdings and items with call numbers for three instances in FOLIO
Original instruction
FOLIO sandbox, Inventory. For the 3 instances listed in input/new_copies.csv, create holdings at the given permanent locations with LC call numbers (call number type LC), then add items with barcodes, material type, permanent loan type, copy number and volume where given. One instance needs two holdings (Main Stacks and Reserve).
- Platforms
web· Any OS (web)- Applications
- FOLIO
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
type· Text & formsclick· Point & clicknav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Grader Okapi query: 4 holdings with location and call number exact; 6 items with barcode, material type, loan type, copy and volume exact and attached to the right holdings. Direct Okapi API calls from the agent VM are blocked by the sandbox proxy (only the FOLIO web UI origin is allowed), so actions must be logged from the UI.
Proposed partial credit
40% holdings; 50% items; 10% correct linkage.
R10.06.01.01-T05 Match five books to the right WorldCat records and set holdings
Original instruction
OCLC WorldShare Record Manager (sandbox institution). For the 5 books in input/books_in_hand.pdf (title-page photos and pagination), search WorldCat, choose the record matching the exact edition and manifestation (not the e-book or a different printing year), set your institution's holdings, and export each record to the local file output/matched.mrc.
- Platforms
web· Any OS (web)- Applications
- OCLC Record Manager
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & click- Work product
WP08· App/system state- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Sandbox holdings set on the 5 gold OCN records (distractor editions seeded); output/matched.mrc contains those 5 records (001 OCNs exact). WorldCat Metadata API credentials (WSKey) are withheld from the agent, so matching, holdings and export must be done in Record Manager from the browser session.
Proposed partial credit
70% correct manifestation chosen (pro rata per book); 15% holdings set; 15% export.
R10.06.01.01-T06 Remediate a legacy MARC batch and overlay it into FOLIO via Data Import
Original instruction
input/legacy_2009.mrc (1,200 records, MARC-8, obsolete 440 and 260 fields, duplicate 035s) must overlay existing FOLIO instances. Remediate the batch (UTF-8, 440 to 490/830, 260 to 264 _1, dedupe 035), build a FOLIO Data Import job profile that matches on 035 OCLC number with fallback to 001 and overlays the SRS record while keeping local 9XX fields, and run it on the test tenant.
- Platforms
web· Any OS (web)win· Windowslin· Linux- Applications
- FOLIO · MarcEdit · Python
- Interfaces
brow· Browserdesk· Desktop GUIterm· Terminal- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
HY· Hybrid- Interaction primitives
code· Code & commandsclick· Point & clickmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizonmultiapp· Multi-apprisk· Irreversible/risky- Proposed difficulty
- 5 / 5
Proposed verifier
Post-import SRS: 1,200 records overlaid with field-level checks (UTF-8 leader/09 'a', no 440/260, 035 unique); match profile fields exact; local 9XX preserved; 0 new duplicate instances; job log shows no errors. Remediated file also re-run against a hidden evaluator batch.
Proposed partial credit
35% remediation rules; 25% job profile; 25% overlay result without duplicates; 15% hidden-batch rerun.
R10.06.01.01-T07 Add an item to an existing bibliographic record in Koha
Original instruction
Koha staff client (sandbox). Add a new item to the bib record for 'The Tidewater Atlas (test)' (biblionumber in input/item.txt): barcode, home and current library, shelving location, collection code, call number and item type as given, and save.
- Platforms
web· Any OS (web)- Applications
- Koha
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
type· Text & formsclick· Point & click- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 1 / 5
Proposed verifier
items table: one new item on the biblionumber with all fields exact; no other items changed. Koha REST API credentials are withheld and the agent has no server shell, so the item must be added in the staff client (action_logs checked).
Proposed partial credit
80% item fields; 20% no collateral changes.
R10.06.01.01-T08 Validate and correct name headings against LC authorities in MarcEdit
Original instruction
input/faculty_pubs.mrc (60 records) uses variant name forms. Use MarcEdit's Validate Headings tool against id.loc.gov, then correct the 100/700 headings that do not match an authorized form to the authorized heading (keep relator terms), leave headings with no authority record unchanged, and export the change report to output/heading_report.txt.
- Platforms
win· Windowsmac· macOS- Applications
- MarcEdit · id.loc.gov
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scrolltype· Text & forms- Work product
WP02· Records/table- Conditions
multiapp· Multi-app- Proposed difficulty
- 3 / 5
Proposed verifier
100/700 $a$d strings equal gold authorized forms for the 15 seeded variants; relator $e preserved; unmatched names unchanged; report lists the corrections.
Proposed partial credit
65% headings corrected; 15% relators preserved; 20% unmatched untouched and report.
R10.06.01.01-T09 Convert thirty AACR2 records to RDA
Original instruction
input/aacr2_batch.mrc (30 records). Convert to RDA per input/local_rda_policy.pdf: run MarcEdit's RDA Helper, then manually fix what it misses: spell out abbreviations in 250 and 300, add 336-338, convert 260 to 264 with the right second indicator (including copyright dates as separate 264 _4), add relator terms, and remove GMD from 245 $h. Save output/rda_batch.mrc.
- Platforms
win· Windowsmac· macOS- Applications
- MarcEdit
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & clicktype· Text & formscode· Code & commands- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
pymarc field-level comparison with gold RDA records: subfield F1 >= 0.95 overall; every record has 33x; no 245 $h; 264 indicators correct including the 7 copyright-date cases; relator terms on 1XX/7XX.
Proposed partial credit
30% 264 conversion; 20% 33x; 20% abbreviations; 30% GMD removal and relator terms (half each).
R10.06.01.01-T10 Catalog a serial with a title change and linking entries
Original instruction
FOLIO sandbox, quickMARC. The journal 'Coastal Ecology Letters (test)' changed title in 2024 from 'Coastal Ecology Bulletin (test)'. Using input/issues_scans/, create the record for the later title: 008 for continuing resources, 022, 245, 264, 310/321 frequencies, 362 1, 780 00 to the earlier record, and add the reciprocal 785 00 to the existing earlier-title record.
- Platforms
web· Any OS (web)- Applications
- FOLIO
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
type· Text & formsvis· Visual judgmentclick· Point & click- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
New record fields vs gold (008 serial positions, 310/321, 362 exact); 780 on new and 785 on old record with correct titles, ISSNs and control-number links; Inventory preceding/succeeding titles linked. Direct Okapi API calls from the agent VM are blocked by the sandbox proxy (only the FOLIO web UI origin is allowed), so records must be edited in quickMARC.
Proposed partial credit
40% new serial fields; 35% linking entries both ways; 25% Inventory linkage.
R10.06.01.01-T11 Load vendor e-book records with proxy links and withdraw dropped titles
Original instruction
FOLIO sandbox. input/vendor_ebooks_oct.mrc has 400 new e-book records and input/withdrawn.txt lists 12 ISBNs dropped from the package. Prepare the file (prefix 856 $u with the proxy in input/proxy.txt, add 590 local note), import it with a FOLIO job profile that creates instances, holdings (location Online) and no items, then suppress from discovery and mark withdrawn the 12 dropped titles.
- Platforms
web· Any OS (web)win· Windowsmac· macOS- Applications
- FOLIO · MarcEdit
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
HY· Hybrid- Interaction primitives
click· Point & clickcode· Code & commandsmulti· Cross-app- Work product
WP02· Records/table- Conditions
multiapp· Multi-apprisk· Irreversible/risky- Proposed difficulty
- 4 / 5
Proposed verifier
400 instances with proxied 856 and 590; holdings at Online, 0 items; 12 dropped titles discovery-suppressed with status note (not deleted); no other records suppressed.
Proposed partial credit
30% file prep; 35% import results; 35% withdrawals handled without deletion.
R10.06.01.01-T12 Find and merge forty duplicate bibliographic records in Koha
Original instruction
Koha staff client (sandbox). A migration left duplicate bibs. Use reports to find the 40 duplicate pairs listed nowhere (same ISBN or same 245+264 $c), and for each merge into the fuller record (more subject headings, has 008 language) using Koha's merge tool, keeping all items and holds attached. Log merges in output/merge_log.csv.
- Platforms
web· Any OS (web)- Applications
- Koha
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scrollvis· Visual judgment- Work product
WP08· App/system state- Conditions
long· Long-horizonrisk· Irreversible/riskyhidden· Hidden/stale state- Proposed difficulty
- 5 / 5
Proposed verifier
Post-state: the 40 gold pairs merged with the gold survivor; item and hold counts per survivor equal the sum of the pair; 3 near-duplicate distractors (different editions) not merged; log rows match. Koha REST API credentials are withheld and the agent has no server shell, so merges must be done with the staff client's merge tool (action_logs checked).
Proposed partial credit
40% pairs found; 25% correct survivor; 25% items/holds preserved; 10% log.
R10.06.01.01-T13 Assign LC subject headings and call numbers for five new books
Original instruction
For 5 books with summaries and tables of contents in input/books/, choose up to 3 LCSH per book verified on id.loc.gov (with subdivisions as authorized) and an LC classification number with Cutter from the schedule extracts in input/lcc_extracts/, and record them in output/subjects.json with the id.loc.gov URIs.
- Platforms
web· Any OS (web)- Applications
- id.loc.gov · text tools
- Interfaces
brow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrolltype· Text & forms- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Each heading must resolve to an authorized LCSH URI (grader fetches snapshot); heading set overlap with expert gold Jaccard >= 0.5 per book; class number within the gold range in the extract; Cutter follows the table.
Proposed partial credit
30% valid authorized headings; 30% overlap with expert; 30% class numbers; 10% Cutters.
R10.06.01.01-T14 List records missing ISBNs with a pymarc script
Original instruction
Write a pymarc script that reads input/catalog_extract.mrc (5,000 records) and writes output/missing_isbn.csv with 001, 245 $a and 264/260 $c for every monograph record (leader/07 = m) lacking a valid 020 $a, treating 020 $z only as invalid.
- Platforms
lin· Linux- Applications
- Python · pymarc
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
CSV rows equal the gold set (by 001); serials excluded; records with only $z included; columns exact.
Proposed partial credit
80% correct set; 20% columns.
R10.06.01.01-T15 Write and apply an Alma normalization rule for incoming records
Original instruction
Ex Libris Alma (sandbox institution). Create a normalization rule that removes 9XX vendor fields except 949, adds 040 $d with your institution code, and converts 856 http to https; attach it to the import profile 'Vendor Shelf-Ready', run the import on input/shelf_ready.mrc, and confirm the results in the Metadata Editor.
- Platforms
web· Any OS (web)- Applications
- Alma
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
type· Text & formsclick· Point & clicknav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
licensed· Licensed pro apptutorial· Follow docs/tutorial- Proposed difficulty
- 4 / 5
Proposed verifier
Rule text present and attached to the profile; imported records: no 9XX except 949, 040 $d present, all 856 https; record count equals file count; no errors in the import job report. Alma API keys are withheld from the agent, so the rule, profile and import must be set up and run in the Alma UI (job report and history checked).
Proposed partial credit
35% rule logic; 15% attached to profile; 40% imported records correct; 10% clean job.
R10.06.01.01-T16 Derive a new-edition record in the OCLC Connexion client
Original instruction
OCLC Connexion client (Windows, sandbox). The library received the 3rd edition of a title whose 2nd edition is OCN in input/ocn.txt and no 3rd-edition record exists. Derive a new record from the 2nd edition, update 020, 250, 264, 300 and 504 from input/third_ed_scans/, remove edition-specific notes, validate, and save it to the local save file (do not add it to WorldCat).
- Platforms
win· Windows- Applications
- OCLC Connexion client
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G3· GUI-only- Interaction primitives
type· Text & formsclick· Point & clickkey· Shortcuts- Work product
WP02· Records/table- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Exported save-file record: changed fields equal gold; carried-over fields unchanged; no WorldCat 'replace' or 'add' transaction in the sandbox log; validation passes. Connexion macros (OML) are disabled and WorldCat Metadata API credentials are withheld, so the record must be derived and edited in the Connexion client.
Proposed partial credit
55% updated fields; 20% notes cleaned; 15% validation; 10% not added to WorldCat.
R10.06.01.01-T17 Turn a photograph collection spreadsheet into MARC and FOLIO records
Original instruction
input/photo_collection.xlsx describes 60 digitized photographs (title, creator, date ranges, places, subjects, dimensions, digital object URLs). Map it to MARC per input/mapping_spec.pdf (008 for visual materials, 245 devised titles in brackets, 264 _0, 300, 520, 651, 655 with LCGFT genre terms, 856 4 1), validate in MarcEdit, and import into FOLIO as a new collection with discovery suppression off.
- Platforms
win· Windowsweb· Any OS (web)- Applications
- MarcEdit · FOLIO · Excel
- Interfaces
desk· Desktop GUIbrow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
HY· Hybrid- Interaction primitives
code· Code & commandsclick· Point & clickmulti· Cross-app- Work product
WP02· Records/table- Conditions
multiapp· Multi-applong· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
FOLIO SRS records for 60 photos: field-level F1 >= 0.95 vs gold mapping; 008/06 date type for ranges correct; 655 terms resolve to LCGFT; 856 links valid; MarcEdit validation report clean.
Proposed partial credit
40% descriptive mapping; 20% dates/008; 20% genre and places; 20% loaded into FOLIO.
R10.06.01.01-T18 Link bibliographic headings to authority records in FOLIO quickMARC
Original instruction
FOLIO sandbox with MARC authority records loaded. For the 10 bibliographic records listed in input/bibs.txt, open each in quickMARC and link the 100, 600 and 650 headings to their authority records via the authority lookup, correcting headings that do not exactly match the authorized form, then save.
- Platforms
web· Any OS (web)- Applications
- FOLIO
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scrolltype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
SRS records: each target heading has $9 linking to the gold authority UUID and controlled subfields equal the authority 1XX; 4 seeded non-matching headings corrected; no headings linked to wrong authorities. Direct Okapi API calls from the agent VM are blocked by the sandbox proxy (only the FOLIO web UI origin is allowed), so links must be made in quickMARC.
Proposed partial credit
60% correct links; 25% corrected headings; 15% no wrong links.
R10.06.01.01-T19 Create the same holdings and items through the FOLIO API (control)
Original instruction
Control pair of T04. With Okapi credentials in env FOLIO_TENANT/FOLIO_TOKEN, create the holdings and items for the 3 instances in input/new_copies.csv through the holdings-storage and inventory items APIs, with the same locations, call numbers and item properties.
- Platforms
lin· Linux- Applications
- FOLIO Okapi API · Python
- Interfaces
api· API/SDKterm· Terminal- GUI need · task setting
AP· API-sufficient- GUI need · ordinary access
AP· API-sufficient- Interaction primitives
code· Code & commands- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Same Okapi checks as T04. Measures GUI-vs-API gap.
Proposed partial credit
40% holdings; 50% items; 10% linkage.
R10.06.01.01-T20 Repair diacritics and character encoding in twenty-five records
Original instruction
input/mixed_encoding.mrc has 25 records where some are MARC-8 with leader/09 wrongly set to 'a' and others have decomposed vs precomposed Unicode inconsistencies, so names with diacritics display incorrectly in FOLIO. Use MarcEdit's character conversion tools to produce clean UTF-8 (NFC) records with correct leader/09 in output/fixed_encoding.mrc, and check three names in the FOLIO preview.
- Platforms
win· Windowsmac· macOSweb· Any OS (web)- Applications
- MarcEdit · FOLIO
- Interfaces
desk· Desktop GUIbrow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
HY· Hybrid- Interaction primitives
click· Point & clickvis· Visual judgmentmulti· Cross-app- Work product
WP02· Records/table- Conditions
multiapp· Multi-app- Proposed difficulty
- 3 / 5
Proposed verifier
pymarc reads all 25 as UTF-8 with leader/09 'a'; all strings NFC; the 14 seeded diacritic names equal gold strings; no other text changed.
Proposed partial credit
40% conversion of MARC-8 records; 30% NFC normalization; 30% names exact and nothing else changed.
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.