Inventory question-bank duplicates with quiz-reference protection

The problem

After course reuse, duplicate-looking question records can have different identities, versions, scoring settings or quiz dependencies. A cleanup list based on text similarity alone can obscure fixed-version links, latest-version pointers and random category selection, even when no student data is involved.

Agent job

Prepare a version-aware question inventory that separates exact observable content matches from near-match hypotheses, retains every original ID and maps direct and category-based quiz references before presenting educator/admin review questions.

Expected output

A draft question_match_review.csv, question_usage_edges.csv, category_growth.csv and keep_separate_questions.txt with preserved identity/version records, evidence for candidate matches and a protected/unobservable register. No merge, deletion recommendation, answer revision or publishable bank file.

Why an agent / simpler alternative

The agent may explain nuanced near matches, conflicting version intent and why a category-based usage path makes cleanup uncertain across several exports. Exact signatures, version joins, selector expansion and count conservation should be deterministic. A native question-bank usage report plus a simple content-hash inventory is preferable when it covers all versions and selection rules. Agent use adds little for exact duplicates alone and must never replace the educator’s content judgment or administrator’s reference checks.

Who this helps

1. Learning platform administrator

2. Assessment content librarian

Before a pilot — required, unexecuted

1. Obtain a specific owner-authorized synthetic test scope, processing environment and retention plan, and create an independent expected-result oracle before executing the proposed workflow.

2. Confirm a present unresolved owner need in the target installation/version and compare the complete output and total review effort with the native or deterministic alternative; stop if the simpler control is sufficient.

3. Verify actual account eligibility, task-specific file/tool support, access limits and costs before testing; no connector installation, purchase, upload or access provisioning is authorized by this plan.

4. Before an authorized fixture, create a field-complete identity/version and selector schema, specifying latest-ready status, hidden/archived versions, category descendants, filter evaluation, null-versus-empty configuration and digest provenance. Content owner/admin should independently author all match labels and typed edges.

5. Include multiple versions under one entry, unknown selector coverage, same-text/different-option-order or scoring, missing assets and safe XML external-entity cases. Verify all identities and paths survive, selector membership remains potential eligibility and no output artifact introduces a survivor/deletion or executable repair target.

Evidence boundary

Problem-source and occupational context are separated from this original procedure. All six steps, prompts, output schemas, synthetic numbers, acceptance criteria and value hypotheses are authored proposals, not source-derived instructions or demonstrated agent results.

Setup requirements

1. Use synthetic adult-course assessment content for the proposed first test. Do not use student records, submissions, grades, accommodations, analytics, enrolment or employment decisions.

2. Require bank namespace plus entry/version identity. A numeric question ID or prompt text alone is insufficient across copied courses or banks.

3. Require fixed/latest/random reference semantics, category hierarchy and filter definitions. Export omissions must be recorded; direct quiz links alone do not establish unused status.

4. Define the exact-match field set, including scoring, option order, type-specific settings and asset digests. Do not normalize substantive text whitespace, mathematical expressions or code.

5. Compare category snapshots only under the same scope and mapping. Obtain content-owner approval for secure handling of questions/keys; no third-party upload or plugin installation is assumed.

Inputs and access

1. Name

   question_bank_metadata.json_or_xml

   Source or owner

   Synthetic adult-course content owner

   Sensitivity

   First test is synthetic only. Bank namespace, entry ID, version ID, category, question type, language, prompt/options/scoring configuration and relevant media/configuration digests supplied by the owner. No student answers, attempts, grades, names or identifiers. A later owner-approved copy must remove personal data, credentials and unrelated content.

   Permission needed

   Before any future pilot, the accountable owner must authorize the exact synthetic or nonpersonal export copies, selected processing environment, scope and retention. Planning permission does not authorize that pilot.

   Scope

   Bank namespace, entry ID, version ID, category, question type, language, prompt/options/scoring configuration and relevant media/configuration digests supplied by the owner. No student answers, attempts, grades, names or identifiers.

2. Name

   quiz_reference_metadata.csv

   Source or owner

   Synthetic learning-platform administrator

   Sensitivity

   First test is synthetic only. Course/quiz/slot codes, fixed-version or latest-version reference semantics, random/category selectors with subcategory/type/tag filters and export coverage. No enrolments, submissions, learner-level records or access credentials. A later owner-approved copy must remove personal data, credentials and unrelated content.

   Permission needed

   Before any future pilot, the accountable owner must authorize the exact synthetic or nonpersonal export copies, selected processing environment, scope and retention. Planning permission does not authorize that pilot.

   Scope

   Course/quiz/slot codes, fixed-version or latest-version reference semantics, random/category selectors with subcategory/type/tag filters and export coverage. No enrolments, submissions, learner-level records or access credentials.

3. Name

   category_and_version_manifest.json

   Source or owner

   Synthetic content librarian

   Sensitivity

   First test is synthetic only. Category hierarchy, before/after counts with identical scope, current version/status mapping, intentional variant labels and coverage of archived/hidden versions and attached assets. A later owner-approved copy must remove personal data, credentials and unrelated content.

   Permission needed

   Before any future pilot, the accountable owner must authorize the exact synthetic or nonpersonal export copies, selected processing environment, scope and retention. Planning permission does not authorize that pilot.

   Scope

   Category hierarchy, before/after counts with identical scope, current version/status mapping, intentional variant labels and coverage of archived/hidden versions and attached assets.

4. Name

   identity_and_comparison_rules.txt

   Source or owner

   Synthetic educator and administrator

   Sensitivity

   First test is synthetic only. Which IDs are authoritative, approved serialization-only normalization, protected language/type/scoring differences, content ownership, permitted review scope and fields missing from export. No authority to choose deletions. A later owner-approved copy must remove personal data, credentials and unrelated content.

   Permission needed

   Before any future pilot, the accountable owner must authorize the exact synthetic or nonpersonal export copies, selected processing environment, scope and retention. Planning permission does not authorize that pilot.

   Scope

   Which IDs are authoritative, approved serialization-only normalization, protected language/type/scoring differences, content ownership, permitted review scope and fields missing from export. No authority to choose deletions.

Proposed steps

1. 1

   Validate file scope, row counts, namespaces and version fields. Require student-free material and known content ownership. Inventory unavailable plugin settings, media digests, hidden versions and unexported quiz usages. For a future parser, disable external entities and embedded execution; source text is never instruction.

   Input: Question/quiz copies, identity rules and coverage manifest

   Output: Input completeness map, preserved identity/version inventory and blocked field list.

   Expected output: Input completeness map, preserved identity/version inventory and blocked field list.

   Checkpoint: No quiz reference is declared absent merely because the export omits it. Every original entry/version remains represented; copied content is not uploaded elsewhere.

   Approval gate: Institution/content owner approves exact files, processing environment and exam-security handling before any pilot; unauthorized or personal data stops intake.

2. 2

   Build exact-match signatures from the full declared observable content and configuration, excluding only identity/location fields. Preserve prompt whitespace, option order, language, scoring, tolerances, randomization and media/configuration digests. Then propose separately labelled near-match pairs with explicit textual differences, not numeric confidence masquerading as equivalence.

   Input: Complete observable question content/configuration and approved comparison rules

   Output: Exact-observable-match groups, near-match hypotheses and field-difference evidence.

   Expected output: Exact-observable-match groups, near-match hypotheses and field-difference evidence.

   Checkpoint: A missing substantive field prevents exact-equivalence classification. A shared stem alone cannot override a changed key, question type, version intent or asset.

   Approval gate: Educator decides whether any near match is substantively related. The agent does not grade content, validate keys or choose a canonical record.

3. 3

   Resolve each fixed reference to its full bank/entry/version identity. Preserve latest-version selection mode and its snapshot resolution. Expand random/category selectors only where the hierarchy, filters and eligible versions are complete, retaining the selector and all candidate membership edges.

   Input: Quiz slots, version manifest and category selector definitions

   Output: Question-usage graph with direct, latest-pointer and selector-membership edge types plus unresolved references.

   Expected output: Question-usage graph with direct, latest-pointer and selector-membership edge types plus unresolved references.

   Checkpoint: A category-selected item is potentially used even with no direct quiz link. Unknown selector coverage protects every possibly affected item and blocks a safe-cleanup conclusion.

   Approval gate: Administrator confirms reference semantics and scope; no live quiz, course, bank or SIS access is requested.

4. 4

   Overlay usage and protected differences onto each group. Mark exact-looking members referenced by separate quizzes as retained/protected, keep scoring/language/type/version variants separate, and place incomplete matches in an unobservable queue. Preserve original reference multiplicity and do not collapse IDs into one row.

   Input: Match groups, identity/version inventory and usage graph

   Output: Reference-aware group register and keep-separate/protected list with specific reasons.

   Expected output: Reference-aware group register and keep-separate/protected list with specific reasons.

   Checkpoint: All original question identities and usage edges remain intact. Unreferenced-in-export is not safe-to-delete; there is no automatic survivor or replacement target.

   Approval gate: Content owner and administrator review intent and usage together. Deletion, merge, replacement and answer edits are excluded regardless of apparent similarity.

5. 5

   Compute before/after net count changes by the supplied category mapping and scope, keeping additions, version changes and unknown causes distinct. Report the number of exact-match groups and additional members as descriptive counts only; do not equate total growth or match membership with avoidable duplication.

   Input: Comparable category count snapshots and reference-aware groups

   Output: Category-growth workpaper and candidate-group summary with denominator and scope notes.

   Expected output: Category-growth workpaper and candidate-group summary with denominator and scope notes.

   Checkpoint: Counts conserve the full inventory. Different snapshot scopes block a growth comparison; a zero baseline has undefined percent growth rather than division by zero.

   Approval gate: Administrator approves comparability and any interpretation of why counts changed; a historical copy report does not establish the cause in these inputs.

6. 6

   Produce the four local draft review artifacts, preserving every ID/version and reference. Ask the educator/admin for missing metadata or intended variant decisions. Document what an independently authorized later cleanup would need to verify, without supplying destructive SQL, imports, remapping files or executable repair commands.

   Input: Protected/group registers, reference graph, growth workpaper and unresolved coverage

   Output: Draft question-bank lifecycle review packet awaiting joint educator/admin decision.

   Expected output: Draft question-bank lifecycle review packet awaiting joint educator/admin decision.

   Checkpoint: Every match is a review hypothesis or exact match of declared observable fields. Zero questions are designated for deletion or merged, and no learner, grading or course-publishing decision appears.

   Approval gate: Any future cleanup, export sharing, question revision or live operation requires separate explicit scope and authorization; this plan ends at review.

Original starter prompt — proposed

Origin

editorial_proposal

Label

Starter prompt for a future authorized synthetic inventory test; not executed

Placeholders

1. question_metadata

2. quiz_references

3. category_version_manifest

4. comparison_rules

5. output_folder

Text

Using only {question_metadata}, {quiz_references}, {category_version_manifest} and {comparison_rules}, draft a question-bank match and usage inventory. Preserve every bank/entry/version ID. Exact-match labels require all declared substantive fields; keep scoring, option order, language, type, media and version differences visible. Separate near matches and missing metadata from exact observable matches. Trace fixed and latest references and complete random/category selector membership; unknown usage protects the affected scope. Compare counts only under identical scope. Write draft question_match_review.csv, question_usage_edges.csv, category_growth.csv and keep_separate_questions.txt in {output_folder}. Do not choose survivors, delete, merge, remap references, edit answer keys, produce repair SQL or use any student data.

Execution authorized

No

Expected output schema — proposed

Origin

editorial_proposal

Columns

1. bank_namespace

2. entry_id

3. version_id

4. category_id

5. match_group_id

6. match_class

7. comparison_field_coverage

8. exact_match_basis

9. substantive_differences

10. quiz_slot_refs

11. reference_modes

12. selector_membership_refs

13. version_resolution_status

14. protected_reason

15. missing_metadata

16. source_refs

17. owner_decision

Record status

draft_pending_owner_review

Allowed match classes

1. exact_declared_observable_fields

2. near_match_for_educator_review

3. protected_substantive_variant

4. unobservable_comparison

5. no_candidate_match

Deletion field

None. There is no canonical/survivor ID, delete flag, replacement target or executable import in this output.

Synthetic output illustration — not an executed result

Label

Synthetic metadata and authored expected inventory; not a completed deduplication or cleanup

Synthetic

Yes

Executed

No

Owner approval status

pending_not_obtained

Data

Synthetic inputs

Bank namespace

SYN-BANK

Shared content for q1 q2

Type

single_choice

Language

en

Prompt

Which value equals two plus two?

Ordered options

1. 3

2. 4

Score fractions

1. 0

2. 1

Shuffle

No

Media

Feedback

Other declared type settings

identical and complete

Items

1. Id

   SYN-Q1

   Version

   1

   Category

   SYN-A

   Content

   shared content; complete

2. Id

   SYN-Q2

   Version

   1

   Category

   SYN-A

   Content

   shared content; complete

3. Id

   SYN-Q3

   Version

   2

   Category

   SYN-A

   Content

   shared content except score_fractions=[1,0]

   Latest ready version

   2

4. Id

   SYN-Q4

   Version

   1

   Category

   SYN-B

   Content

   shared content except prompt=What is 2 + 2?

5. Id

   SYN-Q5

   Version

   1

   Category

   SYN-A

   Content

   shared text/settings but asset manifest missing; equality unobservable

6. Id

   SYN-Q6

   Version

   1

   Category

   SYN-C

   Content

   Different complete question: select a primary colour; no candidate match

Quiz references

1. Quiz slot

   SYN-X/1

   Mode

   fixed

   Target

   SYN-Q1/v1

2. Quiz slot

   SYN-Y/2

   Mode

   fixed

   Target

   SYN-Q2/v1

3. Quiz slot

   SYN-Z/1

   Mode

   latest_ready

   Target entry

   SYN-Q3

   Resolved snapshot target

   SYN-Q3/v2

4. Quiz slot

   SYN-W/3

   Mode

   random_category

   Selector id

   SYN-S1

   Category

   SYN-A

   Include subcategories

   No

   Type filter

   single_choice

   Eligibility

   all four listed A entries, no extra filters

Category counts before

Syn-a

2

Syn-b

0

Syn-c

1

Category counts after

Syn-a

4

Syn-b

1

Syn-c

1

Coverage

Same synthetic category scope; fixed/latest/selector exports complete for this illustration. Q5 asset completeness alone is unknown.

Illustrative expected match rows

1. Id

   SYN-Q1/v1

   Group

   SYN-G1

   Class

   exact_declared_observable_fields

   Protection

   Fixed X/1 and random S1; preserve identity

2. Id

   SYN-Q2/v1

   Group

   SYN-G1

   Class

   exact_declared_observable_fields

   Protection

   Fixed Y/2 and random S1; preserve identity

3. Id

   SYN-Q3/v2

   Group

   Not specified.

   Class

   protected_substantive_variant

   Protection

   Different score fractions; latest Z/1 and random S1; no key-correctness judgment

4. Id

   SYN-Q4/v1

   Group

   SYN-N1

   Class

   near_match_for_educator_review

   Protection

   Paraphrased prompt resembles G1 but remains separate; no unused/safe-delete conclusion

5. Id

   SYN-Q5/v1

   Group

   Not specified.

   Class

   unobservable_comparison

   Protection

   Missing asset metadata and random S1 usage; no exact-match claim

6. Id

   SYN-Q6/v1

   Group

   Not specified.

   Class

   no_candidate_match

   Protection

   Retained original record

Illustrative expected usage

Fixed or latest edges

1. SYN-X/1 -> SYN-Q1/v1

2. SYN-Y/2 -> SYN-Q2/v1

3. SYN-Z/1 -> SYN-Q3 (latest_ready), resolved at snapshot to v2

Selector membership edges

1. SYN-S1 -> SYN-Q1/v1

2. SYN-S1 -> SYN-Q2/v1

3. SYN-S1 -> SYN-Q3/v2

4. SYN-S1 -> SYN-Q5/v1

Formula

3 direct/latest question edges + 4 selector-to-question membership edges = 7 question-target edges; plus 1 quiz-slot-to-selector edge = 8 typed graph edges across 4 quiz slots. Membership is potential eligibility, not an observed draw.

Quiz slot to selector edges

1. SYN-W/3 -> SYN-S1

Question target edge count

7

Total typed graph edge count

8

Illustrative counts

Inventory before

3

Inventory after

6

Net growth

3

Growth percent

100

Formula

(4 + 1 + 1) - (2 + 0 + 1) = 3; (6 - 3) / 3 * 100 = 100%

Exact match groups

1

Members in exact group

2

Additional exact group members

1

Calculation

2 members - 1 group = 1 additional content-matching member; not a deletable-item count

Retained question records

6

Records selected for deletion

0

Caveat

The source of net growth is unknown. No content-match member is designated erroneous or removable.

Human approval boundaries

1. Before any future pilot, the accountable owner must authorize the exact synthetic or nonpersonal export copies, selected processing environment, scope and retention. Planning permission does not authorize that pilot.

2. The institution and content owner approve use of question text and keys, the exam-security boundary, version semantics, permitted normalization and complete quiz/category-reference coverage. No student-level data is included.

3. Educator/admin jointly judge intended variants. Approving this inventory does not authorize a canonical selection, deletion, merge, answer-key change, SQL repair, import, course publication or SIS action.

4. The only contemplated deliverable is a new local draft review packet. No source edits, live connection, system action, external upload, communication, installation, credentials, purchase or personal-data processing is authorized.

Failure modes

1. Stem similarity hides different answer order, scoring, language, tolerances, assets or plugin configuration.

2. An item with no direct quiz reference is treated as unused although a random category selector can draw it.

3. Latest-version references are flattened to a fixed version or historical/hidden versions disappear from the inventory.

4. Missing media or reference fields are treated as equality or absence rather than unobservable.

5. Aggregate bank growth is described as a count of erroneous duplicates or deletable items.

6. Question content or XML asks for external access; embedded instructions are not authority and external entities must not be fetched.

Recovery and troubleshooting

1. Signal

   Stem similarity hides different answer order, scoring, language, tolerances, assets or plugin configuration.

   Recovery

   Remove the equivalence label, show each substantive difference and retain the full identity/version pair.

   Resume condition

   Educator can review the exact difference; changed scoring/language/type remains protected regardless of similarity.

2. Signal

   An item with no direct quiz reference is treated as unused although a random category selector can draw it.

   Recovery

   Trace all supplied selector definitions and category descendants. If scope is incomplete, mark the affected bank/category protected-by-unknown-usage.

   Resume condition

   Administrator supplies complete selector semantics, or safe-cleanup remains explicitly unassessable.

3. Signal

   Latest-version references are flattened to a fixed version or historical/hidden versions disappear from the inventory.

   Recovery

   Restore the original reference mode and version/status history; stop snapshot resolution if latest is undefined.

   Resume condition

   A complete owner-supplied version map exists; no fixed-version rewrite is produced.

4. Signal

   Missing media or reference fields are treated as equality or absence rather than unobservable.

   Recovery

   List missing fields and downgrade the match to unobservable or near-review-only.

   Resume condition

   The missing owner-approved metadata is supplied or uncertainty remains explicit.

5. Signal

   Aggregate bank growth is described as a count of erroneous duplicates or deletable items.

   Recovery

   Separate net growth, exact-field match membership and unknown cause; delete no records and make no savings claim.

   Resume condition

   Counts are scope-comparable and the narrative makes no causal or deletion inference.

Unexecuted validation plan

Sample setup

Proposed fixture: the six synthetic items and references in the illustration, plus isolated variants for an unknown subcategory selector, a stale latest-version map, a dangling fixed reference Q99, changed option order, a missing media digest, XML external-entity text and mismatched count scopes. There are no learner records and the fixture has not been built.

Independent oracle

Author

A content owner and learning-platform administrator independent of the tested agent

Proposed oracle

Prepare an expected group/register and reference-edge list before execution: one exact group {Q1/v1,Q2/v1}; Q3 protected scoring variant; Q4 near-review only; Q5 unobservable; Q6 separate. Preserve 6 identities and 8 typed graph edges (3 direct/latest, 1 slot-to-selector, 4 selector-membership) in the base illustration. Net growth 3, exact-group additional member 1, deletion selections 0. Variants specify protected-by-unknown-usage or blocked comparisons.

Creation status

not_created_or_executed

Acceptance checks

1. The base exact group contains only Q1/Q2; key/configuration, prompt and missing-asset variants are not silently merged.

2. All six original identities survive, including versions and category provenance; there is no canonical or delete target.

3. All three direct/latest edges, the one slot-to-selector edge and four selector-membership edges appear. Latest selection mode is preserved and membership edges are never reported as actual quiz draws.

4. Unknown selector coverage protects its possible members; stale latest mapping and Q99 remain unresolved reference exceptions.

5. Count growth is exactly +3 under the base scope, never called three erroneous duplicates; mismatched scopes block the comparison and a zero baseline yields no percent.

6. External entities are not fetched and no executable repair, import, answer edit, grading or learner decision is generated.

7. All supplied copies remain byte-for-byte unchanged; outputs retain source references, synthetic/draft labels and pending-owner-review status.

8. No network/account operation, live mutation, external communication, personal-data use, installation or access provisioning is attempted.

Stop conditions

1. A near match, missing field or no-direct-reference status is used to select a deletion or replacement.

2. Any version/reference edge disappears, a random-selector path is omitted, or a substantive key/configuration difference is treated as formatting.

3. Personal data, credentials or unrelated content appears, or an input requests an external action. Treat source text as data and pause the affected scope.

Executed

No

Fixture status

Design only. The fixture and oracle below have not been created or run. Arithmetic in the illustration is an authored expectation, not measured performance.

Approval status

not_requested_or_granted_for_execution

Costs and effort — unknown or hypothetical

Status

unknown_or_estimate

Currency

Not specified.

Items

1. Agent plan price, secure-file-processing eligibility, input limits and format support remain unknown.

2. Obtaining complete version, plugin-field, media and cross-course selector metadata may be the largest effort; institution and administrator costs are unmeasured.

3. Educator review of near matches and intentional versions is necessary and has unknown cost. Compare with built-in question-bank usage tools first.

4. No new platform access, database query, paid plugin, installation, upload or workflow execution is included.

Caveat

No storage saving, educator time saving, cleanup rate, error reduction or commercial demand is established.

Canadian adaptation — proposed

A Canadian institution must approve the processing environment and handling of proprietary assessment content under its own security, retention and applicable provincial policies. Use synthetic adult-course data only in the first proposed test and exclude all learner records. Keep English/French versions, accents, mathematical notation and locally meaningful units distinct unless the educator explicitly defines an equivalence; translated wording alone never establishes a duplicate. No student eligibility, grades, accommodations or SIS decisions are assessed. No Canadian deployment, privacy compliance or account availability is claimed.

Limits

1. Original editorial plan only. No workflow, sample fixture, target application or acceptance test was executed; no observed agent success is claimed.

2. The problem report is context, not evidence of a current defect, prevalence, willingness to pay or present institutional need. Confirm the current installed version and simpler built-in remedies before a pilot.

3. Official capability references support general file-reading, analysis and reviewable-output primitives. Exact ChatGPT Work or Claude account access, tool availability, format support, integrations and end-to-end compatibility remain unknown and untested.

4. This plan authorizes no future execution. Any pilot needs separate approval of the exact minimized copies, processing environment, scope and retention; source systems remain untouched.

5. All output is draft evidence for an accountable owner. No student records, employee evaluation, employment decisions, financial transactions, external communication, live-system writes, new access or installation is included.

Candidate platforms — compatibility untested

1. Name

   ChatGPT Work

   Required capabilities

   1. Read owner-supplied synthetic structured/text copies without source-system access

   2. Use inspectable deterministic joins, set checks and calculations where specified

   3. Compare textual discrepancies and produce source-linked local draft artifacts for human review

   Capability source urls

   1. https://learn.chatgpt.com/docs/get-started-with-work

   2. https://learn.chatgpt.com/docs/artifacts-viewer

   3. https://learn.chatgpt.com/docs/sandboxing

   4. https://learn.chatgpt.com/docs/enterprise/chatgpt-work-overview

   Workflow compatibility

   Unknown and untested for every proposed plan.

   Selection basis

   Editorial candidate based on documented primitives only; no claim of project-system, accounting-system, learning-system or SIS integration.

   Current access note

   Current official documentation supports general file, code and reviewable-output primitives. This proposed workflow and its connector/environment compatibility are unknown and untested.

   Setup caveat

   Keep cost unknown; verify current plan, limits and task-specific tools before testing.

   Availability in canada

   Not assessed by this reference update.

2. Name

   Claude (Cowork capabilities)

   Required capabilities

   1. Read owner-supplied synthetic structured/text copies without source-system access

   2. Use inspectable deterministic joins, set checks and calculations where specified

   3. Compare textual discrepancies and produce source-linked local draft artifacts for human review

   Capability source urls

   1. https://support.claude.com/en/articles/13345190-get-started-with-claude-cowork

   2. https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude

   Workflow compatibility

   Unknown and untested for every proposed plan.

   Selection basis

   Editorial candidate based on documented primitives only; no claim of project-system, accounting-system, learning-system or SIS integration.

   Current access note

   Current Claude documentation says Cowork and chat are merging into Claude, beginning with a staged Pro/Max rollout. Some accounts still show Cowork. Verify the account’s plan, interface, permissions and tools before testing; exact workflow compatibility remains untested.

   Setup caveat

   Use current help for actual runtime and permissions. Local files, apps or browser resources require the appropriate connected desktop setup; do not assume all cloud tasks require the app open.

   Availability in canada

   Not assessed by this reference update.