Skip to main content

Review Workbench

14 requirements · 40 scenarios

Filter Reset Button Visibility

JR-review-workbench-001
The review workbench filter controls SHALL display a "Remove All Filters" button when any filter is not set to "All".
3 test scenarios
  • Button appears when compliance filter is active JR-review-workbench-001.1
  • Button appears when status filter is active JR-review-workbench-001.2
  • Button hidden when all filters at default JR-review-workbench-001.3

Filter Reset Button Behavior

JR-review-workbench-002
The "Remove All Filters" button SHALL reset all filters to their default "All" state when clicked.
2 test scenarios
  • Reset all filters to default JR-review-workbench-002.1
  • Clear persisted filter preferences JR-review-workbench-002.2

Citation Link Navigation and Highlighting

JR-review-workbench-004
Citation links in the report pane SHALL navigate to and highlight the referenced paragraph in the contract pane.
3 test scenarios
  • Click citation link in report pane JR-review-workbench-004.1
  • Citation link inside contract pane JR-review-workbench-004.2
  • Target paragraph not found JR-review-workbench-004.3

Edited Recommended Response Persistence

JR-review-workbench-005
When a user edits the Recommended Response in the review workbench, the updated text SHALL be persisted and restored after refresh.
2 test scenarios
  • Edited response persists after reload JR-review-workbench-005.1
  • Edited response is persisted to backend JR-review-workbench-005.2

Row Status Merge Contract

JR-review-workbench-007
Incoming review row updates SHALL be merged with existing rows according to a shared contract, not replaced wholesale.
3 test scenarios
  • SSE update preserves ignored status JR-review-workbench-007.1
  • SSE update applies explicit reopen JR-review-workbench-007.2
  • Full refresh preserves applied status JR-review-workbench-007.3

Row Status State Machine

JR-review-workbench-008
Row status transitions SHALL follow an explicit state machine and reject invalid regressions.
3 test scenarios
  • Applied is terminal unless undone JR-review-workbench-008.1
  • Ignored persists until reopened JR-review-workbench-008.2
  • Reopen explicitly clears ignored JR-review-workbench-008.3

Backend Status Authority

JR-review-workbench-009
The backend SHALL persist row status updates and include status in all review row payloads.
2 test scenarios
  • Ignore persisted and emitted JR-review-workbench-009.1
  • Apply persisted and emitted JR-review-workbench-009.2

Ignore Action Independent of Edit Progress

JR-review-workbench-010
The Ignore button in the review workbench SHALL remain enabled and clickable regardless of whether document edits are currently being applied. The UI SHALL update immediately using optimistic updates.
4 test scenarios
  • Ignore button enabled during edit in progress JR-review-workbench-010.1
  • Ignore updates UI immediately (optimistic update) JR-review-workbench-010.2
  • Ignore rollback on API failure JR-review-workbench-010.3
  • Ignore multiple findings during batch edit JR-review-workbench-010.4

Background Review SSE Coverage

JR-review-workbench-011
The frontend SHALL maintain SSE connections for all reviews with status processing or pending, regardless of which review is active in the workbench.
2 test scenarios
  • Background completion updates status JR-review-workbench-011.1
  • Switching active review does not disconnect processing streams JR-review-workbench-011.2

Edited Badge Accuracy

JR-review-workbench-012
The "(edited)" badge in recommendation cells SHALL only appear when the user has actually modified the text content.
3 test scenarios
  • Badge not shown after opening and closing without changes JR-review-workbench-012.1
  • Badge shown after actual text modification JR-review-workbench-012.2
  • Badge preserved after subsequent no-op edits JR-review-workbench-012.3

Genealogy Persistence Across Apply Paths

JR-review-workbench-013
Apply edit operations SHALL persist genealogy steps regardless of apply path, including opcode replay and direct subagent fallback.

- Precomputed genealogy steps SHALL be stored in apply_data when opcodes are precomputed
- Opcode replay SHALL persist genealogy steps using precomputed genealogy data when available
- Direct subagent apply SHALL persist genealogy steps from edit tracking metadata
2 test scenarios
  • Opcode replay persists precomputed genealogy JR-review-workbench-013.1
  • Direct subagent apply persists edit genealogy JR-review-workbench-013.2

Report Table Column Consolidation

JR-review-workbench-016
The review workbench report table SHALL consolidate Compliance + Analysis into one column and Recommended Response + Suggested Edits into one column.
2 test scenarios
  • Consolidated headers JR-review-workbench-016.1
  • Recommended response remains visible by default JR-review-workbench-016.2

Analysis Details Visibility Controls

JR-review-workbench-017
The report table SHALL collapse analysis details by default and provide both per-row and global controls for expanding analysis details.
4 test scenarios
  • Analysis collapsed by default JR-review-workbench-017.1
  • Per-row analysis toggle JR-review-workbench-017.2
  • Global analysis toggle JR-review-workbench-017.3
  • Persisted analysis default JR-review-workbench-017.4

Batch Edit Application with Progress Events

JR-review-workbench-020
The system SHALL provide "Apply All" and "Apply All with Comments" buttons that dispatch progress events during batch edit operations for cross-view status tracking.
5 test scenarios
  • Apply All dispatches start event JR-review-workbench-020.1
  • Apply All dispatches progress events JR-review-workbench-020.2
  • Apply All dispatches complete event JR-review-workbench-020.3
  • Apply All with Comments follows same pattern JR-review-workbench-020.4
  • Ignored rows excluded from pending count JR-review-workbench-020.5
All Specs