Task: Build the cut screens

A person browses the cut types and their rules, lists the cuts and reads one, with the data the cut model already holds.

• The BFF answers the cut types and the cuts with the reads the screens need.
• The cut types list and record draw the rules and the retention; the cuts list filters by type and groups by month; the cut record draws the body, the acts and the derived state.
• The permissions, the i18n, the tests, the route inventory and the menu design are wired and green.

(Implementation strategy. Written when work starts; key decisions are distilled into the parent story's * Decisions at close, but the plan itself stays — it is the historical record of what we did.)

Every number this task claims about a tree, with the command that produced it and the commit it was measured at. A number without both is not a measurement, because a clean-up record describes a tree and trunk moves under a branch: a consumer count, a gate's component count and a suite's size all change with no code changing, and the reader of a bare number cannot tell which tree it came from. Where the raw output matters, put it in an evidence/ file beside the task and name that file in the Command column.

Manual QA scenarios (scaffolded via compass add test_scenario) that verify this task. Link new ones here as they're created; the scenario doc itself links back via its "Verifies task" field.

The cut screens' first unit is built, reviewed and repaired.

• The reads: list_cut_rows (the derived state and label per row, the replacing body, the page's run tally) and get_cut_record (the body, the acts, the gaps, the state), with the cut read service, the paged read and the refusal of a malformed month as a value.
• Seven BFF routes: the cut types (list, record, choices, write, remove) and the cuts (list, record, runs). The signed-or-not is derived from the approval kind; the choices answer the retention destinations the store holds.
• The screens: Reference data > Cut types (the rules, the retention, the type's cuts grouped by the type's frequency, the edit through the standard dialog, the delete through the shared remove dialog) and Reporting > Cuts (the filters, the state badges with their basis and the indicative mark, the body cards, the gaps, the authorisation tab with the request link, the Runs tab, the history).
• The i18n in three locales, the server paging, the menu and paths, the route inventory and the menu design.
• The two-axis review was run and its findings closed: the choices contract, the server paging, the standard dialogs, the tabs and the version line, the permission constants, the test logger, the std::expected refusal, the Runs tab, the reader-note gap, the indicative mark, the request link, the grouping by frequency, the superseded basis, the draft notice, the retired-kind sentence, and the smells named in the review. One finding is deliberately left: the trigger's three resolution flags are generated from the contract main already holds; removing them is the selection work's change.
• Checks: the web build, BFF 85 files/1179 tests, web 188 files/2369 tests, the codegen drift check, the route inventory (417 routes) and lint. The reporting core suite's cut-read cases pass; two unrelated tests fail against the shared database, which is behind main (missing canonical_value and the report metrics tables) and is recreated at the deploy.

Left for the following units: the take flow, the authorisation raise and decide, the report's cut selection and the LIVE mint UI; the screens state those gaps.