Task: Seed the ladder family and let it fail like the rest

The ladder family is seeded for ACME and Northwind with its configs and cut, fails at the engine with the gap named, and appears in the reports list.

• The design's report list is seeded, dual-cycle rows split into EOD and LIVE definitions.
• Each definition binds its six configuration types, the sensitivity analytic and the files the configs state.
• The docs and the index state the definitions and what still blocks the engine; a live list shows them.
• The seed generators, pgTAP and the gates are green.

The product owner overruled the design's decision to hold the ladder family back, so the design's * Reports to build list is the seed list. Rows 1 to 10 are the family; rows 11 to 14 wait on swaption, cap and floor, credit and equity volatility data and stay stated and unseeded.

• the catalogue gains the ten reports through report_catalogue.json and the two generators, so the artefact populates are regenerated rather than hand written. A row the design marks eod and intraday is two catalogue entries, because a definition states one cut selection. That is fifteen definitions per party.
• the report names are the ones the ladder pages already carry, because the index lists a report by its page. The design's rates delta ladder is the Rho ladder and its currency pair spot ladder is the Delta ladder, which is the family's general first-order reader.
• every ladder runs npv and sensitivity, except the par-rate DV01 ladder, which reads the par column and asks for no valuation. The two analytics already have profiles, so a ladder binds the run settings the sensitivity reports already bind: the same rows, which name sensitivity.xml and simulation.xml.
• the engines are Risk Sensitivities, the same set the delta, gamma and DV01 reports use, because a bump wants the prices that carry analytic sensitivities. A close twin takes the default official engines, as the pricing table says.
• the definition binds the six configuration types through the existing artefact pattern, so nothing about the binding mechanism is new.
• the ladder family is decided before the party rules. A definition that produces no figure costs nothing to hold, and the catalogue should state what the platform runs. Every party therefore holds the fifteen, the holding company included, and Northwind holds them through the London rows its catalogue reads.
• the environment is shared and already provisioned, so the seeds reach it by being applied: the artefact populates are staged again and the publish carries them onto the rows the parties hold. evidence/apply_the_ladder_family.sql does that and states the two secrets of the order: run settings before definitions, and the Northwind rows staged and not published, because this environment holds no Northwind tenant.

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 ladder family is stated in the report catalogue and applied to this environment. The ten rows of the design's list are fifteen definitions per party, because five of the rows run on both cycles and each cycle is a definition:

Each definition binds the six configuration types: conventions, todays_market, curve_configuration, pricing_engines — Risk Sensitivities for the plain name and the official engines for a close twin — risk_report_config and engine_settings. The run settings are the party's existing sensitivity rows, so the sensitivity analytic carries recalibrateModels Y, outputSensitivityThreshold 0.000001, parSensitivity Y, sensitivityConfigFile = sensitivity.xml and marketConfigFile = simulation.xml. The definitions run through the existing publish, so no table had to be invented for the shift: the size is still a setting no table holds, and the pages say so.

Rows 11 to 14 are stated and not seeded, because they wait on swaption, cap and floor, credit and equity volatility data.

What the run stops on. Nothing was run. No table holds a SensitivityAnalysis document and no mapper reads one, no table states a shift size, the dataset holds no volatility quote and no FX spot, and the report inputs story is open. That is the same list that stops "Delta and Gamma" and every other sensitivity report, so the ladder fails like the rest rather than in a way of its own.

Docs. The index's ladder table states the definitions and their cuts, "what is not built" names the four gaps in one place, the pages of the five dual-cycle ladders state both cycles, and the design page records the overrule and the name each row is seeded as.

Tests. The two seed generators run and a re-run produces no diff. The five reporting pgTAP suites pass (62 tests) and the two sensitivity analytics suites pass (18 tests). The docs checks pass: lint, site-page on the changed pages, org-links, heading-numbers, structure-notes and pattern-uses. The SQL and codegen checks pass: sql-hygiene, populate-refs, sql-reachability, generated-column-refs, entity-writes, handler-permissions, rls-widenings and cmake-sources. er-diagram fails on ores_schema.dq_staging.puml, which the commit that added cut_kind_code to the DQ artefact table left unregenerated; that failure is on origin/main and this branch does not touch a schema file. The codegen checks need the codegen virtual environment, which this worktree does not hold, so they were run with the one in the main checkout.

Evidence. evidence/apply_the_ladder_family.sql is the apply, and evidence/the_ladder_family_live_list.txt is what the database answers: the counts by party and cut, the live list with each definition's schedule and run settings, the six bindings, the two files, and Northwind's staged rows.