Task: Price the bonds a report run carries

A bond report run completes with its bonds priced. Today it builds no curve and no trade: the curve configuration names quote identifiers no producer publishes, so ORE cannot bootstrap the discount curves, and every bond then fails to resolve one.

• The Acme curve configurations name only quotes the deployment's synthetic feeds publish
• Every OIS pillar publishes under the tenor its curve configuration names
• pricingstats.csv reports a non-zero NumberOfPricings per bond

(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.

A bond report run prices its bonds. Verified end to end on a clean `compass db recreate`, a restart and one `bootstrap setup –profile acme_demo`, with nothing run by hand: all eight bonds show `NumberOfPricings=2` in `pricingstats.csv`, the log carries no `Failed to run analytics` and no `did not find object`, and the exported portfolio carries `ReferenceCurveId=GBP-SONIA` on each bond.

Four independent causes had to fall, and only the first was expected.

The curve series did not exist. A tenant held 271 market series and not one expressed a curve, so a bond could name nothing and the soft key had no target. The curve's definer now publishes its series, keyed by the oresmd URI of the curve's own coordinates and scoped by party.

The market feed and the curve configuration spelled the same pillar two ways, and the configuration asked for a `1Y` tenor the market's convention does not define and a money-market segment the feed cannot produce. The one-year tenor is `12M` and the segment is gone, because an overnight RFR curve is OIS throughout.

A live feed gave a fresh environment no close for the valuation date an end-of-day report asks for. The ingest loop now records each observation under the last completed business day's close as well as the day it arrived.

And the trade named no curve. `bond_issue.reference_curve_id` holds an oresmd URI, a translator turns it into the ORE index name on export, and the name must be the market's own forwarding key, which direct engine runs confirmed: `GBP-SONIA` prices, while `GBP1D`, `Yield/GBP/GBP1D` and empty all reproduce the failure.

The validator is projection-independent on purpose. The identity projection is written only on the C++ path, so a series inserted by SQL has no identity row until a backfill runs; a validator joining it would have refused every trade naming a freshly published curve.

Left for a reviewer. The new column appears as a position on the generated shell command surface. `find_reference_curve_id` skips a party's curve whose URI carries no `curve_id`. And the commit trails a task in the product backlog whose story owns the wider pipeline rather than this arc's story.