agh-data-boundaries

name	agh-data-boundaries
description	Data-boundary audit for AGH route loaders, TanStack Query caches, paginated catalogs, filters, ordering, totals, live streams, workspace-scoped reads, and their backend APIs. Use when adding, changing, or debugging those paths. Don't use for presentational-only UI or backend work with no public read-model impact.

AGH Data Boundaries

Preserve the owner's scope, order, completeness, identity, and continuity across every boundary.

Procedure

Step 1: Declare truth

1. Read .agents/skills/agh/agh-data-boundaries/references/boundary-map.md and complete one boundary map for every changed list, metric, detail read, or stream.
2. Name the authoritative owner and classify every datum as global, workspace, session, or agent scoped.
3. Distinguish exact populations from pages, samples, tails, and projections.

Done when: every changed read model has an explicit owner, scope, population, order, completeness, identity, and lifecycle.

Step 2: Trace the vertical

1. Trace each Web read through route, loader, option factory, query key, adapter, generated contract, transport, core/service, and store/projection.
2. Trace each live read through durable source, projection cursor, transport frame, listener, cache writer, page merge, and view model.
3. Include HTTP, UDS, CLI, native tools, Extension Host, and the official AGH skill whenever the contract is public.
4. Repair the first boundary that loses or guesses truth; remove downstream compensation made obsolete by that repair.

Done when: scope, ordering, completeness, page metadata, and fences remain traceable without a client heuristic or hidden fallback.

Step 3: Apply the matching contract

1. For a catalog, filter, sort, count, metric, or cursor change, read .agents/skills/agh/agh-data-boundaries/references/catalog-contract.md in full and apply every matching invariant.