01
Every SKU lives in exactly one lifecycle state
ALLOWED_TRANSITIONS enforce the spec — no orphan drafts, no zombie listings, no SKUs sitting in 'testing' for six months because no one had time.
// for heads of ecommerce
Run a tighter store, with less context-switching.
Catalog hygiene, velocity calls, briefs, repricing, archive — the loop your team almost-catches-up-to every Monday. Magistry's Catalog Specialist runs it on a schedule, inside your margin floor, with one lifecycle state per SKU.
Tier-of-evidence gates ALLOWED_TRANSITIONS enforced Margin floor is policy
// why heads of ecommerce pick magistry
A catalog spine that catches up to itself, every night.
You don't need another dashboard. You need the SKUs in your store to actually behave the way the spec says they should — and to keep behaving when no one's watching.
02
Tier-gated velocity calls, not gut calls
Tier A (verified cost_per_item) only writes prices. Tier B (reverse-image match) drafts. Tier C parks for review. Margin floor is policy, not preference.
03
Weekly digest that fits in one Slack thread
What launched. What archived. What got repriced. What needs a human call. Linked straight into the decision_log row that prompted each one.
04
Brief-to-PDP pipeline, audited end-to-end
Researcher surfaces candidates from its discovery lanes — ad libraries, marketplaces, competitor catalogs, search demand. Catalog Specialist drafts copy in your voice. You approve. Magistry pushes to Shopify with the trademark and uniqueness checks already done.
// this week
What Magistry did for your catalog this week.
An illustrative week with the Catalog Specialist in live mode — every change a row, every row a one-click reverse.
- Archived 27 long-tail SKUs that slipped under the 90-day sell-through floor — €0 revenue lost, 14% of catalog noise gone, all reversible from one screen.
- Repriced 41 Tier-A SKUs (verified cost_per_item) with margin-aware deltas — no SKU dropped under its floor; one needed a manual call and Otto flagged it before it ran.
- Drafted PDPs for 6 new Researcher candidates and queued them for your approval — copy in your voice, trademark cleared, uniqueness verified.
// from your dashboard
One lifecycle state. One transition. One audit row.
SKUs move through ALLOWED_TRANSITIONS only. Below: an archive call on a long-tail SKU that missed its 90-day sell-through window, sitting in your weekly digest.
weekly digest ── catalog_specialist ── wk 21
─────────────────────────────────────────────
launched 6 (Researcher → drafts)
repriced 41 (Tier A, all in margin)
drafted 27 (under 90d sell-through)
held for review 3 (Tier C, awaiting cost)
escalated to you 1 (trademark conflict)
─────────────────────────────────────────────
row #84,201 ACTION draft
sku ceramic-mug-sage-12oz
state_from ACTIVE
state_to DRAFTED (transition OK)
evidence sell_through 4% / floor 18%,
90d revenue EUR 412, cost EUR 380
revert republish in 1 click (derived)
status APPLIED
// what you actually do
Read the digest. Approve the holds. Move on.
You're not auditing 41 reprices on Monday morning. You're reviewing the three rows Magistry held because the evidence wasn't strong enough — and clicking through to the row that already ran if you want to.
otto: weekly digest ready · 1 held for trademark · 3 Tier-C awaiting cost · view 41 reprices →
// why context matters
An agent is only as good as the context it sees.
An agent is only as smart as the data it sees. Magistry works from the full context of your store — orders, refund history, margin tiers, lifecycle state — so its calls are grounded in your reality, not a generic playbook, and every one is a row you can audit or reverse.
// run a tighter catalog
Hand the SKU lifecycle to the spine, keep the call on the bets.
Spin up a tenant, point Magistry at your Shopify store, and watch the lifecycle states snap into place — every transition logged, every change reversible.
Operator-owned kill switch · Dry-run from day one