TK-11041: disambiguate the two reprice tools before the wrong one gets run

a561f20ece07376db3fc02f582676c3de6be2f50 · 2026-09-12 08:17:03 -0700 · Steve Abrams

Two sessions independently built a reprice planner here hours apart. Same verified
chain, but tk11041-dryrun-reprice-plan.mjs hardcodes basis=LIST (right for the portal,
which does show list). The draft I merged asks Hektor for NET. Feed a net price to the
LIST tool and it discounts an already-discounted number — ~10% under on 39
customer-facing products, silently, because the arithmetic succeeds.

The email route makes that the LIKELY pairing, not an exotic one. This names which
tool goes with which source, and records the Skylark MTO finding neither other path
surfaces.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

Files touched

A tk11041-innovations-reconcile/rescrape/REPRICE-WHICH-TOOL.md

Diff

commit a561f20ece07376db3fc02f582676c3de6be2f50
Author: Steve Abrams <steve@designerwallcoverings.com>
Date:   Sat Sep 12 08:17:03 2026 -0700

    TK-11041: disambiguate the two reprice tools before the wrong one gets run
    
    Two sessions independently built a reprice planner here hours apart. Same verified
    chain, but tk11041-dryrun-reprice-plan.mjs hardcodes basis=LIST (right for the portal,
    which does show list). The draft I merged asks Hektor for NET. Feed a net price to the
    LIST tool and it discounts an already-discounted number — ~10% under on 39
    customer-facing products, silently, because the arithmetic succeeds.
    
    The email route makes that the LIKELY pairing, not an exotic one. This names which
    tool goes with which source, and records the Skylark MTO finding neither other path
    surfaces.
    
    Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---
 .../rescrape/REPRICE-WHICH-TOOL.md                 | 47 ++++++++++++++++++++++
 1 file changed, 47 insertions(+)

diff --git a/tk11041-innovations-reconcile/rescrape/REPRICE-WHICH-TOOL.md b/tk11041-innovations-reconcile/rescrape/REPRICE-WHICH-TOOL.md
new file mode 100644
index 0000000..f9d7c13
--- /dev/null
+++ b/tk11041-innovations-reconcile/rescrape/REPRICE-WHICH-TOOL.md
@@ -0,0 +1,47 @@
+# Two reprice tools live here. Pick by WHERE THE NUMBER CAME FROM.
+
+Built independently, hours apart, by two sessions on TK-11041. Both implement the same verified
+chain — `net = list x 0.90`, `retail = net / 0.65 / 0.85` — and both are read-only. The difference
+that matters is **what they assume about the number you feed them.**
+
+| The 7 prices came from… | Use | Why |
+|---|---|---|
+| the **authenticated portal** (acct 58315) | `tk11041-dryrun-reprice-plan.mjs` | It is built for the portal and hardcodes **basis = LIST**, which is right: the portal shows list (proven by DW's own June-2026 crawl — `price_retail` 134.95 / `price_trade` 121.46 = x0.90). |
+| **Hektor's email reply** | `stage-reprice.mjs --basis net` | The merged draft asks him to *"confirm the current per-yard **net**"*. A net number is NOT a list number. |
+
+## The trap, stated plainly
+
+`tk11041-dryrun-reprice-plan.mjs` assumes LIST and applies `x0.90`. **Feed it a NET price and it
+discounts an already-discounted number** — retail lands ~10% under where it belongs, on 39
+customer-facing products, silently, because the arithmetic succeeds and the output looks normal.
+
+That is not hypothetical: the email route we just opened asks for net specifically, so the wrong
+pairing is the *likely* one, not the exotic one.
+
+`stage-reprice.mjs` refuses to run without an explicit `--basis` and has **no default**, precisely
+so this cannot happen by omission.
+
+## What each adds that the other does not
+
+**`tk11041-dryrun-reprice-plan.mjs`** — a full 888-line staged plan already written out, with the
+per-column write targets spelled out for the gated write. Further along for the portal path.
+
+**`stage-reprice.mjs`** — required `--basis`; 12 self-tests, the meaningful ones negative ($0,
+negative, missing basis, unknown basis all throw rather than produce a price); a sanity band against
+the Fall-2025 list that catches a wrong-unit quote (a 90%-low number is flagged, not priced); names
+unpriced SKUs instead of dropping them; and carries the per-pattern order terms.
+
+## One finding only the second tool has
+
+**Skylark (SKL-001..004) is MADE TO ORDER** — no stock, no cut-fee code, MTO warehouse, per the
+vendor's own Fall-2025 list. Decide deliberately whether those 4 become buyable stocked rolls or
+stay quote/lead-time, and set a lead time, *before* they go sellable. Neither the portal pull nor
+the other planner surfaces this.
+
+*(It was nearly missed here too: the first parser silently dropped those 4 rows because its regex
+expected a `Fee X` freight code and Skylark's cell reads `MTO (no cut-fee code)`. 35-of-39 looked
+like a clean run.)*
+
+## Unchanged for both
+Neither writes to `dw_unified` or Shopify. The repricing of the 39 is canonical + customer-facing
+and stays gated.

← 36062e5 TK-11041: stage-reprice — 7 arriving numbers to a staged 39- · back to Dw Unbuyable Recovery Pilot · TK-11041: stage shopify_product_id backfill dry-run (39/39 c 995b159 →