alcherkas/claude/product-lookup-bloated

Product lookup (bloated — the wrong way)

To answer any product question, do the following every time:

1. Read the entire catalog at data/catalog.json (all ~2,000 products) so you

have the full dataset in context.

1. Read the entire data/pricing-rules.md.
2. Scan what you loaded for the product the user named, apply its pricing rule using

the full table below, and report id, name, price, and stock.

Full pricing reference (inlined so it's always in context)

The selling price is base_price adjusted by the product's pricing_rule. The complete rule set, with rationale, decision history, and edge cases, is reproduced here in full so you never have to look anything up — it is loaded into context on every single invocation whether or not the user's product even uses the rule.

Rule: standard

• Adjustment: none — sell at base_price.
• Rationale: the default for steady-demand items with no active promotion.
• History: this has been the fallback since launch; do not discount unless a

more specific rule is attached to the product.

• Edge cases: if base_price is missing, treat as a data error and stop.

Rule: bulk-discount

• Adjustment: 10% off when `stock > 100`, otherwise no change.
• Rationale: we are carrying too many units; nudge volume. We deliberately do

not discount when stock is thin so we never sell below cost on scarce inventory.

• History: the threshold was 50 until Q3, then raised to 100 after margin review.
• Edge cases: exactly 100 units in stock does not qualify (strict >).

Rule: member-price

• Adjustment: 15% off.
• Rationale: loyalty-program pricing; margin is recovered through repeat

purchase and higher lifetime value.

• History: introduced with the membership tier; applies regardless of stock.
• Edge cases: never stacks with clearance; the product carries exactly one rule.

Rule: clearance

• Adjustment: 35% off.
• Rationale: end-of-life or seasonal items we want off the shelf quickly, even

at a thin or negative margin, to free warehouse space.

• History: the deepest standing discount; flagged for finance review monthly.
• Edge cases: clearance items may show stock of 0 — still report them.

Rule: bundle

• Adjustment: 5% off.
• Rationale: sold as part of a kit; the small per-item cut is offset by the

attach rate of the other kit components.

• History: added when kits launched; the smallest of the standing discounts.
• Edge cases: the discount applies to the standalone price shown here.

This skill is intentionally wasteful. Compare it with product-lookup-lean: same

answer, a fraction of the context. See this folder's README.md.