/quality-goods
what it is
a shopping method. before anything from a mass retailer is allowed on the page, it checks a curated list of heritage and small-batch vendors you actually trust, then hands back a tight comparison - price, materials, repairability, and why the maker deserves the money. adapted from nathan williams' method.
when to use it
any time you're buying a physical thing you want to outlast the receipt - kitchen tools, clothes, gifts, the good version of anything. it handles budgets, and a gift-card-at-one-store mode.
when not to: commodity purchases where you genuinely don't care.
why it works
good makers get forgotten at exactly the moment you're buying. their sites are sparse, they don't bid on your search terms, and the algorithm serves whoever paid. the fix is memory: a vendor list that's read at the start of every search, so taste accumulates instead of evaporating. the trust math is explicit too - cross-vendor agreement beats reviews. if two shops you already trust independently stock the same maker, that's the buy signal. a 150-year-old can opener doesn't need a five-star average.
what you get
a scannable comparison led by the strongest trust signal, honest gaps ("none of your saved vendors carry this"), and a vendor file that gets better every time you shop. the folder ships with my actual starter list - 25+ pre-vetted vendors across furniture, kitchen, kids, linens, and hardware, with my notes on each, plus category shopping rules (including why your dining rug needs 24 inches past the table on every side). prune it to your taste; the list is the product, the skill is just the librarian.
install
needs claude code. unzip the folder into ~/.claude/skills/ so the file lands at ~/.claude/skills/quality-goods/SKILL.md - the vendor list rides along in the same folder. then say "find me a quality [thing]".
the file
the whole skill, unabridged (~900 words) - click to read
---
name: quality-goods
description: "Find and buy quality goods from trusted small-batch / heritage vendors instead of mass retailers. Searches a curated vendor list FIRST, surfaces a concise comparison with trust signals (cross-vendor agreement, brand longevity, real reviews), then extrapolates to similar makers if needed. Use when the user says 'where can I buy X', 'find me a quality X', 'I need a new X', 'help me shop for X', 'gift idea for X', '/quality-goods', or is shopping for any physical good and wants well-made, made-to-last, made-to-repair options. Also use to add a vendor to the list ('add X to my vendor list')."
source: "https://x.com/nwilliams030/status/2048526120039952540"
---
# /quality-goods — Buy Well-Made Things from People Who Make Them
Adapted from Nathan Williams' method ([source](https://x.com/nwilliams030/status/2048526120039952540)).
The idea: searching for quality, made-to-last goods from small/legacy businesses is the hard
part — their sites are sparse and hard to search, and good makers get forgotten at purchase time.
This skill solves that by always checking a curated, pre-vetted vendor list FIRST, then surfacing
a tight comparison with explicit trust signals before anything from Amazon/Target shows up.
## What this skill is for
- "Where can I buy a good [can opener / wool socks / linen napkins / kids gift]?"
- "Find me a quality [item] under [$X]."
- "I have a $50 gift card to [store] — what fits my taste?"
- "Add [shop] to my vendor list."
## The vendor list (source of truth)
`quality-goods-vendors.md` — keep it wherever your notes live (next to this skill file works
too); read it at the start of every search. It's the curated shortlist of trusted heritage /
small-batch / made-to-repair sellers and pre-vetting aggregators. The user edits it directly;
it grows over time.
**A starter list ships in this folder** — `quality-goods-vendors.md`, a real working list of
25+ pre-vetted vendors (heritage makers, small-batch furniture, kids goods, hardware, linens)
plus category-specific shopping rules and the inclusion bar. On first run, confirm where the
user wants it to live, copy it there, and invite them to prune it to their taste and add shops
they already trust. The list is the product; the skill is just the librarian.
## Workflow
### 1. Read the list
Read `quality-goods-vendors.md`. These vendors get searched and surfaced **before**
any other option.
### 2. Pin down the item
Confirm: what is it, and any constraints — price bracket, recipient (e.g. kids), materials
preference, deadline. Ask only if a constraint would change the results. Don't over-interrogate.
### 3. Search the listed vendors first
Use web search and page fetches (or a browser tool for sites that block plain fetches) to
check the listed vendors' sites for the item. Match against their actual inventory — these
stores are small, so a real check beats guessing.
### 4. Surface a concise comparison
For each match, output a **bulleted** entry with:
- **Product name** + vendor
- **Price**
- **Materials** it's made of
- **Care / maintenance** notes (and whether it's repairable / parts available)
- **Why the maker is trustworthy** — brand history / years in business / craftsmanship reputation
- Image of the product when available
Keep it tight. The point is a fast, scannable decision — not a wall of text.
### 5. Read the trust signals out loud
- **Cross-vendor agreement** = strongest signal. If two+ listed vendors carry the *same*
product/maker, say so explicitly and lead with it. (Worked example from the source: a Nogent
can opener showing up at two saved sites → 150+ yr maker → instant buy.)
- **Brand longevity** — decades of craftsmanship, still making the thing.
- **Real reviews** — organic, across multiple trusted sites. Flag and discount reviews that read
AI-generated or paid. Their presence is a signal the product is genuinely good.
### 6. If it's not on the list — extrapolate (with caution)
Use the list as a *taste profile* to find similar makers:
> "I'm looking to buy from businesses with several years of history, ideally still making things
> in small batches. Vet items through organic, positive reviews across multiple sites — exclude
> anything that reads AI-generated. Avoid trendy DTC brands that spend heavily on marketing and
> haven't been around long."
⚠️ **This mode is weaker protection** against drop-shipping and AI-slop storefronts than the
curated list. Low-quality sellers are very good at *looking* high-quality, and the techniques
shift constantly. Vet anything from extrapolation through a second method before recommending a buy.
### 7. Offer to capture new vendors
If a search surfaces a great new maker, or the user mentions a shop they loved in person / a
friend recommended — offer to add it to `quality-goods-vendors.md` (same table format, with a
`(via Person)` or source note). Curation over volume: only add ones that clear the inclusion bar
at the bottom of that file.
## Query tips (apply when relevant)
- **Price bracket** — always honor a stated budget; if nothing in the approved vendors fits, say so
and offer to extrapolate to similar stores within budget.
- **Gift-card / single-store mode** — "What at [store] best fits my criteria? I have a $X gift card."
Narrow by category on request ("just the kids items").
- **Category narrowing** — kids, kitchen, hardware, desk, linens, etc. Note in the list which
vendors specialize where.
## Output discipline
- Lead with the strongest trust signal (usually cross-vendor agreement), not the cheapest option.
- Bulleted, scannable, image where possible.
- Never pad results with mass-retailer options unless the curated + extrapolated search comes up empty
and the user asks.
- Be honest about gaps: "none of your saved vendors carry this" is a useful answer.
built in claude code, tuned on real use. the personal plumbing is stripped - what's here runs anywhere.