# Scoops, spatulas, ladles, tongs and whisks: research notes

Date: 2026-10-25. Scope: hand tools that move, turn, scrape, grip, whisk or portion food: spatulas (fish spatulas, turners, silicone scrapers, offset icing spatulas, grill spatulas), ladles, kitchen and grill tongs, whisks (balloon, French, flat, ball, coil, silicone-coated), and scoops (ice cream scoops, dishers or portion scoops, transfer scoops for flour and sugar); not spoons, spoodles, bench or bowl scrapers (sibling `kitchen-other`), measuring scoops and spoons (sibling `measuring-cups`), skimmers and spider strainers (sibling `strainers`), bar spoons, ice tongs and ice scoops for drinks (siblings `cocktail-mixing-tools`, `bottle-wine-serving-tools`), or spoon rests and lid holders (sibling `cookware-stove-accessories`). Taxonomy: `supplies > kitchen > scoops`.

Overlaps left to siblings: a slotted spoon or spoodle sold in a utensil set with a turner and ladle belongs to whichever leaf the lead item does; this register only describes the turner, ladle, tongs, whisk or scoop in it. Grill tongs and grill spatulas stay here because the catalog has no grilling leaf (searched `catalog.json` for "grill", "bbq", "barbecue": no match). Ice scoops sold for bars go to the bar siblings; a disher used for ice cream stays here.

**2026-10-25, after the blind check:** `tool_type` now has a `set` value, because this leaf's top queries by volume ("kitchen utensil set", "spatula set", "wooden spoon set") carry several tool nouns at once and previously matched no value (check.md B1-d). This register still does not break a set into its per-piece facets; that is logged as unresolved, not fixed. Other scope conflicts the check found — spider strainers and skimmers claimed by both this register and `strainers`; dough scrapers sent by `molds` to `knives` or `scoops` while this register's own scope sends them to `kitchen-other`, a catch-all with no register; coffee scoops not excluded from `transfer scoop` — need `catalog.json` or a sibling register changed, which is outside `scoops/` and out of scope for this task. See "Corrections after the blind check" near the end of this file.

Quote labels: **verbatim** = the sentence in quotation marks was in raw text I read myself (curl plus tag stripping, pdftotext, or the PubMed E-utilities plain-text abstract); **fetched excerpt** = paraphrased from raw text I read, because the passage is long; **search excerpt** = the claim came only from a search-engine summary. No search excerpt is used anywhere: the session's shared web-search budget (200 calls across concurrent agents) was already spent when this run started, so every source below was reached by a direct fetch of a known URL, a sitemap, the PubMed E-utilities API, Crossref, or the Internet Archive's CDX index.

Research constraints this run: no WebSearch calls were available (see above). The Internet Archive went "Temporarily Offline" partway through, after the BfR and EUR-Lex lookups, so no marketplace snapshot could be read. OXO, Rubbermaid Commercial and WebstaurantStore returned 403 to curl. Serious Eats and Wirecutter were not tried, because sibling runs today logged both as blocked for this agent (`kitchen-boards`, `graters`, `mashers` notes).

## Seed

Seed output, verbatim:

```
=== Example entry (from an unrelated category) ===
category: kitchen-knives
what this shows: a numeric spec that only makes sense as one side of a trade-off, and a visual grammar borrowed from the hobbyist community that puts the trade-off on one axis.

<example>
{
  "id": "steel_hardness",
  "label": "Steel hardness",
  "definition": "How hard the blade steel is, measured on the Rockwell C scale, which sets how long an edge stays sharp and how easily it chips.",
  "why_experts_care": "Harder steel holds a thinner, sharper edge for longer but is brittle, so it chips on bones and hard squash and is harder to sharpen at home. Softer steel dulls faster but bends rather than breaks and comes back with a few strokes on a honing rod. Japanese knives cluster at the hard end and German knives at the soft end, which is most of what the Japanese-versus-German debate is actually about.",
  "kind": "numeric",
  "unit": "HRC",
  "range": { "min": 52, "max": 66, "typical": "German-style knives sit around 56 to 58. Japanese-style knives sit around 60 to 63." },
  "roles": ["annotation", "sort_key"],
  "novice_answerable": false,
  "source": "listing",
  "listing_field_hint": "Often in the specs table as HRC or Rockwell. Missing on cheaper knives, which usually means soft.",
  "backfill_cost": "low",
  "backfill_note": "Manufacturer pages carry it for anything mid-range and up. Below that, infer soft from the steel name if given.",
  "visual": { "chartable": true, "grammar": "edge retention versus toughness scale", "encoding": "position on a single horizontal scale, hard on the right, with chip risk shaded in" },
  "priority": { "narrows_set": 3, "cost_to_answer": 4 },
  "evidence": ["knifesteelnerds_hardness", "seriouseats_knives"]
}
</example>

Trade-off entry it belongs to:
{
  "optimise": "steel_hardness",
  "give_up": "toughness and ease of sharpening",
  "plain": "Push hardness up and you get a knife that stays sharp for months but needs a whetstone and care around bones. Push it down and you get a knife you can abuse and hone in the kitchen but will sharpen more often.",
  "see_more_branch": "See more low-maintenance knives"
}

=== Persona ===
A car mechanic, who thinks in wear, serviceability and what fails first

Ask yourself two things and write the answers under 'Seed' in notes.md:
  1. What is the scoops equivalent of the example above?
  2. What would this persona notice first about scoops that a typical buyer would not?
Keep what survives the data check. Drop the rest.
```

1. **Equivalent of steel hardness: the head material's heat rating, which only makes sense against how thick the head is.** On a knife, hardness buys edge retention and costs toughness. On a turner, scraper, ladle or whisk the head material buys either a thin edge or heat resistance, and it decides whether the tool may touch a nonstick pan. ATK's nonstick-spatula test puts numbers on both sides: nylon and fiberglass-resin heads 1.3 to 1.8 mm thick that melted between 400 and 500 °F, silicone heads 2.5 to 7 mm thick that survived 600 °F; bare metal fish spatulas are thinner still (0.8 to 0.9 mm) but scratch nonstick coatings. Survived as `head_material`, `max_heat_temp`, `head_thickness`, the "Head thickness versus heat" visual grammar (one scale, like the knife example's) and the first trade-off. Only the heat ratings carry numbers in the JSON, because they have brand pages behind them (Vollrath, GIR); the thickness figures are one publisher's measurements and stay here.
2. **What the car mechanic notices first: the joint and the moving part fail before the head does.**
   - Two-piece silicone spatulas trap water and food where head meets handle; one ATK model "collected dishwater in a small notch ... that water poured out into freshly whipped cream", another's handle "melted in our heat test". Silicone over a steel core: "water got under the silicone coating and visibly rusted the metal core within." Vollrath sells "one-piece" as the fix on turners, tongs, dishers and ladles. Survived as `seamless_construction`.
   - Dishers have a spring, gears and a sweeping strip; ATK warns their "thin metal levers are more prone to breaking when used on firm, frozen ice cream", and Vollrath prints "Not intended for hard or frozen foods". Zeroll's selling point is "no springs to replace", but its heat-conducting fluid means "DO NOT PLACE IN DISHWASHER". Survived as `release_mechanism` and the scoop trade-off.
   - Tong locks were where ATK's tongs differed most in use (sliders that jam, plugs that creep back to locked, all-metal locks that grind). **Dropped** as a facet: one publisher, no field on brand pages, and the failure is a design detail a buyer cannot filter on. Kept in Trimmed detail.
   - A whisk loop "detached from the handle" while whipping butter (ATK nonstick whisks), and a multipart whisk handle "held water". Folded into `seamless_construction` (it covers whisks) rather than a separate facet.
   - Nylon wears by melting and chipping ("showing scratches and nicks after only a few cooking tests and melting when left on a hot skillet", ATK). Folded into `max_heat_temp`.
   - Wood dries and cracks in the dishwasher (ATK wooden spoons: "most spoons looked bleached and felt as dry as driftwood" after 10 cycles). Kept as a note on `dishwasher_safe`; spoons themselves are out of scope, but wooden spatulas are in.

## Sources and what each supports

### Regulators and standards

**21 CFR 177.2600, Rubber articles intended for repeated use** (`cfr_177_2600`). https://www.law.cornell.edu/cfr/text/21/177.2600 (Cornell LII, curl). Confidence high.
- **verbatim** "Rubber articles intended for repeated use may be safely used in producing, manufacturing, packing, processing, preparing, treating, packaging, transporting, or holding food , subject to the provisions of this section."
- **verbatim** the elastomer list includes "Silicone (Si) elastomers containing methyl groups" and "Silicone (Psi) elastomers containing methyl and phenyl groups".
- **verbatim** (e) aqueous foods: "total extractives not to exceed 20 milligrams per square inch during the first 7 hours of extraction, nor to exceed 1 milligram per square inch during the succeeding 2 hours"; (f) fatty foods: "not to exceed 175 milligrams per square inch during the first 7 hours". Kept here only; the extraction limits are not something a buyer or a listing ever shows.
- Supports: the US rule that governs silicone heads.

**21 CFR 177.1500, Nylon resins** (`cfr_177_1500`). https://www.law.cornell.edu/cfr/text/21/177.1500 (Cornell LII, curl). Confidence high.
- **verbatim** "The nylon resins listed in paragraph (a) of this section may be safely used to produce articles intended for use in processing, handling, and packaging food , subject to the provisions of this section".
- **verbatim** one grade is limited to "articles intended for repeated use in contact with food at temperatures not to exceed 212 °F" (Nylon 612). Not used in the JSON: no listing names the nylon grade.
- Supports: the US rule that governs nylon heads.

**Commission Regulation (EU) No 284/2011** (`eu_284_2011`). EUR-Lex HTML via Internet Archive snapshot https://web.archive.org/web/2024/https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32011R0284 (legislation.gov.uk returned an empty 202 to curl). Confidence high.
- **verbatim** "These notifications and alerts primarily concern polyamide and melamine plastic kitchenware that does not comply with the requirements concerning the release of primary aromatic amines and formaldehyde into food".
- **verbatim** "Primary aromatic amines (hereinafter ‘PAA’) are a family of compounds, some of which are carcinogenic, while others of these compounds are suspected carcinogens."
- **verbatim** "Polyamide and melamine plastic kitchenware originating in or consigned from China and Hong Kong shall be imported into the Member States only if the importer submits ... a declaration ... confirming that it meets the requirements concerning the release of primary aromatic amines".
- **verbatim** "as regards polyamide kitchenware, analytical results demonstrating that they do not release into foods or food simulants primary aromatic amines in a detectable quantity. ... the detection limit for primary aromatic amines is set at 0,01 mg/kg food or food simulants"; checks include "laboratory analysis of 10 % of consignments".
- Supports: nylon (polyamide) utensils are the kitchenware the EU singled out for import checks. The 0.01 mg/kg figure is single-source and stays out of the JSON. I did not read the current text of Regulation 10/2011, so the register does not state today's PAA limit.

**BfR Opinion No. 036/2019, "Polyamide Kitchen Utensils: Keep contact with hot food as brief as possible"** (`bfr_polyamide_2019`). https://www.bfr.bund.de/cm/349/polyamide-kitchen-utensils-keep-contact-with-hot-food-as-brief-as-possible.pdf (direct curl, 200; pdftotext). Found through the Internet Archive CDX index of bfr.bund.de URLs. Confidence high.
- **verbatim** "Cooking spoons, spatulas or whisks: polyamide (PA) kitchen utensils provide valuable baking, roasting and cooking assistance. However, components of this plastic can migrate from the utensils into the food and consequently be ingested by consumers."
- **verbatim** "the amount of 5 mg/kg of food was assessed as being toxicologically acceptable as group migration value"; "In 23 out of 33 items the group migration of cyclic PA oligomers originating from kitchen utensils, investigated in 2016/2017, was less than 5 mg/kg of food. However, in 10 of the 33 items the release exceeded 5 mg/kg of food."
- **verbatim** "The BfR recommends that consumers keep contact with food as brief as possible when using PA kitchen gadgets, especially at high temperatures (above 70 °C)."
- Supports: the nylon food-contact judgement. The 70 °C and 10-of-33 figures are from the regulator's own opinion (single source, which the brief allows for a governing standard); listed as single-source in the report.

### Academic

**Liu, Brandsma, Schreder 2024, "From e-waste to living space: Flame retardants contaminating household items add to concern about plastic recycling"**, Chemosphere 365:143319, PMID 39271080 (`liu_2024_black_plastic`). PubMed E-utilities plain-text abstract. Confidence high for what the abstract says.
- **verbatim** "A total of 203 products were screened for bromine (Br), and products containing >50 ppm Br were analyzed for BFRs, OPFRs, and plastic polymers"; "FRs were found in 85% of analyzed products, with total FR concentrations ranging up to 22,800 mg/kg."
- **verbatim** "Plastic typically used in electronics (styrene-based) contained significantly higher levels of ∑FRs than plastics less typically used for electronics (polypropylene and nylon)."
- **verbatim** "Estimation of exposure to BDE-209 from contaminated kitchen utensils indicated users would have a median intake of 34,700 ng/day".
- PubMed lists two errata: "Chemosphere. 2025 Feb;370:143903" and "Chemosphere. 2025 Sep;385:144552". Crossref gives both titles as "Corrigendum to ‘From e-waste to living space ...’"; I could not read the corrigendum text.
- Note: the 85% is of the products that passed the bromine screen, not of all 203. ATK's article says "85 percent of all black plastic household products tested", which overstates the base. The register follows the abstract.

### Specialist reviewers who test (America's Test Kitchen, raw HTML via curl; URLs from ATK's own `sitemap/reviews.xml` and `sitemap/articles.xml`)

**ATK, The Best Nonstick-Safe (Plastic and Silicone) Spatulas, Bromberg and Brightwell, 24 June 2025** (`atk_nonstick_spatulas`). https://www.americastestkitchen.com/equipment_reviews/2047-nonstick-safe-spatulas. Confidence high.
- **verbatim** "when we’re using nonstick cookware, we prefer to use spatulas with heads (flat surfaces) made from nylon, silicone, or other nonmetal materials. That’s because they’re gentler and less likely to scratch nonstick surfaces."
- **verbatim** "We prefer models with thin heads measuring from 1.3 to 1.8 millimeters at their thickest."
- **verbatim** "Spatula heads made from nylon and resin/fiberglass are almost always thinner and more flexible than silicone heads."
- **verbatim** "Silicone alone is too soft and floppy to support food, so manufacturers reinforce silicone spatula heads with metal or fiberglass frames. ... Silicone heads measured between 2.5 and 7 millimeters at their thickest ... The silicone models we tested all withstood temperatures up to 600 degrees without melting or warping, whereas the nylon and resin composite spatulas all began to melt when left in a pan heated to between 400 and 500 degrees. Still, performance trumped durability for our testers, who universally agreed that they’d rather be more careful with a thin nylon spatula than fight to remove food with a clumsily thick but more heat-resistant silicone spatula."
- **verbatim** winner (Matfer Exoglass) "Made from a resin/fiberglass composite ... it melted at 450 degrees"; another nylon model "wasn’t heatproof past 410 degrees"; another "melted, albeit at a comparatively high temperature of 500 degrees".
- **verbatim** "water got under the silicone coating and visibly rusted the metal core within."
- **verbatim** "Most testers had a slight preference for heads that were fairly long and tapered toward the handle—the shape of a traditional metal fish spatula."
- **verbatim** FAQ: "Yes—as long as they're made from virgin (new) plastic, not recycled plastic".
- Supports: the head-material trade-off, nylon melt range, silicone heat tolerance and thickness, fish-spatula shape, silicone core rust.

**ATK, The Best Metal Spatulas, Brightwell and Bromberg, 25 June 2025** (`atk_metal_spatulas`). https://www.americastestkitchen.com/equipment_reviews/1913-metal-spatulas. Confidence high.
- **verbatim** "We prefer fish spatulas to conventional metal spatulas because they have long, slim heads that are thin and flexible and relatively short, easy-to-control handles."
- **verbatim** "Often referred to as a turner or flipper—likely to avoid confusion with thicker silicone spatulas that are used to fold and scrape—it’s used to flip or transfer foods whenever we’re working with metal cookware or bakeware. (When cooking in more delicate nonstick pans, we prefer plastic spatulas .)"
- **verbatim** "Most of the fish spatulas we tested had heads that were from 0.8 to 0.9 millimeters thick"; conventional heads "too rigid and thick (measuring more than 1 millimeter)"; one "too flexible and thin (0.2 millimeters), and it buckled".
- **verbatim** "Models made entirely of stainless steel are safe to clean in the dishwasher. Wood-handled spatulas should always be hand-washed".
- Supports: fish spatula over conventional turner; metal for metal cookware; thickness numbers (notes only); dishwasher note.

**ATK, The Best Silicone Spatulas, Levine, Sandler and Brightwell, 17 December 2025** (`atk_silicone_spatulas`). https://www.americastestkitchen.com/equipment_reviews/1748-silicone-spatulas. Confidence high.
- **verbatim** "Each model has a solid internal frame made of nylon, steel, glass, or silicone ... The frame is sheathed in a layer of soft, flexible silicone."
- **verbatim** "In general, we preferred models made from one seamless piece of silicone. These were more durable and easier to clean than spatulas with discrete plastic handles, as they had no nooks or crannies for water or food remnants to get trapped in."
- **verbatim** "Another model collected dishwater in a small notch where its head met the handle; it failed to dry, and that water poured out into freshly whipped cream as we folded it."; "The two-part design traps food and water, and the handle melted in our heat test."
- **verbatim** "The spatulas we tested were rated safe to use up to a range of about 450 to 600 degrees Fahrenheit."
- **verbatim** "Lighter-colored heads showed stains, so we generally preferred darker models".
- **verbatim** "Top-performing models had heads with thicker centers and thinner edges".
- Supports: seamless construction, silicone heat ratings, scraper design.

**ATK, The Best Offset Spatulas** (`atk_offset_spatulas`). https://www.americastestkitchen.com/equipment_reviews/1630-offset-spatulas. Confidence high.
- **verbatim** "These useful spatulas have long, thin, round-tipped blades that are offset (or bent at an angle) away from their handles. The angle creates enough clearance to keep your fingers free of whatever you’re spreading".
- **verbatim** "Most offset spatulas have stainless-steel blades with handles made of plastic or wood."
- **verbatim** "Manufacturers usually specify the overall length of the blades, rather than that of the flat portions".
- Single-source thresholds kept here: blades "no more than 1.25 inches wide and about 8 inches long"; angle "about 35 degrees" (the winner's is 28); thickness "less than 1 millimeter".
- Supports: offset spatula definition and its job.

**ATK, The Best Kitchen Tongs, Phares, López and Li, 31 July 2026** (`atk_kitchen_tongs`). https://www.americastestkitchen.com/equipment_reviews/1859-kitchen-tongs. Confidence high.
- **verbatim** "There are two basic types of tongs: coated and uncoated. On coated tongs, the heads are typically coated with silicone, making them a good choice for use with nonstick cookware because they won’t scratch the coatings. Uncoated tongs have bare metal heads, often made from stainless steel."
- **verbatim** "Uncoated tongs with bare metal heads are our go-to for everyday use. When using nonstick cookware, reach for tongs with coated heads."
- **verbatim** "We found that tongs with 12-inch arms were the most useful and versatile for home cooks. ... A pair of smaller 9-inch tongs is also great for tasks involving low heat or no heat, such as tossing a salad".
- **verbatim** "Scalloped edges won, hands down."; "Tongs with this design [broad, flat heads] offered no grip whatsoever".
- **verbatim** MIT test: "The tongs needed between 0.44 and 0.84 pounds of force to close." (single source, notes only).
- **verbatim** on coated tips: "The coated pincers didn’t grip as well as the uncoated version" on a roast, but "the silicone held the ceramic ramekins more securely".
- Supports: tong length bands, coated versus bare, scalloped heads.

**ATK, The Best Grill Tongs, Bromberg and McManus, 19 June 2026** (`atk_grill_tongs`). https://www.americastestkitchen.com/equipment_reviews/1806-grill-tongs. Confidence high.
- **verbatim** "we tested a variety of tongs specifically designed for use with grills, each with arms measuring about 14 to 21 inches in length."
- **verbatim** "we preferred models with arms on the shorter side of the spectrum, measuring about 16 inches. This length was enough to keep our hands out of harm’s way but provided better control than longer pairs".
- **verbatim** "If grill tongs are too short, your hands get uncomfortably hot. Too long, and the tongs are hard to control."
- Supports: 16 in for grilling; the tong-length trade-off.

**ATK, The Best All-Purpose Whisks** (`atk_allpurpose_whisks`). https://www.americastestkitchen.com/equipment_reviews/2449-all-purpose-whisks. Confidence high.
- **verbatim** "There are two basic styles of all-purpose whisk: French whisks have narrower heads, while balloon whisks have more bulbous heads. Many whisks on the market are hybrids that fall somewhere between the two styles. Ultimately, the whisk’s style isn’t the most important factor in its success".
- **verbatim** "Narrow French whisks were especially easy to maneuver ... Wider balloon whisks covered more territory in a single swoop ... But they sometimes felt clumsy and awkward, especially for detail work."
- **verbatim** "At Least 10 Moderately Flexible, Evenly Spaced Wire Loops : The more wire loops the whisk had, the more even and consistent the coverage it provided ... With whisks that had at least 10 loops, there were never any gaps bigger than half an inch between the wires".
- **verbatim** "Several of the whisks we tested had hooks or rings at the top of the head ... they just served to collect cheese sauce or whipped cream, requiring finicky cleaning afterward."
- Single-source thresholds kept here: length "11 to 12 Inches"; "our top three choices all weigh less than 4 ounces".
- Supports: whisk shape trade-off; loop count; the judgement that shape matters less than build.

**ATK, The Best Nonstick Whisks** (`atk_nonstick_whisks`). https://www.americastestkitchen.com/equipment_reviews/2253-nonstick-whisks. Confidence high.
- **verbatim** "They’re built like traditional wire whisks , but each loop is coated in silicone to be gentler on pans."
- **verbatim** "their manufacturer-provided maximum temperature ratings ranged from 450 to 600 degrees Fahrenheit (this is because not all silicone is identical; it can have different compositions and properties). However, none of the whisks’ silicone loops melted in hot pans during normal use."
- **verbatim** "Our highest-rated whisks had eight loops, while the least effective had just four."
- **verbatim** "a loop detached from the handle of one whisk as it worked the firm butter"; "the dual-purpose flat-and-balloon whisk, whose multipart handle held water."
- Supports: silicone heat ratings (second ATK test); loop count; joints fail.

**ATK, The Best Flat Whisks, 2025** (`atk_flat_whisks`). https://www.americastestkitchen.com/equipment_reviews/1677-flat-whisks. Confidence high.
- **verbatim** "Sometimes called a gravy or roux whisk, the flat whisk excels at helping you produce any kind of sauce that you might make in a skillet or saucepan."
- **verbatim** "One type has plain metal tines (loops) and is meant for use with metal cookware ... The other type has metal tines that are coated in silicone. This type is best for use with nonstick cookware ... we don’t recommend them for use with other types of cookware."
- Supports: flat whisk for skillets; bare metal on metal cookware.

**ATK, The Best Ladles** (`atk_ladles`). https://www.americastestkitchen.com/equipment_reviews/2402-ladles. Confidence high.
- **verbatim** "if you’re concerned about scratching your enameled or nonstick cookware, we recommend getting a ladle with a silicone or plastic bowl. These materials will be gentler on any surfaces they touch, though they can melt if exposed to high heat. We tested both stainless-steel and nonstick-safe versions, all with capacities of 4 to 7 ounces, the most commonly available sizes."
- **verbatim** "A 5-Ounce Capacity: We found that bowls this size were ideal for dishing out soups and stews."
- **verbatim** "Floppy silicone and nylon bowls buckled and bounced under the weight of heavy stew".
- **verbatim** "We also particularly liked handles made from a single piece of metal, silicone, or plastic, as they had no seams to collect debris."
- **verbatim** "several models we tested melted along their handles when left in a hot pot".
- Single-source details kept here: pouring rim preferred; handles 9.5 to 10 in at about 45°.
- Supports: ladle capacity, enamelled cookware in the nonstick question, one-piece handles.

**ATK, The Best Ice Cream Scoops, 2026** (`atk_icecream_scoops`). https://www.americastestkitchen.com/equipment_reviews/1595-ice-cream-scoops. Confidence high.
- **verbatim** "this scoop’s hallmark technology is the heat-conducting fluid (which a representative from the company told us was food-grade mineral oil) inside the body, which transfers heat from your hand to slightly soften the ice cream as you scoop. Aluminum itself is also a great conductor of heat."
- **verbatim** "Some people may reach for a portion scoop to scoop ice cream, but their thin metal levers are more prone to breaking when used on firm, frozen ice cream or sorbet; we recommend leaving those for portioning cookies and muffins."
- **verbatim** "If you’re solely scooping from pint containers, a 2- to 3-ounce scoop will fit best".
- **verbatim** "All the scoops we tested had heads made of metal—either aluminum, zinc, or stainless steel".
- **verbatim** "One ice cream scoop had a unique lever that hinged forward to propel the ice cream out of the scoop. It was unnecessary".
- Supports: ice cream scoop versus disher; heat-conducting scoops; scoop heads are metal.

**ATK, The Best Portion Scoops, Bromberg and Brightwell, 11 June 2025** (`atk_portion_scoops`). https://www.americastestkitchen.com/equipment_reviews/2157-small-and-medium-portion-scoops. Confidence high.
- **verbatim** "To operate them, you apply pressure to a spring-loaded handle, and a system of gears turns a thin, curved strip of metal inside each disher’s bowl."
- **verbatim** "Thumb-press scoops operate using a lever that extends diagonally outward from the handle ... Squeeze-style scoops have triangular handles".
- **verbatim** "scoops are given numerical “sizes” based on the number of level scoops it would take to empty a quart. It would take 20 level scoops with a #20 size scoop, for example, and each scoop would hold about 3.2 tablespoons."
- **verbatim** "manufacturers play fast and loose with this numbering system. We found that scoops of the same “size” could have capacities that varied by as much as a tablespoon. To get a portion scoop with the capacity we actually wanted, it was much more useful to look for information on the scoop’s volume, which each manufacturer usually lists alongside the inconsistent and unreliable numerical size."
- **verbatim** "scoops with listed capacities of about 3 tablespoons, which often corresponded to a #24 size scoop—the most common size".
- **verbatim** "Squeeze-style scoops were easier for testers of different hand sizes to operate."; thumb-press levers "jutted too far outward—45 degrees or more".
- Winner names give the cookie-size bands: "Best 3-Tablespoon Scoop OXO Good Grips Large Cookie Scoop", "Best 1.5-Tablespoon Scoop OXO Good Grips Medium Cookie Scoop", "Best 1-Tablespoon Scoop OXO Good Grips Small Cookie Scoop".
- Supports: disher number definition, unreliability, squeeze over thumb, OXO small/medium/large volumes.

**ATK, "Are Black Plastic Kitchenwares Safe to Use?", 9 December 2024** (`atk_black_plastic`). https://www.americastestkitchen.com/articles/8464-black-kitchenwares-toxic. Confidence medium (journalism summarising a study).
- **verbatim** "This study, and others we looked at, indicates that the products at fault were made with recycled plastic."
- **verbatim** "A review of the study in Chemosphere by scientists at McGill University identified a math error in the results published within the study."
- **verbatim** "Currently, we recommend getting rid of any kitchen utensil made from recycled black plastic. If your kitchen utensil is made from virgin (new) black plastic, however, it might still be fine to use".
- **verbatim** "(OXO tells us that while they are committed to using more recycled materials generally, they use no recycled plastic in food-contact applications.)"
- **verbatim** "Spatulas, ladles, and spoons made from silicone usually have thicker edges than do nylon or ABS models too."
- **verbatim** "Because stainless steel is so strong, you don’t need as much of it to create a ladle or spatula with a thin bowl or head ... metal tools aren’t great for use with nonstick or enameled cookware".
- Supports: recycled versus virgin plastic; the disagreement; silicone thicker (second ATK page).

**ATK, "Can You Clean Wooden Spoons in the Dishwasher?"** (`atk_wooden_spoons`). https://www.americastestkitchen.com/articles/2326-can-you-clean-wooden-spoons-in-the-dishwasher. Confidence high.
- **verbatim** "After 10 trips through the dishwasher, most spoons looked bleached and felt as dry as driftwood."; "Wash your wooden spoon by hand if you want to prolong its lifespan."
- Supports: wood heads are hand-wash (applied to wooden spatulas).

### Second opinions on nonstick and utensils

**Consumer Reports, "You Can't Always Trust Claims on 'Non-Toxic' Cookware"** (`cr_nonstick`). https://www.consumerreports.org/toxic-chemicals-substances/you-cant-always-trust-claims-on-non-toxic-cookware-a4849321487/ (curl). Confidence high.
- **verbatim** "Just be sure not to scrape it with metal utensils or steel wool, which can abrade the surface".
- Supports: nonmetal utensils on nonstick (second source after ATK).

**The Rational Kitchen, "Ceramic Frying Pans: Are They Better than Teflon (PTFE)?"** (`rational_kitchen_ceramic`). https://therationalkitchen.com/ceramic-frying-pans-better-than-teflon/ (curl). Confidence medium.
- **verbatim** "ceramic nonstick and PTFE both require low-medium heat, hand washing, and plastic or wooden utensils in order to get the longest possible life span out of the nonstick coating".
- **verbatim** "And while many makers say you can use their pans with metal utensils and put them in the dishwasher, you shouldn't do either of these if you want your ceramic nonstick pan to last as long as possible".
- Supports: third source for nonmetal on nonstick; the disagreement with makers who say metal-utensil safe.

**McGill Office for Science and Society, "Are Black Plastic Spatulas and Serving Spoons Safe to Use?"** (`mcgill_oss_black_plastic`). https://www.mcgill.ca/oss/article/critical-thinking-health-and-nutrition/are-black-plastic-spatulas-and-serving-spoons-safe-use (found via the OSS sitemap). Confidence high.
- **verbatim** "For a 60-kg adult, they calculated this to be 42,000 nanograms a day and opined that this was frighteningly close to the 34,700 nanograms of daily intake. But they made a mathematical error: 7,000 times 60 is not 42,000, but 420,000. The supposed exposure is not close to the reference dose. It is one-tenth of it."
- **verbatim** "For me, this risk would not be enough to discard a black plastic kitchen item if I had one ... That being said, there should be no flame retardants in such items."
- **verbatim** "immersion of bits of black plastic in olive oil at 160 C for 15 minutes was found to transfer flame retardants to the oil."
- Supports: the black-plastic judgement and its disagreement with ATK.

### Spec sheets (fields brands list)

**Vollrath foodservice product pages** (URLs from `vollrath.com/sitemap.xml`; curl). Confidence high for each product's own claim.
- Squeeze disher (`vollrath_squeeze_disher`) https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/dishers/color-coded-squeeze-dishers/47397. **verbatim** "1.52-ounce disher with red squeeze handle"; spec rows "Size | 24", "Capacity (OZ) | 1.52", "Color | Red"; "simple two-part construction that eliminates the concern over the loss of parts"; "Color-coded handle allows for quick size identification"; "NSF-certified"; "Not intended for hard or frozen foods". Sibling sizes on the page: "2.4-Ounce Disher With Sky Blue Squeeze Handle", "2.8-Ounce Disher With Green Squeeze Handle".
- Thumb disher (`vollrath_thumb_disher`) https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/dishers/nsf-certified-dishers-one-piece-color-coded-han/47147. **verbatim** "0.75-ounce one-piece thumb disher with purple handle"; "Molded, one-piece handle provides additional durability, prevents unsanitary food traps"; "Dishwasher-safe for easy cleaning"; "Not recommended for ice cream or other hard, frozen foods".
- Nylon turner (`vollrath_nylon_turner`) https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/turners/high-heat-nylon-turner/5284120. **verbatim** "The strong, one-piece nylon construction doesn’t scratch or mar surfaces, is heat-resistant up to 475°F and features a smooth design that eliminates crevices that trap bacteria."; "This durable nylon spoon is ideal for nonstick cookware."
- Silicone spatula (`vollrath_silicone_spatula`) https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/spatulas-softspoons-whips/high-temperature-spatula-utensils/52026. **verbatim** "Both the silicone blade and nylon handle are heat-resistant up to 500°F, and they are molded together to ensure easy cleaning and prevent unsanitary food traps."
- Nylon-end tongs (`vollrath_nylon_tongs`) https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/tongs/high-temperature-nylon-end-one-piece-tongs/4781612. **verbatim** "16-inch high-temperature nylon-end one-piece tongs"; "The nylon end won’t scratch or mar nonstick surfaces and is heat-resistant up to 450°F"; "One-piece design prevents unsanitary food traps".
- Ladle (`vollrath_ladle`) https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/ladles/one-piece-heavy-duty-ladles/4982410. **verbatim** "24-ounce one-piece heavy-duty stainless steel ladle"; "Capacity stamp on the handle allows for quick identification"; "One-piece design is easy to clean".
- French whip https://www.vollrathfoodservice.com/products/smallwares/kitchen-essentials/spatulas-softspoons-whips/stainless-steel-french-whips/47284. **verbatim** "The elongated design is ideal for deep vessels and straight-sided pans"; lengths offered "12-inch", "14-inch", "18-inch", "22-inch". No wire count is listed. Used only in notes (supports French whisk for deep pots alongside Wikipedia; not cited in JSON to keep the reference list short). 
- Field order Vollrath lists: dimensions, size (disher number), capacity in cup and oz, weight, colour, construction material, handle colour, shape. Capacity and colour sit in spec rows; the disher number sits in a "Size" row.

**Zeroll** (`zeroll_scoop`, `zeroll_care`). https://zeroll.com/products/model-1020-zeroll-aluminum-ice-cream-scoop and https://zeroll.com/pages/care-use (curl; URLs from the Shopify sitemap). Confidence high.
- **verbatim** "Constructed of sturdy aluminum alloy, this ice cream scoop has non-toxic, heat-conducting fluid in the handle that uses the warmth of your hand to make scooping easy and fast."; "The easy-to-use one-piece design has no springs to replace."; "this #20 ice cream scoop features a gold end cap for a portion size of 2 ounces"; "Do not place this ice cream scoop in the dishwasher; wash it by hand only."; "Do not expose to temperatures above 140°F".
- Care page **verbatim** "Not Dishwasher Safe"; "Do Not Expose to temps in excess of 140F".
- Supports: heat-conducting scoop, not dishwasher-safe, Zeroll numbering differs (#20 = 2 oz against the standard 1.6 oz).

**GIR** (`gir_tools`). https://www.gir.co/products/ultimate-spatula (curl). Confidence medium (marketing copy).
- **verbatim** "Heat-Resistant to 425°F"; "Made from stainless steel and premium-grade silicone"; "Platinum-Grade Silicone".
- Supports: the low end of silicone heat ratings (425 °F, below ATK's 450 to 600 °F range).

### Primers

**Wikipedia, "Scoop (utensil)"** (`wikipedia_scoop`). https://en.wikipedia.org/wiki/Scoop_(utensil) (curl). Confidence medium (the article flags "citation needed" on the table).
- **verbatim** "there is a clear distinction among three types of scoop: the disher , which is used to measure a portion (such as cookie dough ) ...; the ice cream scoop ; and the transfer scoop which is used to measure or to transfer an unspecified amount of a bulk dry foodstuff such as rice, flour, or sugar."
- **verbatim** "Traditional dishers are sized by volume either by the number of scoops per quart , or by its volume in fluid ounces or tablespoons , or the diameter of the tool's bowl."
- **verbatim** "The table below is the standard definition [ citation needed ] in the US food industry, but actual capacity varies by manufacturer."
- Table rows used (colour, number, fl oz): Gray 8, 4.0; Green 12, 2.7; Sky Blue 14, 2.4; Royal Blue 16, 2.0; Yellow 20, 1.6; Red 24, 1.3; Black 30, 1.1; Orchid 40, 0.8; Plum 70, 0.46.
- **verbatim** caption: "Zeroll scoops use a non-standard color coding, where the size # is the number of double scoop servings in a gallon".
- Supports: three scoop types; disher numbering; the colour code (matches Vollrath: red 24, green 12, sky blue 14); Zeroll non-standard.

**Wikipedia, "Whisk"** (`wikipedia_whisk`). https://en.wikipedia.org/wiki/Whisk (curl). Confidence medium.
- **verbatim** "Balloon whisks are best suited to mixing in bowls, as their curved edges conform to a bowl's concave sides."; "the French whisk has a more cylindrical profile, suiting it to deep, straight-sided pans."; "A flat whisk has the loops arranged in a flat successive pattern. It is useful for working in shallow vessels like skillets"; ball whisk "easier to clean than traditional looped varieties"; twirl whisk "is not suitable for whisking in the traditional sense ... but is ideal for beating eggs."
- **verbatim** "The wires are usually metal, but some are plastic for use with nonstick cookware."
- Supports: whisk shape values and the vessel mapping.

### Could not reach or did not use

- WebSearch: no calls left in the shared session budget.
- OXO (403), Rubbermaid Commercial (403), WebstaurantStore disher guide (403). Matfer Bourgeat's US home page loaded; I did not look for the Exoglass spec page, so ATK's melt test (450 °F) is the only Exoglass figure here.
- Internet Archive: went offline during the run, so no Amazon or Walmart snapshot was read.
- legislation.gov.uk (empty 202 bot challenge).
- The two Chemosphere corrigenda (titles only, via Crossref).
- ATK offset turner review (https://www.americastestkitchen.com/equipment_reviews/2071-offset-turners) and grill spatula review were fetched but are not cited: they add diner-style and grill turners without a new facet.

## Where "on listing" comes from

Retail marketplaces were not fetched (see constraints). Facets marked `listing` rest on product names, which are what marketplace titles carry: brand product names on fetched pages (Vollrath "16-inch high-temperature nylon-end one-piece tongs", "1.52-ounce disher with red squeeze handle", "24-ounce one-piece heavy-duty stainless steel ladle", "13.5-inch slotted nylon turner"; Zeroll "Zeroll Original 2 oz Ice Cream Scoop") and the product names ATK prints for every tested model ("OXO Good Grips 12-Inch Tongs", "OXO Good Grips 16" Locking Tongs", "Wüsthof Gourmet 7" Slotted Fish Spatula", "OXO Good Grips 11" Balloon Whisk", "OXO Good Grips Flat Whisk", "Cuisipro Silicone Flat Whisk—10 inch", "Cuisinart Curve Handle Line Curve Nylon Ladle", "Rösle Hook Ladle with Pouring Rim", "Di Oro Living Seamless Silicone Spatula—Large", "Rubbermaid 13.5" High-Heat Scraper", "OXO Good Grips Bent Icing Knife", "OXO Good Grips Large Cookie Scoop", "Dexter-Russell Sani-Safe 8 x 3" Cake Turner"). From those names: tool type, spatula style (fish, turner, scraper, icing/offset), head material (nylon, silicone, stainless), tong and whisk length, whisk shape (balloon, flat, French) and scoop capacity are in the title text. Facets whose value sits in bullets or spec rows (`max_heat_temp`, `dishwasher_safe`, `ladle_capacity`, `disher_number`, `release_mechanism`, `seamless_construction`) are marked `researched` with low backfill. OXO cookie scoops are named only Small, Medium and Large, so `portion_capacity` needs a lookup for that brand; the hint says so.

## Data points used for derived metrics

`nominal_portion = 32 / disher_number` (US fluid ounces; a US quart is 32 fl oz).

| Check | Inputs | Result | Source |
|---|---|---|---|
| ATK's worked example | #20 | 32 / 20 = 1.6 fl oz = 3.2 tbsp | ATK portion scoops: "each scoop would hold about 3.2 tablespoons" |
| Wikipedia chart row | #24 | 1.33 fl oz | Wikipedia table: "Red 24 ... 1.3" |
| Wikipedia chart row | #16 | 2.0 fl oz | Wikipedia table: "Royal Blue 16 ... 2.0" |
| Real Vollrath #24 | spec row "Size 24", "Capacity (OZ) 1.52" | 1.52 against 1.33 nominal, about 0.38 tbsp over | Vollrath squeeze disher page |
| Zeroll #20 | "#20 ... portion size of 2 ounces" | 2 oz against 1.6 nominal; Zeroll's scale is not the chart's | Zeroll model 1020 page; Wikipedia caption "Zeroll scoops use a non-standard color coding" |

So the metric is reliable as a nominal figure for food-service dishers and wrong for Zeroll ice cream scoops. The register says both.

## Against the existing add-on schema

Read: `supplies/kitchen/add-ons/scoop.jsonc` (one field: `sizing.disher_size_number`), composed with `kitchen-standard.jsonc` (`materials[]` with material and role, `max_heat_temp` with f/c unit, `dishwasher_safe` enum yes/top_rack_only/no/unspecified) and `supplies-standard.jsonc` (`high_signal_markdown`, `sizing.size_stated`, `pack_count`, `pieces_included_text`, `dimensions`, `capacity`, `weight`, `measurements_text`). Kinds below are the four from `zamana-expertise/AGENTS.md` rule 4 (vertical anchor, category quality currency, hard filter, sizing) plus "visual" (rule 1) and "markdown" (high_signal_markdown) for fields that should not be columns. Proposals only; nothing in zamana-expertise was edited.

| Facet in this register | In schema today | Proposed kind | Note |
|---|---|---|---|
| tool_type | no | hard filter | Proposal: `utensil_type` enum [spatula, ladle, tongs, whisk, scoop]. The leaf bundles five tools; every other field depends on it. Outside `sizing` (variant-invariant). |
| spatula_style | no | hard filter (spatulas only) | Proposal: `spatula_style` enum [fish_spatula, turner, scraper, offset_spatula]. Photos usually show it, but the word "spatula" means a scraper on some listings and a turner on others, so text extraction plus a photo tie-break. |
| nonstick_safe | no | derived at query time | From `materials[]` with role `working_head`: metal = no, everything else = yes. Do not store. |
| head_material | yes, `materials[]` role `working_head` | vertical anchor | Works today. Gap: no value for fiberglass-resin composite (Matfer Exoglass) beyond `composite`, which is fine. `aluminum` exists. `silicone` over a `core` is expressible with two entries (silicone working_head + stainless_steel core). |
| max_heat_temp | yes, kitchen standard | hard filter | Keep. Store bare metal as not applicable rather than null, so "unknown" and "no limit" differ. |
| head_thickness | no | not a column | Reviewer-measured only; no PDP states it (rule 3: Jev material at best). Keep in this register as a researched annotation for the top N. |
| seamless_construction | no | markdown | "One-piece" and "seamless" are construction claims; rule 5 puts curated claim vocabularies in `high_signal_markdown`. The comment there already asks for "durability and construction signals". |
| tong_length | partly: `sizing.dimensions.length` | sizing | Bought by number (9, 12, 16 in), so it belongs in sizing. The generic `dimensions.length` covers it; no new field needed, but extraction should prefer the named length in the title. |
| whisk_shape | no | visual, or hard filter | A photo shows balloon versus French versus flat (rule 1). Titles also state it. Leave to the image pipeline unless shape filtering is wanted on text alone. |
| wire_loops | no | visual | Countable from the photo (rule 1 and rule 5 "counts ... a photo answers"). Not a column. |
| ladle_capacity | partly: `sizing.capacity` (fl_oz) | sizing | Bought by number in food service ("24-ounce ladle", capacity stamped on the handle). The standard `capacity` field covers it. |
| scoop_type | no | hard filter | Proposal: `scoop_type` enum [ice_cream_scoop, disher, transfer_scoop]. A disher is "not intended for hard or frozen foods" (Vollrath), so this is a fit fact, not a style. |
| portion_capacity | partly: `sizing.capacity` | category quality currency (scoops) | The number bakers compare. Use `sizing.capacity` with unit fl_oz; add `tbsp` to the capacity unit enum, since OXO and ATK quote tablespoons. |
| disher_number | yes, `sizing.disher_size_number` | sizing, but demote | Keep the field because recipes and food service use it, but it should not be the scoop's only size: ATK found same-number scoops vary by up to a tablespoon, and Zeroll's numbering differs. Pair it with `capacity` and a brand note. |
| release_mechanism | no | hard filter or markdown | Proposal: `release_mechanism` enum [squeeze, thumb_lever, none]. A photo shows it, so it may be visual (rule 1); squeeze versus thumb mattered in ATK's test. |
| dishwasher_safe | yes, kitchen standard | hard filter | Keep. The trap the standard flags ("not dishwasher safe" contains "dishwasher safe") matters for Zeroll, whose page says "DO NOT PLACE IN DISHWASHER", which a plain phrase match would read as dishwasher-safe. |

Add-on and standard fields this register treats as noise for this leaf: `supplies-standard.sizing.weight` (ATK weighed whisks and grill tongs, but no buyer filters by it; it is a review detail); `pack_count` and `pieces_included_text` matter only for utensil sets, which are a minority here; `measurements_text` duplicates the named length and capacity. Certifications (NSF on Vollrath dishers) stay in markdown under rule 5.

## Branches

Question-to-value rules: split 1 offers its five values verbatim. Split 2 offers three answers: "flipping food" maps to `fish spatula` and `turner` (both stay in play), "scraping bowls" to `scraper`, "spreading icing" to `offset spatula`. Split 3 is a yes/no: yes filters `nonstick_safe` to yes; no leaves both values in play, so bare metal is allowed. Every `when` value is a value of its split facet (or yes/no for the boolean).

**Split 1, tool_type ("Spatula, ladle, tongs, whisk, or scoop?")**
- when `spatula`: then_ask spatula_style; makes_irrelevant tong_length, whisk_shape, wire_loops, ladle_capacity, scoop_type, portion_capacity, disher_number, release_mechanism.
  - then_ask: the four styles have opposite builds (ATK metal spatulas: turners "often referred to as a turner or flipper—likely to avoid confusion with thicker silicone spatulas that are used to fold and scrape"; ATK offset spatulas; ATK silicone spatulas).
  - makes_irrelevant: each facet is defined on another tool in the register ("The tongs' overall length", "How many wire loops a whisk has", "one level ladleful", "one level scoop", "on a disher", "How the scoop lets go"). Trivial.
- when `ladle`: then_ask nonstick_safe, ladle_capacity; makes_irrelevant spatula_style, head_thickness, tong_length, whisk_shape, wire_loops, scoop_type, portion_capacity, disher_number, release_mechanism.
  - nonstick_safe: ATK ladles "if you’re concerned about scratching your enameled or nonstick cookware, we recommend getting a ladle with a silicone or plastic bowl". ladle_capacity: ATK "A 5-Ounce Capacity ... ideal"; Vollrath sells by stamped capacity.
  - head_thickness is defined as "How thick a turner's head is". Trivial. Others trivial as above.
- when `tongs`: then_ask tong_length, nonstick_safe, tip_shape; makes_irrelevant spatula_style, head_thickness, whisk_shape, wire_loops, ladle_capacity, scoop_type, portion_capacity, disher_number, release_mechanism.
  - tong_length: ATK kitchen tongs (9 and 12 in) and grill tongs (16 in). nonstick_safe: ATK "There are two basic types of tongs: coated and uncoated ... coated ... a good choice for use with nonstick cookware".
  - **Added leg (2026-10-25):** tip_shape. ATK: "Scalloped edges won, hands down"; flat-tipped tongs "offered no grip whatsoever". Required fix 13: this facet was dropped in the original run for lacking a listing field, but the check asked for it back as a photo-derived annotation, since it was ATK's most decisive finding. Annotation only, not a filter, so it narrows nothing on its own; shown alongside the other tong facets.
- when `whisk`: then_ask whisk_shape, nonstick_safe, wire_loops; makes_irrelevant spatula_style, head_thickness, tong_length, ladle_capacity, scoop_type, portion_capacity, disher_number, release_mechanism.
  - whisk_shape: Wikipedia shape-to-vessel lines; ATK flat whisk for skillets. nonstick_safe: ATK nonstick whisks "each loop is coated in silicone to be gentler on pans"; ATK flat whisks "two main types". wire_loops: ATK all-purpose "At Least 10 ... Wire Loops"; nonstick "Our highest-rated whisks had eight loops, while the least effective had just four."
- when `scoop`: then_ask scoop_type, portion_capacity, release_mechanism; makes_irrelevant spatula_style, head_thickness, tong_length, tip_shape, whisk_shape, wire_loops, ladle_capacity.
  - then_ask: Wikipedia's three scoop types; ATK "buy by volume"; ATK squeeze versus thumb.
  - **Changed leg (2026-10-25):** nonstick_safe is no longer irrelevant for all scoops. The blind check found no source for blanket irrelevance, and ATK's portion-scoop page uses dishers for "muffin batter", which goes into muffin tins, often nonstick (check.md C, D-level finding). nonstick_safe now moves to the new scoop_type split's disher branch instead (see Split 4 below); ice cream and transfer scoops, which only touch tubs and bowls, still skip it by not being asked.
  - **Dropped leg (2026-10-25, before publishing):** max_heat_temp irrelevant on scoops. Zeroll states "Do not expose to temperatures above 140°F", which is a heat limit, so the facet still applies; it is covered in the register through dishwasher_safe.

**Split 4, scoop_type ("Ice cream, dough and batter, or flour and sugar?") — added 2026-10-25 after the blind check (check.md required fixes 3 and 7).**
- when `ice cream scoop`: then_ask portion_capacity.
  - ATK "a 2- to 3-ounce scoop will fit best" for pints; this is a volume question, not the cookie-size wording, which is why portion_capacity's cold_start_question was reworded to "Small, medium, or large scoop?" rather than "... cookies?" (it now also serves disher buyers for the same reason).
  - The facet's own definition keeps dishers eligible for "ice cream" answers too (ATK portion scoops names "muffin batter" and ice cream among disher uses), so this branch narrows the question, not the underlying tool.
- when `disher`: then_ask nonstick_safe, portion_capacity, disher_number, release_mechanism.
  - nonstick_safe: new leg, see the changed tool_type/scoop leg above. portion_capacity, disher_number, release_mechanism: same sourcing as before (ATK portion scoops; Wikipedia; Vollrath; Zeroll).
- when `transfer scoop`: makes_irrelevant portion_capacity, disher_number, release_mechanism.
  - The register's own scoop_type definition says a transfer scoop "measures nothing", so none of the three sizing or mechanism facets apply. This was the leg check.md's Part C table marked WEAK ("the register itself says a transfer scoop measures nothing, and there is no scoop_type branch to drop the facet for it"); this split is that missing branch.

**Split 2, spatula_style ("Flipping food, scraping bowls, or spreading icing?")**
- when `fish spatula`, `turner`: then_ask nonstick_safe, head_thickness.
  - nonstick_safe: ATK split its turner testing into a metal review ("whenever we’re working with metal cookware") and a nonstick-safe review ("when we’re using nonstick cookware, we prefer ... nonmetal materials"). head_thickness: ATK nonstick "thin heads measuring from 1.3 to 1.8 millimeters"; metal "0.8 to 0.9 millimeters".
- when `scraper`: then_ask seamless_construction, max_heat_temp; makes_irrelevant nonstick_safe, head_thickness.
  - seamless_construction: ATK silicone spatulas "we preferred models made from one seamless piece of silicone". max_heat_temp: ATK "rated safe to use up to a range of about 450 to 600 degrees", used for "high-heat cooking ... hot caramel".
  - nonstick_safe irrelevant: chain of two facts. (1) Scrapers are silicone-headed (ATK "Made using food-grade silicone"; "All silicone spatulas share the same basic design"). (2) Silicone heads are nonstick-safe (the register's nonstick_safe definition; ATK nonstick-safe spatulas). WEAK-acceptable: a hard plastic scraper could exist; none was in my sources.
  - head_thickness irrelevant: defined on turners. Trivial.
- `offset spatula`: no branch. **Dropped legs (2026-10-25):** "offset → nonstick_safe irrelevant" (ATK says offset spatulas "help loosen and remove brownies ... from baking pans", which may be nonstick, so the leg is not supported) and "offset → max_heat_temp irrelevant" (rests on my inference that icing tools never meet heat; no source says so).

**Split 3, nonstick_safe ("Will it touch nonstick or enamelled pans?")**
- when `yes`: then_ask head_material, max_heat_temp.
  - Chain of two facts. (1) On nonstick the choices are nylon, composite, silicone or wood (ATK nonstick spatulas; ATK black plastic "Silicone Tools", "Wood and Bamboo Tools"). (2) **Corrected 2026-10-25** (check.md A3, CONTRADICTED; this leg was WEAK in Part C): maker ratings for nylon and silicone overlap (Vollrath rates nylon 450 to 475 °F and a nylon handle to 500 °F; GIR rates silicone from 425 °F), so the rating alone does not separate them. Only ATK's hot-pan melt test does: nylon and composite melted at 400 to 500 °F, silicone did not melt up to 600 °F. max_heat_temp is still worth asking once metal is out, but the register no longer claims the maker rating is "the number that separates nylon from silicone".
- when `no`: then_ask head_material.
  - Bare metal becomes eligible and is what reviewers prefer on metal pans: ATK tongs "Uncoated tongs with bare metal heads are our go-to for everyday use"; ATK flat whisks silicone tines "we don’t recommend them for use with other types of cookware"; ATK metal spatulas for "cast iron, stainless steel, or carbon steel". 
  - **Dropped leg:** "no → max_heat_temp irrelevant". The buyer may still pick nylon or silicone on a metal pan, and those melt; not supported.

## Numbers in category.json and how many sources stand behind each

| Number | Sources | Status |
|---|---|---|
| Nylon heads melt around 400 to 500 °F | ATK melt test (400 to 500; 410, 450, 500 per model); Vollrath ratings 450 and 475 °F fall inside | two |
| Silicone rated about 425 to 600 °F | GIR 425; Vollrath 500; ATK silicone spatulas 450 to 600; ATK nonstick whisks 450 to 600 | two or more |
| Tongs 9, 12, 16 in; grill pairs tested to 21 in | ATK kitchen tongs (9, 12); ATK grill tongs (16, 14 to 21); Vollrath 16 in tongs; OXO product names "12-Inch", "9-Inch", "16"" as printed by ATK | 16 in: two. 9 and 12 in: one publisher plus OXO names seen only on ATK's page (listed as single-source) |
| Disher number = scoops per quart; #20 about 1.6 fl oz; #16 = 2; #24 = 1.33 | ATK; Wikipedia; arithmetic | two |
| Red #24, green #12 (sky blue #14) | Vollrath product pages `.../47397` (red #24) and `.../47393` (green #12, **added 2026-10-25**: the old citation only showed page 47397, which does not carry green #12 — check.md A6); Wikipedia chart | two |
| Vollrath #24 holds 1.52 oz | Vollrath spec row; compared against the chart | one brand fact, used as an example |
| Zeroll #20 holds 2 oz | Zeroll product page only | single-source, brand-only (**corrected 2026-10-25**: Wikipedia's caption confirms only that Zeroll's scale is non-standard; its own double-scoop-per-gallon rule would give 3.2 oz for a #20, so it is not a second source for 2 oz — check.md A7) |
| Same-number scoops differ by up to a tablespoon | ATK only | single-source (attributed) |
| Cookie scoops 1, 1.5 and 3 tbsp; 2 to 3 oz for pints | ATK only | single-source (attributed) |
| Home ladles 4 to 7 oz, 5 oz preferred | ATK only | single-source (attributed) |
| Vollrath ladle to 24 oz | Vollrath only | single brand fact |
| Zeroll keep below 140 °F | Zeroll product page and care page (same brand) | single-source |
| BfR: 70 °C; 10 of 33 items above 5 mg/kg | BfR opinion 036/2019 | regulator, single |
| Flame retardants in 85% of analysed products | Liu et al. abstract; ATK article (which misstates the base) | two |
| Exposure a tenth of the reference dose | McGill OSS (arithmetic shown); PubMed lists two corrigenda | one plus corroboration |

## Dropped

- **Tong lock type as a facet.** ATK's kitchen and grill tong reviews found locks the main usability difference (sliders that stick, a plug that "gradually worked its way back to the locked position", an all-metal lock that grinds, a gravity lock that "popped open or locked shut at the wrong moments"). One publisher and no brand field; kept in Trimmed detail.
- ~~Tong tip shape (scalloped, flat, toothed).~~ **Un-dropped 2026-10-25** (required fix 13): now the `tip_shape` facet, annotation role, in the tongs branch's then_ask. Still a single-source, photo-derived fact (AGENTS rule 1); kept because it was ATK's most decisive tong finding, not because a second source was found.
- **Ladle pouring rim and handle angle.** ATK only, and photo facts.
- **Whisk length (11 to 12 in) and weight (under 4 oz).** ATK only; Vollrath sells French whips from 12 to 22 in for food service, so no single band fits the catalog.
- **Spring force on tongs (0.44 to 0.84 lb).** One MIT measurement in one review.
- **Offset spatula blade size and angle.** ATK only ("no more than 1.25 inches wide and about 8 inches long", "about 35 degrees"), and ATK itself notes listings give overall rather than flat length.
- **Silicone "pinch test" for fillers.** No fetchable source; not used.
- **A recycled-plastic facet.** ATK's advice is to phone the maker; no listing states it. Kept as a judgement.
- **Current EU PAA limit under Regulation 10/2011.** I read only 284/2011's 0.01 mg/kg detection limit, which may have been superseded; kept out of the JSON.
- **NSF certification.** Vollrath dishers say "NSF-certified"; rule 5 puts certifications in markdown, and home buyers do not filter on it.

## Trimmed detail

- `head_thickness`: ATK numbers: preferred nonstick-safe heads 1.3 to 1.8 mm; silicone 2.5 to 7 mm; metal fish spatulas 0.8 to 0.9 mm; conventional metal turners over 1 mm; one at 0.2 mm buckled. Head area 12 to 15 sq in preferred; handles 4.75 to 5.5 in (nonstick) and 4.5 to 5 in (metal).
- `head_material`: ATK's nonstick winner is fiberglass-resin (Matfer Exoglass) and "melted at 450 degrees"; ATK's FAQ says such plastics are fine "as long as they're made from virgin (new) plastic". Lighter silicone heads stain; darker ones hide it (ATK).
- `seamless_construction`: Vollrath's two-part squeeze disher is sold as "simple two-part construction that eliminates the concern over the loss of parts", so the claim exists for mechanisms too.
- `tong_length`: grill tongs weighing 8 to 9 oz were preferred and over 12 oz tiring (ATK). Optional Kool Touch handles on Vollrath tongs are "color safe up to 225°F".
- `whisk_shape`: twirl or coil whisks are "not suitable for whisking in the traditional sense ... but ... ideal for beating eggs" (Wikipedia); ball whisks are "easier to clean than traditional looped varieties" (Wikipedia, unsourced in the article).
- `portion_capacity`: ATK tested scoops "with listed capacities of about 3 tablespoons, which often corresponded to a #24 size scoop—the most common size"; that is 1.5 fl oz, which sits between the chart's #20 and #24.
- `scoop_type`: ice cream is best scooped at about 5 °F, per ATK's science editor; most home freezers sit near 0 °F. Not a product attribute.
- Judgement on black plastic: the study screened 203 products, analysed those above 50 ppm bromine, and found styrene plastics far worse than nylon or polypropylene, which matters here because most black utensils are nylon.

## Corrections after the blind check (2026-10-25)

1. Added a `set` value to `tool_type`; changed its `backfill_cost` from `none` to `low` and added a `backfill_note`, because the leaf's top queries ("kitchen utensil set", "spatula set", "wooden spoon set") carry several tool nouns at once and matched no value (check.md B1-d, required fix 1).
2. Retracted "It is the number that separates nylon from silicone" from `max_heat_temp.why_experts_care` and split `range.typical` into a rated band and a measured melt band: maker ratings for nylon (450 to 475 °F, Vollrath) and silicone (425 to 600 °F, GIR/Vollrath/ATK) overlap, and only ATK's hot-pan melt test actually separates them (check.md A3, CONTRADICTED; required fix 2).
3. Added a fourth split on `scoop_type` with branches: ice cream → ask `portion_capacity`; disher → ask `nonstick_safe`, `portion_capacity`, `disher_number`, `release_mechanism`; transfer scoop → makes `portion_capacity`, `disher_number`, `release_mechanism` irrelevant. Removed `nonstick_safe` from `tool_type`'s scoop-branch `makes_irrelevant` list so the new disher leg can ask it (check.md C, two WEAK legs; required fixes 3 and 7).
4. Reworded `portion_capacity`'s `cold_start_question` to "Small, medium, or large scoop?" (was cookie-only) and its `backfill_note` to keep OXO's 1/1.5/3 tbsp line separate from generic marketplace 1/2/3 tbsp sets; fixed the OXO listing hint, since OXO's own Amazon bullet states the tablespoon figure directly and needs no size-table lookup (check.md D5 and B facet-level notes; required fixes 4 and 9).
5. Changed `tong_length`'s definition from "overall length" to "arm length, as reviewers measure it", added a note that most tong listings are 2- or 3-pair sets, and changed `backfill_cost` from `none` to `low` with a `backfill_note` on picking the length nearest the asked band (check.md D2; required fix 5).
6. Dropped the "Titles often add 'Nonstick-Safe'" listing hint from `nonstick_safe` and limited its derivation to `head_material` with role `working_head`, because a stuffed title labelled a bare-metal whisk "Non Stick" (check.md D4; required fix 6).
7. Softened `nonstick_safe.why_experts_care` so it no longer excludes all scoops: ice cream and transfer scoops still don't touch pans, but a disher now asks `nonstick_safe` (via the new `scoop_type` branch) because dishers fill nonstick muffin tins too (check.md C, WEAK leg "scoop → nonstick_safe irrelevant"; required fix 7).
8. Repointed the `gir_tools` reference from the redirecting `/products/ultimate-spatula` URL to the live `/products/spatula` page and retitled it. Refetched the new URL this run and confirmed "Heat-resistant to 425°F" and the silicone materials claim are both still there (check.md A2; required fix 8).
9. Changed `max_heat_temp`, `dishwasher_safe`, `disher_number` and `release_mechanism` from `source: "researched"` to `source: "listing"` with updated `listing_field_hint`s, since the checker's live Amazon sample found each on most listing bullets, care lines, spec rows or titles. Mapped marketplace "trigger" and "spring-loaded" wording to `release_mechanism`'s squeeze-handle value. Replaced `head_thickness`'s infeasible "backfill the top 50 turners" plan with "use ATK's own figures where they exist, else rank by head_material", since ATK's two turner reviews measured only about two dozen models (check.md D8 to D13; required fix 9).
10. Reworded judgement 8 to "85% of the products that passed the bromine screen (>50 ppm)" rather than "bromine-screened products", to match Liu et al.'s abstract rather than ATK's overstated base (check.md Part A note; required fix 10).
11. Added a new reference, `vollrath_squeeze_disher_12` (Vollrath page `.../47393`), for the green #12, 2.8 oz disher claim, since the previously cited page (`.../47397`) only shows red #24 and the green #12 sits on its sibling page. Reworded `disher_number`'s `why_experts_care` and `backfill_note` so the Zeroll #20 = 2 oz claim is attributed to Zeroll's own pages only; Wikipedia's caption confirms only that Zeroll's numbering scale differs, not the 2 oz figure, which its own rule would put at 3.2 oz (check.md A6, A7; required fix 11). Refetched `.../47393` this run: confirmed "Size 12", "Capacity (OZ) 2.8", "Color Green".
12. Four minor wording fixes (required fix 12): (a) `mechanism.plain`'s silicone-core sentence now reads "a core of nylon, steel, glass or fiberglass", matching ATK's own wording, not "a steel or nylon core". (b) Added `hybrid` to `whisk_shape.values` and reworded its `why_experts_care` and `listing_field_hint`, since ATK found many all-purpose whisks are balloon/French hybrids with no shape word in the name. (c) Reworded `dishwasher_safe`'s `cold_start_question` to "Hand-wash this tool, or dishwasher is fine?" and added a trap note about "DO NOT PLACE IN DISHWASHER" reading as a false positive on a plain phrase match. (d) Added the 21 CFR 177.1500 212 °F nylon-grade caveat to the CFR judgement, and an Internet Archive snapshot / unconfirmed-current-status caveat to the EU 284/2011 judgement.
13. Added a new annotation-only facet, `tip_shape` (scalloped, flat, toothed), wired into the `tongs` branch's `then_ask`, and repointed the "Scalloped tongs" vocabulary entry's `facet` field to it from `tong_length`. Previously dropped for lacking a listing field; kept now as a photo-derived annotation per the check's instruction (check.md required fix 13; see "Dropped" below, updated).

**Rejected finding.** check.md's "Script problems" section and Part A item 15 both say `schema.json`'s `vertical` enum excludes `"supplies"` and that `validate.py` rejects it. Re-running `validate.py` against the schema and category.json actually installed at `/Users/amnbh/.claude/skills/category-expertise/` shows the enum already lists `formulations, machines, fashion, furnishing, supplies, other`, and the file validates clean both before and after this run's edits. This register did not touch `schema.json` (out of scope), and the finding does not reproduce against the schema as it stands today; logged as a false positive, not fixed because there was nothing in scope to fix.

**Partially addressed, not fully resolved.** Required fix 1's cross-register conflicts — spider strainers and skimmers claimed by both this register and `strainers`; dough scrapers sent by `molds` to `knives` or `scoops` while this register's own scope sends them on to `kitchen-other`, a catch-all with no register; coffee scoops not excluded from `transfer scoop` — need `catalog.json` or a sibling register changed, which is outside `scoops/` and out of scope for this task. Added the `set` value (see item 1 above) as the one piece fixable from inside this register; the rest is logged here for the taxonomy owner, not fixed.

## Script problems

- The skill's `schema.json` `vertical` enum is `formulations, machines, fashion, furnishing, other`. The brief requires `"vertical": "supplies"`, which `validate.py` rejects with exactly one schema error ("'supplies' is not one of [...]"). I set `supplies` as instructed. A scratch copy with only that value changed to `other` validates clean: "16 facets, 3 splits, 33 references". Because the brief's build line chains the three scripts with `&&`, render and viewer were run directly after validate; both ran clean. Reported, not fixed, since the skill's files are out of scope. The sibling `kitchen-boards` run logged the same problem today.