Prose-style & consumer-language sweep — 2026-06-09 (batch 7, ingredients)
Counts: P1 0 · P2 20 · P3 18. Report-only; no pages were edited.
Pages scanned (25): milk-and-dairy.md, mixed-meals.md, molluscs.md, non-apple-fruit.md, non-dairy-creamer.md, non-rice-grains.md, non-root-vegetable-purees.md, non-root-vegetables.md, non-soy-infant-formula.md, non-soy-protein-source.md, noodles.md, oat-ring-cereal.md, oat.md, oatmeal.md, olive-oil.md, olives.md, onions.md, orange-juice.md, oranges.md, organ-meats.md, parsley.md, parsnips.md, pasta.md, peach.md, peanut-butter.md
Stubs skipped (5): milk-based-infant-formula.md (provisional_scaffold), mushrooms.md (provisional_scaffold), nectarine.md (provisional_scaffold), niger-seed-oil.md (provisional_scaffold), palm-oil.md (provisional_scaffold).
Clean pages (no findings): oatmeal.md, orange-juice.md, organ-meats.md.
Next run resumes at wiki/ingredients/peanut-oil.md.
Dominant pattern this batch
The same two systemic patterns continue: unanchored magnitude qualifiers and the bold-lever-label template pattern. Three observations specific to this batch:
Cascade qualifiers. milk-and-dairy.md:148 packs three unanchored qualifiers into a single paragraph (“generally a low-metal-load food commodity,” “Lead in milk is typically very low,” “Cadmium in milk is also typically very low”), each without a ppb anchor. Milk-and-dairy is a high-traffic page referenced by many downstream ingredient and product pages. The contamination profile table (immediately below) would anchor all three claims; they should draw from it.
Oil pages share a pattern. olive-oil.md, oranges.md, and onions.md each use “negligible metals” for extract/oil derivatives that carry the mechanism (selective lipophilic extraction) but no ppb data. These are low-concern from a defensibility standpoint, but the pattern is consistent enough to flag.
Sparse-data caveat on a risk label. non-apple-fruit.md:126 explicitly acknowledges “sparse monitoring data” in the same sentence that asserts “generally low-risk.” This is the most actionable finding in the P2 set — the sentence both makes a risk claim and undermines it with a data-quality caveat.
P1 — consumer safe/dangerous without anchor
| Page | Line | Offending text | Why | Suggested fix | Claim-adjacent? |
|---|---|---|---|---|---|
| (none found) | — | — | — | — | — |
P2 — bullets-in-prose & other qualifiers
| Page | Line | Offending text | Why | Suggested fix | Claim-adjacent? |
|---|---|---|---|---|---|
| milk-and-dairy.md | 148 | Dairy is generally a low-metal-load food commodity / Lead in milk is typically very low / Cadmium in milk is also typically very low | Three unanchored magnitude qualifiers in the same paragraph; contamination profile data (below the paragraph) would anchor all three | Anchor from the profile table: “Lead in fluid milk is typically below [X ppb] and Cd below [Y ppb] (per FSA/Fera and TDS data in profile below), placing dairy among the lower-metal commodity groups” | ⚠ yes — this is the lede paragraph of a page referenced heavily by infant-formula and dairy-product downstream pages |
| milk-and-dairy.md | 170 | plastic-bottle and glass-bottle milk have very low packaging-migration profiles | ”Very low” for packaging migration from plastic and glass without a ppb migration value | Rephrase: “plastic-bottle and glass-bottle milk have packaging-migration profiles at or below detection limits under normal storage conditions; no significant Pb or Cd contribution from these packaging types is documented” | no — mechanism is sound; “very low” is the trigger |
| non-apple-fruit.md | 126 | Tropical fruits (mango, pineapple, papaya) are generally low-risk based on the sparse monitoring data available | ”Generally low-risk” acknowledged as based on sparse monitoring data in the same sentence — the sentence simultaneously makes a risk claim and flags inadequate evidence | Rephrase: “Monitoring data for tropical fruits in the current corpus are sparse; preliminary results suggest lower Pb and Cd burdens than berry or grape categories, but this should not be treated as an established finding until more representative survey data are ingested” | ⚠ yes — risk label for a category consumed by infants, explicitly acknowledged as under-evidenced |
| non-dairy-creamer.md | 99 | refined vegetable oils carry very low metal concentrations / non-dairy creamer has one of the lower heavy metal profiles among processed food categories | ”Very low” for refined oil metals without a ppb anchor; “lower heavy metal profiles” characterizes the finished product without dose/population | Anchor: “refined vegetable oils carry metals at or below FDA TDS reporting limits (TDS below-detection for margarine/oil fractions); non-dairy creamer metals are similarly at or below detection (TDS n=3, all analytes ND)“ | no — both claims are supported by TDS below-detection data; “very low” and “lower” are the triggers |
| non-dairy-creamer.md | 151 | consistent with the expectation that this is a low-risk matrix | ”Low-risk matrix” characterisation without dose/population anchor; TDS data (all below detection) are cited in the same paragraph | Rephrase: “consistent with the TDS finding of all analytes below detection limits; no routine monitoring priority is indicated for this commodity” | no — TDS data cited in same paragraph; “low-risk” is the trigger |
| non-root-vegetables.md | 95 | non-root vegetables generally accumulate lower concentrations of heavy metals than root vegetables / tomatoes and cucumbers generally show lower metal concentrations than brassicas | Two unanchored comparative magnitude claims without ppb for either category in the same sentence | Anchor: “non-root vegetables carry Cd typically in the range of [X–Y ppb] vs root vegetables at [A–B ppb] per TDS data; tomatoes and cucumbers carry lower Cd than brassicas based on the available TDS-comparative data” | ⚠ yes — category-comparison claims that inform formulation decisions for baby-food and mixed-vegetable products |
| non-root-vegetables.md | 130 | the effect on Cd and Pb concentrations is modest because these metals are not primarily in the cytoplasmic water fraction | ”Modest” for blanching/steaming Cd/Pb reduction without a ppb or percentage anchor; mechanism is given | Rephrase: “the effect on Cd and Pb concentrations from blanching/steaming is typically less than 10–15 percent (based on general cereal and vegetable literature), because these metals are bound to cellular structures rather than dissolved in cytoplasmic water” — or note “magnitude not quantified for this category in the current corpus” | no — the mechanism is the main point; “modest” is the trigger |
| non-soy-infant-formula.md | 115 | Pb, Cd at low background levels reflecting forage and water inheritance / generally low metal loads for carbohydrate sources | ”Low background levels” and “generally low metal loads” without ppb for either pathway in this sentence | Anchor to the finished-formula corpus data: “the cow-milk ingredient contributes Pb typically below [X ppb] and Cd below [Y ppb] at the infant-formula inclusion ratio; carbohydrate sources (lactose, maltodextrin) carry metals near detection limits per TDS data” | ⚠ yes — this is the opening orientation paragraph for infant formula; these are Closer to Zero and EU ML compliance-relevant characterisations |
| noodles.md | 95 | Lead in wheat grain is generally low / Asian-style wheat noodles...carry a modest but non-negligible Cd burden | ”Generally low” for wheat Pb and “modest” for noodle Cd without ppb; the “non-negligible” qualifier in the second phrase is appropriate framing but “modest” is not anchored | Anchor: “Lead in wheat grain is typically below [X ppb] in commercial non-contaminated sources; noodle Cd reflects the grain Cd (typically [A–B ppb] in commercial semolina) and is relevant for high-frequency consumers” | ⚠ yes — both claims are load-bearing for the risk characterization of this commodity |
| oat.md | 151 | Pb (generally low) / iAs in oat is at trace levels, not a meaningful contributor to dietary iAs | ”Generally low” for oat Pb without ppb anchor; “at trace levels, not a meaningful contributor” makes a dietary significance claim without dose/frequency context | Anchor: “Pb in oat grain is typically below [X ppb] in commercial supply per corpus data (profile below); iAs in oat is typically below 5 ppb in aerobic-soil production, substantially lower than rice” — or state that no corpus data are available if the profile is pending | ⚠ yes — “not a meaningful contributor to dietary iAs” is a dietary significance statement that requires quantitative backing |
| oat-ring-cereal.md | 99 | Lead concentrations in oat grain are generally low / iAs is not a significant analyte of concern for oats under typical production conditions | ”Generally low” for Pb and “not a significant analyte of concern” for iAs, both without ppb anchor or dose context | Anchor: “Pb in oat grain is typically below [X ppb]; iAs in oats from aerobic soils is typically below 5 ppb (versus 50–200 ppb in paddy rice), making it a low-priority analyte in this matrix” | ⚠ yes — “not a significant analyte of concern” is a monitoring-priority statement read by QA practitioners |
| olive-oil.md | 127 | Refined olive oil in glass adds modest Ni from refining-clay contact | ”Modest” for Ni pickup from bleaching earth without a ppb value; mechanism is given | Rephrase: “Refined olive oil in glass may carry trace Ni from bleaching-earth contact; no specific corpus data quantify this increment, but it is expected to be below regulatory Ni limits for olive oil” — or anchor to any available survey data | no — Ni from refining clay is a documented pathway; “modest” without ppb is the trigger |
| onions.md | 155 | carry negligible metals because the extraction process is selective for volatile and lipophilic compounds | ”Negligible” for onion oil/oleoresin metals without ppb; mechanism is given but not data-anchored | Rephrase: “carry metals below detection limits, consistent with the expected lipophilic selectivity of the extraction process; no specific corpus data for onion oil metals are available” | no — mechanism is clear; “negligible” without data is the trigger |
| oranges.md | 155 | Orange extract and orange oil (cold-pressed or steam-distilled) carry negligible metals because the extraction process is selective for lipophilic volatile compounds | ”Negligible” for orange oil/extract metals without ppb; same pattern as onion oil | Rephrase: “Orange extract and orange oil carry metals below detection limits, consistent with lipophilic-selective extraction; no specific corpus data for orange oil metals are available” | no — same pattern as onion oil; mechanism is clear |
| oranges.md | 177 | Given the very low metal concentrations in orange flesh documented across monitoring programs | ”Very low” partially anchored by reference to monitoring programs cited earlier in the page; specific ppb not in this sentence | Replace with the actual profile values: “Given the orange Pb and Cd concentrations (profile below, all near or below detection limits in TDS data)“ | no — monitoring data are cited earlier; “very low” is the trigger |
| parsnips.md | 148 | lower-risk root vegetables (for example, swede or celeriac from well-managed soils) | ”Lower-risk” for swede/celeriac without ppb comparison to parsnips | Rephrase: “root vegetables with documented lower Cd uptake than parsnips (such as swede or celeriac from well-managed soils, per [specific source if available])“ | no — comparative claim for formulation substitution; anchoring would make the recommendation more actionable |
| pasta.md | 126 | provides a modest reduction in dietary Cd exposure from this food | ”Modest” for the pasta cooking-water Cd leaching effect; the paragraph notes estimates of 10–20 percent reduction from Italian pasta studies | Rephrase: “provides an estimated 10–20 percent reduction in dietary Cd exposure from pasta per Italian pasta studies (no corpus data specific to US-market boxed macaroni)“ | no — the percentage range IS given in the same paragraph; moving it inline anchors the “modest” claim |
| pasta.md | 146 | provides a modest additional reduction in dietary Cd exposure | ”Modest” for the drain-cooking Cd reduction; 10–20 percent estimate appears in the same paragraph but not this sentence | Anchor to the estimate: “provides an estimated 10–20 percent reduction in dietary Cd exposure (per Italian pasta leaching studies)“ | no — same pattern as pasta:126 |
| peach.md | 95 | interior flesh concentrations are very low / peach flesh cadmium concentrations are among the lowest observed in monitored food groups | ”Very low” for flesh Pb without ppb; “among the lowest observed in monitored food groups” for Cd without an absolute value | Anchor from the contamination profile: “interior flesh Pb concentrations are typically at or below detection (TDS data, profile below); peach flesh Cd is similarly at or near detection limits, consistent with the lowest-Cd tiers of the monitored food system” | no — the table carries the data; anchoring it here makes the claim self-contained |
| peanut-butter.md | 134 | Refined peanut oil...contains negligible heavy metal residues because metals partition to the aqueous and solid phases during refining | ”Negligible” for refined peanut oil metals without ppb; mechanism is given | Rephrase: “Refined peanut oil contains metals at or below detection limits, consistent with the partition of metal-bearing fractions to the aqueous and solid phases during refining” | no — mechanism is correct; “negligible” without ppb is the trigger |
P3 — em dashes & inline bold
| Page | Line | Offending text | Why | Suggested fix | Claim-adjacent? |
|---|---|---|---|---|---|
| milk-and-dairy.md | 148 | Dairy is generally a low-metal-load food commodity compared to muscle, organ, or grain commodities — the mammary gland excretes only a small fraction of the animal's dietary metal intake into milk | Em dash connecting the claim to its mechanism explanation in running prose | Replace with a semicolon or period: “Dairy is generally a low-metal-load food commodity compared to muscle, organ, or grain commodities; the mammary gland excretes only a small fraction of dietary metal intake into milk” | no |
| milk-and-dairy.md | 164–170 | **Concentration steps**..., **Fermentation**..., **Butter and butterfat extraction**..., **Packaging migration**... | Four bold paragraph-opening labels for processing steps in the “Processing effects” section; same pattern as the coffee/coconut processing-step labels flagged in batches 4–6. | Convert to a bullet list with bold labels, or remove bold and let the “Processing effects” heading carry the section structure. | no |
| milk-and-dairy.md | 182–192 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers in the Mitigation section; systemic pattern. | Same fix as the general lever-label pattern. | no |
| mixed-meals.md | 157–167 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| molluscs.md | 114 | bivalves filter large volumes of water — an oyster filters 5-50 liters of water per day — and concentrate dissolved and particle-bound metals | Em-dash pair around an illustrative parenthetical clause in running prose | Replace with parentheses: “bivalves filter large volumes of water (an oyster filters 5–50 liters per day) and concentrate dissolved and particle-bound metals” | no |
| molluscs.md | 114–118 | **Bivalve molluscs**..., **Cephalopods**..., **Gastropods**... | Three bold paragraph-opening labels for mollusc subgroups in the “Ranges by source, region, and variety” section; each names a discrete taxonomic entity with quantitative data embedded. Prose paragraphs, not a bullet list. | Convert to bullet list with bold definition labels (legitimate catalog) or remove bold. | no |
| molluscs.md | 142–148 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Four bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| non-rice-grains.md | 173–183 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| non-root-vegetable-purees.md | 114 | differentiates non-root vegetable purees (10 ppb Pb) from root vegetable purees (20 ppb Pb) — the tighter limit on non-root reflects the lower baseline plus consumer expectation of lower contamination | Em dash connecting a regulatory distinction to its rationale in running prose | Replace with a semicolon: “differentiates non-root vegetable purees (10 ppb Pb) from root vegetable purees (20 ppb Pb); the tighter limit on non-root reflects the lower baseline plus consumer expectation of lower contamination” | no |
| non-root-vegetable-purees.md | 130–140 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| non-soy-infant-formula.md | 154–164 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| non-soy-protein-source.md | 157–167 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| oat.md | 175–185 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |
| olive-oil.md | 98 | found significant differences across six packaging types — glass, PET, tin, porcelain, aluminium, cardboard — with the porcelain and aluminium packages contributing more measurable Pb, Cd, Ni, and Cr | Em-dash pair around an enumerated list of packaging types in running prose | Replace with parentheses: “found significant differences across six packaging types (glass, PET, tin, porcelain, aluminium, cardboard), with the porcelain and aluminium packages contributing more measurable Pb, Cd, Ni, and Cr” | no |
| olive-oil.md | 123 | olive pomace oil — solvent-extracted from the spent pomace — recovers oil from the fraction | Em-dash pair around an appositive describing the production method | Replace with parentheses: “olive pomace oil (solvent-extracted from the spent pomace) recovers oil from the fraction” | no |
| olives.md | 119 | Malta (4 cultivars — Carolea, Cipressina, Leccino, Bidni — from a single cultivator) | Em-dash pair around a list of cultivar names inside a parenthetical expression | Replace with a colon or comma: “Malta (4 cultivars: Carolea, Cipressina, Leccino, Bidni, from a single cultivator)“ | no |
| parsley.md | 76 | Hamburg parsley — Hamburg parsley grown for the root accumulates root-vegetable Pb in the root and leafy-vegetable Pb in the leaves | Em dash inside a parenthetical list item, connecting the cultivar name to an elaborating clause | Replace with a comma: “Hamburg parsley (grown for the root, which accumulates root-vegetable Pb levels while the leaves carry the leafy-vegetable Pb profile)“ | no |
| parsley.md | 88–98 | **Sourcing levers** (...)... through **Packaging and storage levers** (...)... | Six bold lever-label paragraph openers; systemic pattern. | Same fix as above. | no |