Prose-style & consumer-language sweep — 2026-06-03 (batch 3, ingredients)
Counts: P1 0 · P2 24 · P3 4. Report-only; no pages were edited.
Pages scanned (24): butternut-squash.md, cabbage.md, camellia-sinensis.md, canned-corn.md, canned-fruit-cocktail.md, canned-green-beans.md, canned-mushrooms.md, canned-spaghetti.md, canned-tomatoes.md, canned-tuna.md, cantaloupe.md, cashews.md, cauliflower.md, celery.md, cereals.md, chapattis.md, cheddar-cheese.md, chicken-noodle-soup.md, chicken-potpie.md, chicken.md, chili-con-carne-with-beans.md, chocolate-cake.md, chocolate-chip-cookies.md, chocolate-reduced-fat-milk.md
Stubs skipped (5): casein.md (provisional_scaffold + literature_scope: thin), cereal-bars.md (provisional_scaffold + literature_scope: thin), cereal-based-infant-formula.md (provisional_scaffold + literature_scope: thin), chitosan.md (provisional_scaffold + literature_scope: thin), cereal.md (autonomy-created stub, body is stub notice + all-dash contamination table).
Next run resumes at wiki/ingredients/cinnamon.md (skipping chocolate.md and cocoa.md — already covered in batch 1).
P1 — consumer safe/dangerous without anchor
| Page | Line | Offending text | Why | Suggested fix | Claim-adjacent? |
|---|---|---|---|---|---|
| (none found) | — | — | — | — | — |
P2 — bullets-in-prose & other qualifiers
The dominant pattern this batch: ingredients with genuinely low metal burdens (dairy, white-muscle chicken, most fruits, canned goods with clean intrinsic fractions) are described with unanchored magnitude qualifiers — “low”, “modest”, “trace”, “negligible”, “minimal” — in introductory and derivative-risk paragraphs that precede or are separate from the data tables that would anchor them. Editors applying fixes should verify the claim is supported before choosing an anchor number.
| Page | Line | Offending text | Why | Suggested fix | Claim-adjacent? |
|---|---|---|---|---|---|
| cabbage.md | 123 | carry the parent cabbage's metal load with negligible processing-driven change | ”negligible” characterizes magnitude of fermentation’s effect on metals without a quantitative anchor; mechanism is given (“fermentation does not concentrate or dilute metals”) but no ppb comparison | Rephrase to state the mechanism only: “carry the parent cabbage’s metal load unchanged; fermentation does not concentrate or dilute metals on a per-mass basis” | no — mechanical statement, not a risk claim |
| canned-corn.md | 99 | maize is classified as a low Cd accumulator at typical agricultural soil concentrations | ”low Cd accumulator” is a magnitude claim without a ppb anchor in this sentence; TDS data (Cd median 1.1 ppb) appears later in the page | Anchor to TDS range in same sentence, or add parenthetical: “a low Cd accumulator (TDS: Cd median 1.1 ppb, range 0–2.2 ppb)” | ⚠ yes — “low” implies a contamination level the editor should verify against the corpus data |
| canned-corn.md | 151 | lacquered cans maintain much lower levels | ”much lower” characterizes Sn in lacquered cans relative to unlacquered (50–200 mg/kg given), but no number is given for the lacquered-can Sn level | Add a typical lacquered-can Sn figure, e.g., “lacquered cans typically maintain Sn below 10 mg/kg in the same storage scenarios” — or cite a source for the specific reduction | ⚠ yes — “much lower” implies a specific reduction magnitude |
| canned-fruit-cocktail.md | 99 | Lead, cadmium, and arsenic in the fruit fractions themselves contribute modestly | ”modestly” characterizes Pb/Cd/As magnitude from fruit fractions; TDS values (Pb median 5 ppb, Cd ≤ LOD) appear later in the page | Rephrase to reference TDS values directly: “Pb in the fruit fractions sits at ~5 ppb (TDS median) and Cd below detection, with tAs also at or below detection” | ⚠ yes — “modestly” implies a low contribution that should be anchored to the corpus data |
| canned-green-beans.md | 99 | a modest intrinsic pathway from the bean pod and seed tissue | ”modest” characterizes intrinsic Cd/Pb contamination level for canned green beans without a ppb anchor in this sentence; the sentence after gives “single-digit to low tens of ppb” for Cd | Merge the anchor forward: “a modest intrinsic Cd and Pb pathway from the bean tissue (typically single-digit to low tens of ppb for Cd in non-contaminated soils)” | ⚠ yes — “modest” implies a specific magnitude range |
| canned-mushrooms.md | 99 | cultivated mushrooms...typically show lower but still detectable heavy metal levels | ”lower but still detectable” is a relative magnitude comparison to wild mushrooms without a quantitative anchor for cultivated Cd/Hg in this sentence; literature ranges (50–500 ppb Cd for Agaricus bisporus) appear later in the page | Add the literature range: “cultivated Agaricus bisporus typically shows Cd of 50–500 ppb fresh weight in European surveys, substantially lower than wild species but well above most vegetable baselines” | ⚠ yes — the “lower” claim needs anchoring; the TDS single composite (1.8 ppb) conflicts with literature values, which is acknowledged later |
| canned-spaghetti.md | 95 | Tomato is a relatively low metal accumulator as a fruit-vegetable | ”relatively low metal accumulator” characterizes tomato’s contamination magnitude for the reader without a ppb anchor; the mechanism is explained but no number is given in this sentence | Add typical Pb/Cd range for fresh tomato, e.g., “Tomato, carrying Pb typically below 10 ppb and Cd below 20 ppb fresh weight, is a relatively low metal accumulator among vegetables” — see also canned-tomatoes.md | ⚠ yes — the “low” claim should match what corpus sources support for tomato fruit |
| canned-tomatoes.md | 95 | Fresh tomatoes are among the lower-accumulating common vegetables for Pb and Cd / Cd accumulation in tomato fruit is detectable but generally low relative to leafy vegetables or root crops | Two unanchored relative magnitude claims in the same paragraph; the numbers (Pb < 10 ppb, Cd < 20 ppb) appear in line 124 | Move the numeric anchor to line 95 or cross-reference: “Pb typically below 10 ppb and Cd typically below 20 ppb fresh weight (per general literature), placing fresh tomatoes among the lower-accumulating common vegetables” | ⚠ yes — “lower-accumulating” and “generally low” are the claim; editor should confirm the corpus supports these specific thresholds |
| canned-tomatoes.md | 124 | iAs, tHg, and Al also at low levels / Lacquered cans reduce Sn to trace levels | Two qualifiers in the same sentence: “low levels” for iAs/tHg/Al lacks ppb anchors (only Pb and Cd are anchored in this sentence); “trace levels” for lacquered-can Sn is similarly unanchored — unlacquered Sn is given as tens to hundreds of mg/kg, but the lacquered-can number is not | For iAs/tHg/Al: cite TDS values or note “at or below reporting limits in survey data.” For Sn: give a typical lacquered-can Sn figure or rephrase to “substantially reduced, typically below the EU 200 mg/kg limit even at extended storage” | ⚠ yes — “trace” for lacquered Sn implies a specific level the editor should verify |
| canned-tuna.md | 94 | mostly as low-toxicity organoarsenic forms | ”low-toxicity” characterizes the risk to the reader from this arsenic species form without a dose or population anchor; while technically accurate for organoarsenic vs iAs, it implies a safety conclusion without quantitative grounding | Replace with the mechanism: “mostly as organoarsenic forms (arsenobetaine, arsenocholine) that are substantially less bioavailable than inorganic arsenic” — and note the iAs fraction separately per CLAUDE.md Part 14 | ⚠ yes — “low-toxicity” makes a risk claim that depends on species fractionation; the iAs fraction should be quantified rather than minimized by this qualifier |
| canned-tuna.md | 127 | carry the same fish-meat mercury load with negligible packaging-source metal contribution | ”negligible” characterizes Pb/Sn/Al from tuna pouches without a number; modern fully-welded steel cans are the reference point, but pouch-specific metal migration data are not cited | Replace with “with lower packaging-source metal contribution than tinplate cans, given the absence of a tin interior” or cite a pouch-specific migration figure if available in the corpus | no — the mechanical claim (“pouches have no tin interior”) is defensible; “negligible” is a conclusion not a claim about food safety per se |
| cantaloupe.md | 99 | Nickel and total arsenic are detectable but at modest concentrations | ”modest” characterizes Ni and tAs magnitude without a ppb anchor in this sentence; TDS data (Ni up to 270 ppb, tAs up to 26 ppb) are in the table section below | Move the numbers forward: “Nickel (TDS median 75 ppb, max 270 ppb) and total arsenic (TDS median 7 ppb) are detectable” — removing “modest” since 270 ppb Ni is not obviously modest | ⚠ yes — the TDS Ni data (up to 270 ppb) arguably contradicts “modest”; editor should reconcile |
| cantaloupe.md | 159 | Given its very low intrinsic metal load for most analytes, it contributes minimally / from a very low baseline this remains well within regulatory limits for most analytes / low-risk profile | Three unanchored magnitude qualifiers in the Ingredient-derivative risk section; TDS data (Pb=0, tHg=0, Cd median 7.1 ppb, but Ni up to 270 ppb) are in an earlier section | Anchor each claim to the TDS distribution. Note the Ni exception explicitly in the first sentence rather than at the end: “For all analytes except Ni, the TDS data (Pb=0, Cd median 7.1 ppb) support a low derivative contribution at typical inclusion rates; Ni (up to 270 ppb) is the exception and drives the product-level estimate” | ⚠ yes — “low-risk profile” and “very low baseline” are claim-adjacent; Ni data at up to 270 ppb weakens both characterizations |
| cashews.md | 69 | trace Pb from production and post-harvest handling | ”trace Pb” is an unanchored magnitude qualifier for this contamination pathway; no ppb figure is given for production/post-harvest Pb in cashews | Add typical Pb range for cashews or replace with “low but variable Pb depending on growing-region soil and handling practices” — omitting “trace” until corpus data supports it | ⚠ yes — “trace” implies a specific low magnitude; if post-harvest Pb is not negligible in some origins, this characterization misleads |
| cheddar-cheese.md | 95 | accumulates heavy metals at very low concentrations relative to plant foods / a small fraction partitions into milk / the overall metal burden in aged cheddar remains extremely low relative to plant-derived ingredients | Three unanchored magnitude qualifiers in the opening paragraph; the TDS data (all metals at or below reporting limits except U at max 1.3 ppb) appears later and would anchor these claims | Anchor “very low” and “extremely low” to TDS data inline: “accumulates heavy metals at concentrations at or below detection limits in FDA TDS surveys (27 composites; Cd, Pb, tAs, tHg, Ni, Cr all ≤ LOD).” For “small fraction,” cite a dairy-partition factor or remove | ⚠ yes — though TDS data supports “very low” for cheddar, citing the specific data is more defensible |
| chicken.md | 95 | generally a low accumulator of heavy metals / retaining only a small fraction / Cadmium and lead in commercial poultry muscle are consistently low in global surveillance data / residual total arsenic and inorganic arsenic remain detectable at low levels | Four unanchored qualifiers in the opening “Why” paragraph; the contamination profile table (Pb typical 0 ppb, Cd typical 0–3 ppb) follows immediately after | For Pb and Cd, anchor to the contamination profile values. For “at low levels” for post-withdrawal arsenic, add a note: “residual tAs and iAs remain detectable at concentrations up to ~20 ppb tAs in post-2015 surveys (lasky2013-arsenic-chicken-correspondence)” — or replace “low levels” with “at concentrations above pre-withdrawal baselines but below FDA action levels” | ⚠ yes — all four qualifiers are claim-adjacent; the arsenic “low levels” characterization is particularly sensitive given the US broiler withdrawal history |
| chicken.md | 132 | this pathway is expected to be low but is not quantified | ”expected to be low” for chicken-fat metal pathway is an unanchored prediction; the page itself acknowledges it is unquantified | Rephrase to remove the magnitude prediction: “this pathway has not been quantified in the current corpus; it is left as a data gap” — reserving magnitude characterization until corpus data supports it | no — the “not quantified” hedge is explicit; this is more of a literature gap than a consumer risk claim |
| chicken-noodle-soup.md | 99 | may contribute trace lead through leaching from the skeletal calcium matrix | ”trace lead” from bone broth leaching is unanchored; no ppb range given for this pathway | Add a parenthetical if corpus data supports it, or rephrase to “may contribute lead through leaching from bone surfaces (the lead bone-mobilization pathway is documented in the literature; concentrations in broth are not quantified in the current corpus)” | ⚠ yes — the magnitude of bone-lead leaching into broth is a meaningful claim for products marketed to children |
| chicken-potpie.md | 99 | The chicken muscle fraction contributes very little to any of the ten profiled metals / neither of which is a documented source of significant metal migration | Two unanchored qualifiers: “very little” for chicken’s metal contribution; “significant metal migration” for packaging (metal pie tin, cardboard sleeve). TDS data (Cd median 6.3 ppb, Ni median ~47 ppb) is given in the same paragraph for the whole product | Anchor “very little” to the per-ingredient estimates if available, or soften: “The chicken muscle fraction, with typical Pb 0 ppb and Cd 0–3 ppb (per chicken), contributes less than the vegetable and pastry fractions.” For packaging, replace “significant” with “documented” | no — the packaging claim (“neither is a documented source”) is a literature absence claim; the chicken claim is relative to the vegetable fraction |
| chili-con-carne-with-beans.md | 155 | tomato is generally a low-metal ingredient and this effect is small | ”low-metal ingredient” and “small effect” characterize tomato’s metal contribution and the concentration effect of tomato processing without quantitative anchors | Anchor to tomato Pb/Cd data (Pb typically below 10 ppb, Cd typically below 20 ppb fresh weight per canned-tomatoes page) or remove the characterization | ⚠ yes — “low-metal” for tomato as an ingredient in chili is claim-adjacent with respect to the product-level risk |
| chili-con-carne-with-beans.md | 159 | leaving a product with a much lower metal burden per serving | ”much lower metal burden per serving” for bean-free chili is an unanchored comparative claim; the bean-driven Ni (median 330 ppb) for chili-with-beans is given, but no estimate is given for the hypothetical bean-free variant | Anchor: “eliminating the bean fraction would reduce Ni by approximately the bean-fraction proportion (roughly 60–70% of the product by inclusion), based on the TDS Ni median of 330 ppb” — or remove the comparative claim | ⚠ yes — “much lower” is a magnitude claim that should reflect the actual ingredient ratios |
| chocolate-cake.md | 99 | Eggs, butter, sugar, and milk contribute negligible metals | ”negligible” is an unanchored magnitude qualifier for these ingredient streams; no ppb values are given for their contributions in this context | Rephrase: “Eggs, butter, sugar, and milk contribute metal loads well below detection in TDS surveys (see eggs, butter)” — or remove the comparative claim and simply note the cocoa fraction is dominant | no — the relative claim (other ingredients contribute less than cocoa) is defensible; “negligible” is the only trigger word |
| chocolate-chip-cookies.md | 99 | Butter, sugar, eggs, and leavening agents contribute negligible metals | Same pattern as chocolate-cake.md:99: “negligible” unanchored for the non-cocoa, non-wheat ingredient streams | Rephrase: “Butter, sugar, eggs, and leavening agents contribute metal loads below detection in TDS surveys” | no |
| chocolate-reduced-fat-milk.md | 155 | the practical effect on metal concentrations of using reduced-fat versus whole milk is minimal | ”minimal” characterizes the fat-reduction effect on metal concentrations without a quantitative comparison; no ppb difference between whole-milk and reduced-fat versions is cited | Rephrase: “the fat-fraction metal contribution is small relative to the cocoa-solids fraction; no quantitative comparison between whole-milk and reduced-fat versions is available in the current corpus” | no |
P3 — em dashes & inline bold
| Page | Line | Offending text | Why | Suggested fix | Claim-adjacent? |
|---|---|---|---|---|---|
| cabbage.md | 157 | and Cd 0.050 mg/kg for the same category — substantially tighter than Codex | Em dash connecting the EU limit statement to its interpretive gloss in running prose | Replace with a comma or semicolon: “and Cd 0.050 mg/kg for the same category, substantially tighter than Codex” | no |
| cashews.md | 73 | (Côte d'Ivoire, Ghana, Nigeria — the largest production region globally) | Em dash inside a parenthetical geographic list; used to attach a relative-clause gloss to one list item | Replace with a nested parenthesis or comma: “(Côte d’Ivoire, Ghana, Nigeria (the largest production region globally))” or drop the gloss | no |
| canned-tuna.md | 94 | The defining heavy-metals concern is mercury — specifically methylmercury (MeHg) | Em dash introducing a parenthetical elaboration in the opening paragraph of running prose | Replace with a comma or colon: “The defining heavy-metals concern is mercury, specifically methylmercury (MeHg)“ | no |
| camellia-sinensis.md | 160, 162, 164, 166 | **Brew temperature and time** determine Al..., **Water source** matters..., **Tea bag vs loose leaf** is..., **Lemon juice or other acid** added... | Bold used at the start of four consecutive paragraphs to highlight the brewing-factor subject; the paragraphs are not in a bullet list. Per check #2, definition labels opening a glossary entry are not violations; these are closer to bold topic-sentence labels. Borderline: they name discrete entities but the bold serves as a structural cue, not a definition label. | If the section is meant as a catalog, convert to a bullet list with explicit bold definition labels (as in the Mitigation section). If it is prose, remove the bold and let the sentence structure carry the topic. | no |