The recordsAll teaching accords

Station 8 of 13 · Accord · the record

A muguet (lily of the valley), and what regulation took from it

A muguet composed from synthetic materials, because lily of the valley gives no oil, and a lesson in what regulation removed.

Teaching accord A teaching accord for learning: not a Radler formula, not a finished perfume, not safety-assessed; do not make it for skin.

Two of its classic materials are now prohibited in EU cosmetics. Remove them and see what the heart loses; the substitutions show the usual response.

Read the essay

The record

Each field carries its evidence: open a chip for the source, the quoted words and the confidence. A field with nothing behind it says so and stays empty.

Name
A muguet (lily of the valley), and what regulation took from it
house interpretation

the house’s own reading, labelled as such

the house’s name for this teaching accord

confidence C

Purpose
teaching
house rule

the house’s own reading, labelled as such

the site’s rule: every accord on record is a teaching accord; the house’s own formulas never enter the repository (brief §12.7)

confidence B

Structure
A muguet built from aldehydes and alcohols: the flower gives no oil, so the note is composed from synthetic materials.
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s convention; the Osmothèque classes Diorissimo (1956) as a muguet soliflore ("Solinote Muguet"). Two of the classic materials are now prohibited in EU cosmetics (p-BMHCA, HICC): the accord keeps them to teach the change

confidence B

Family
floral (the Osmothèque’s "Solinote Muguet")
perfumery convention

Diorissimo DIOR 1956 EDT Floral Solinote Muguet Edmond Roudnitska

OSMO-1 · Osmothèque, Conservatoire international des parfums · read 2026-09-30

the family as the Osmothèque classes a fragrance of this structure

confidence B

Components
Under each component, the evidence for its place in the accord. Teaching parts are a house interpretation for learning: modeled, not measured, not a formula.
ComponentRoleTeaching partsVolatility
Hydroxycitronellal
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s convention; no odour description is filed for it

confidence B

heart
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s top / heart / base convention for this material

confidence B

40 parts
modeled

modeled

teaching proportions, a house interpretation for learning: modeled, not measured, not a formula, not for making

confidence C

not on record
why

no volatility class on record for hydroxycitronellal

Butylphenyl methylpropional
documented

3-(4-(tert-Butyl)phenyl)-2-methylpropanal, trade name Lilial or Lysmeral, is an aroma chemical with a sweet lily like odor.

ENE-1 · Fan XY, et al · read 2026-09-30

prohibited in EU cosmetic products since 1 March 2022 (eu-2021-1902)

confidence A

heart
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s top / heart / base convention for this material

confidence B

20 parts
modeled

modeled

teaching proportions, a house interpretation for learning: modeled, not measured, not a formula, not for making

confidence C

not on record
why

no volatility class on record for p-bmhca

Hydroxyisohexyl 3-cyclohexene carboxaldehyde
perfumery convention

the trade’s shared vocabulary; no single source

prohibited in EU cosmetic products (eu-2017-1410); kept here to teach what was lost

confidence C

heart
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s top / heart / base convention for this material

confidence B

5 parts
modeled

modeled

teaching proportions, a house interpretation for learning: modeled, not measured, not a formula, not for making

confidence C

not on record
why

no volatility class on record for hicc

2-phenylethyl alcohol
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s convention

confidence B

heart
documented

Floral notes, such as rose, jasmine, orange blossom, or lavender, are typically middle notes defining the “heart” of many perfumes

KLE-1 · Kleoff M, Kiler P, Heretsch P · read 2026-09-30

the review names rose among the floral heart notes

confidence A

20 parts
modeled

modeled

teaching proportions, a house interpretation for learning: modeled, not measured, not a formula, not for making

confidence C

not on record
why

no volatility class on record for phenylethyl-alcohol

Linalool
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s convention

confidence B

top
perfumery convention

the trade’s shared vocabulary; no single source

the trade’s top / heart / base convention for this material

confidence B

15 parts
modeled

modeled

teaching proportions, a house interpretation for learning: modeled, not measured, not a formula, not for making

confidence C

top
modeled

modeled

the molecule record’s volatility class (modeled: the simulator’s data is a set of site constants, not measurements)

confidence C

Substitutions
Regulations that apply to a component
Odour descriptors
In history
Fragrances of this family

Causal Debugger why did this change matter?

Teaching accord · modeled A teaching accord for learning: not a Radler formula, not a finished perfume, not safety-assessed; do not make it for skin. Everything below is computed by a teaching model from the records: modeled, not measured, and not a prediction of how any real perfume smells. Regulation lines are flags, not compliance advice.

Change one thing and follow it down the chain: the proportions, the molecules, the words on record, a modeled reading, the dry-down, odour activity where a threshold is measured, the regulations, the provenance. Then: why it mattered.

One change

Without JavaScript these controls stay still: every single change is worked out below.

Modeled · before → after

Choose one component and one change: each layer of the chain appears here, labelled modeled, each input with its evidence.

What happens if… every single change, worked out

The same model, run when this page was built. Figures are modeled: parts, and shares per 100 parts; odour space is the share of the accord whose components carry each word; the dry-down uses the half-lives top 1.5 h, heart 6 h, base 12 h (the simulator’s site constants, not measurements); odour activity is share ÷ measured threshold, compared within one medium and unit, never across (a teaching ratio, not a measurement of this accord). This teaching accord is not a finished product and is not assessed. Nothing is for sale.

Double hydroxycitronellal (×2)
1 · Composition
hydroxycitronellal 40 → 80 parts (40.0 → 57.1 per 100); the others ×0.71.
2 · Molecules
hydroxycitronellal itself rises.
3 · Odour space
“floral” 20.0 → 14.3, “honey” 20.0 → 14.3, “lily” 20.0 → 14.3, “sweet” 20.0 → 14.3; no descriptor on record: 45.0 → 60.7.
4 · Perceived character
A modeled reading, not a prediction: less “floral” and “honey”, with less of it in the opening.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 10.7 / 89.3 / 0.0; half gone at 4.9 h → 5.2 h.
6 · Odour activity
hydroxycitronellal: no threshold on record.
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. The modeled dry-down moves: heart-class components go from 85.0 to 89.3 per 100 parts, and half the accord is gone at 5.2 h instead of 4.9 h (hydroxycitronellal is heart, from its role in the accord).
  2. Components carrying “floral” go from 20.0 to 14.3 per 100 parts, because hydroxycitronellal, which does not carry it, now takes more of the accord.
Remove hydroxycitronellal
1 · Composition
hydroxycitronellal 40 → 0 parts (40.0 → 0.0 per 100); the others ×1.67.
2 · Molecules
hydroxycitronellal itself leaves.
3 · Odour space
“floral” 20.0 → 33.3, “honey” 20.0 → 33.3, “lily” 20.0 → 33.3, “sweet” 20.0 → 33.3; no descriptor on record: 45.0 → 8.3.
4 · Perceived character
A modeled reading, not a prediction: more “floral” and “honey”, with more of it in the opening.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 25.0 / 75.0 / 0.0; half gone at 4.9 h → 4.2 h.
6 · Odour activity
hydroxycitronellal: no threshold on record.
7 · Regulation flags
Flags leaving the accord: IFRA Standard: Hydroxycitronellal.
Why did this change matter?
  1. It takes the “IFRA Standard: Hydroxycitronellal” flag (restriction) off the accord: it names hydroxycitronellal and no other component.
  2. The modeled dry-down moves: heart-class components go from 85.0 to 75.0 per 100 parts, and half the accord is gone at 4.2 h instead of 4.9 h (hydroxycitronellal is heart, from its role in the accord).
  3. Components carrying “floral” go from 20.0 to 33.3 per 100 parts, because hydroxycitronellal, which does not carry it, now takes less of the accord.
Double butylphenyl methylpropional (×2)
1 · Composition
butylphenyl methylpropional 20 → 40 parts (20.0 → 33.3 per 100); the others ×0.83.
2 · Molecules
butylphenyl methylpropional itself rises.
3 · Odour space
“lily” 20.0 → 33.3, “sweet” 20.0 → 33.3, “floral” 20.0 → 16.7, “honey” 20.0 → 16.7; no descriptor on record: 45.0 → 37.5.
4 · Perceived character
A modeled reading, not a prediction: more “lily” and “sweet” and less “floral” and “honey”.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 12.5 / 87.5 / 0.0; half gone at 4.9 h → 5.1 h.
6 · Odour activity
butylphenyl methylpropional: no threshold on record.
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. Components carrying “lily” go from 20.0 to 33.3 per 100 parts, because butylphenyl methylpropional carries it (its descriptors field).
  2. The modeled dry-down moves: heart-class components go from 85.0 to 87.5 per 100 parts, and half the accord is gone at 5.1 h instead of 4.9 h (butylphenyl methylpropional is heart, from its role in the accord).
Remove butylphenyl methylpropional
1 · Composition
butylphenyl methylpropional 20 → 0 parts (20.0 → 0.0 per 100); the others ×1.25.
2 · Molecules
butylphenyl methylpropional itself leaves.
3 · Odour space
“lily” 20.0 → 0.0, “sweet” 20.0 → 0.0, “floral” 20.0 → 25.0, “honey” 20.0 → 25.0; no descriptor on record: 45.0 → 56.3.
4 · Perceived character
A modeled reading, not a prediction: more “floral” and “honey” and no “lily” and “sweet” left, with more of it in the opening.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 18.8 / 81.3 / 0.0; half gone at 4.9 h → 4.6 h.
6 · Odour activity
butylphenyl methylpropional: no threshold on record.
7 · Regulation flags
Flags leaving the accord: Regulation (EU) 2021/1902 (substances classified CMR, butylphenyl methylpropional among them); IFRA Standard: p-tert-Butyl-α-methylhydrocinnamic aldehyde (p-BMHCA).
Why did this change matter?
  1. It takes the “Regulation (EU) 2021/1902 (substances classified CMR, butylphenyl methylpropional among them)” flag (a prohibition (Annex II)) off the accord: it names butylphenyl methylpropional and no other component.
  2. It takes the “IFRA Standard: p-tert-Butyl-α-methylhydrocinnamic aldehyde (p-BMHCA)” flag (restriction / prohibition) off the accord: it names butylphenyl methylpropional and no other component.
  3. butylphenyl methylpropional carried “lily” and no other component does on record, so the word leaves the modeled profile.
Double HICC (×2)
1 · Composition
HICC 5 → 10 parts (5.0 → 9.5 per 100); the others ×0.95.
2 · Molecules
HICC itself rises.
3 · Odour space
“floral” 20.0 → 19.0, “honey” 20.0 → 19.0, “lily” 20.0 → 19.0, “sweet” 20.0 → 19.0; no descriptor on record: 45.0 → 47.6.
4 · Perceived character
A modeled reading, not a prediction: less “floral” and “honey”.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 14.3 / 85.7 / 0.0; half gone at 4.9 h → 4.9 h.
6 · Odour activity
HICC: no threshold on record.
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. Little moves: at 5 parts, HICC’s share goes from 5.0 to 9.5 per 100, so in this share-weighted model the change is nearly invisible, and no measured threshold is on record to show its strength.
Remove HICC
1 · Composition
HICC 5 → 0 parts (5.0 → 0.0 per 100); the others ×1.05.
2 · Molecules
HICC itself leaves.
3 · Odour space
“floral” 20.0 → 21.1, “honey” 20.0 → 21.1, “lily” 20.0 → 21.1, “sweet” 20.0 → 21.1; no descriptor on record: 45.0 → 42.1.
4 · Perceived character
A modeled reading, not a prediction: more “floral” and “honey”.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 15.8 / 84.2 / 0.0; half gone at 4.9 h → 4.8 h.
6 · Odour activity
HICC: no threshold on record.
7 · Regulation flags
Flags leaving the accord: Regulation (EU) 2017/1410 (HICC, atranol and chloroatranol prohibited); IFRA Standard: 3 and 4-(4-Hydroxy-4-methylpentyl)-3-cyclohexene-1-carboxaldehyde (HMPCC).
Why did this change matter?
  1. It takes the “Regulation (EU) 2017/1410 (HICC, atranol and chloroatranol prohibited)” flag (a prohibition (Annex II)) off the accord: it names HICC and no other component.
  2. It takes the “IFRA Standard: 3 and 4-(4-Hydroxy-4-methylpentyl)-3-cyclohexene-1-carboxaldehyde (HMPCC)” flag (restriction) off the accord: it names HICC and no other component.
  3. Components carrying “floral” go from 20.0 to 21.1 per 100 parts, because HICC, which does not carry it, now takes less of the accord.
Double 2-phenylethyl alcohol (×2)
1 · Composition
2-phenylethyl alcohol 20 → 40 parts (20.0 → 33.3 per 100); the others ×0.83.
2 · Molecules
2-phenylethyl alcohol itself rises.
3 · Odour space
“floral” 20.0 → 33.3, “honey” 20.0 → 33.3, “lily” 20.0 → 16.7, “sweet” 20.0 → 16.7; no descriptor on record: 45.0 → 37.5.
4 · Perceived character
A modeled reading, not a prediction: more “floral” and “honey” and less “lily” and “sweet”.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 12.5 / 87.5 / 0.0; half gone at 4.9 h → 5.1 h.
6 · Odour activity
2-phenylethyl alcohol (140 µg/kg in water, measured): ×1.67; no ratio with linalool (air, ng/L).
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. Components carrying “floral” go from 20.0 to 33.3 per 100 parts, because 2-phenylethyl alcohol carries it (its descriptors field).
  2. The modeled dry-down moves: heart-class components go from 85.0 to 87.5 per 100 parts, and half the accord is gone at 5.1 h instead of 4.9 h (2-phenylethyl alcohol is heart, from its role in the accord).
Remove 2-phenylethyl alcohol
1 · Composition
2-phenylethyl alcohol 20 → 0 parts (20.0 → 0.0 per 100); the others ×1.25.
2 · Molecules
2-phenylethyl alcohol itself leaves.
3 · Odour space
“floral” 20.0 → 0.0, “honey” 20.0 → 0.0, “lily” 20.0 → 25.0, “sweet” 20.0 → 25.0; no descriptor on record: 45.0 → 56.3.
4 · Perceived character
A modeled reading, not a prediction: more “lily” and “sweet” and no “floral” and “honey” left, with more of it in the opening.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 18.8 / 81.3 / 0.0; half gone at 4.9 h → 4.6 h.
6 · Odour activity
2-phenylethyl alcohol (140 µg/kg in water, measured): ×0; no ratio with linalool (air, ng/L).
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. 2-phenylethyl alcohol carried “floral” and no other component does on record, so the word leaves the modeled profile.
  2. 2-phenylethyl alcohol carried “honey” and no other component does on record, so the word leaves the modeled profile.
  3. The modeled dry-down moves: heart-class components go from 85.0 to 81.3 per 100 parts, and half the accord is gone at 4.6 h instead of 4.9 h (2-phenylethyl alcohol is heart, from its role in the accord).
Double linalool (×2)
1 · Composition
linalool 15 → 30 parts (15.0 → 26.1 per 100); the others ×0.87.
2 · Molecules
linalool itself rises.
3 · Odour space
“medicinal” 15.0 → 26.1, “sharp” 15.0 → 26.1, “floral” 20.0 → 17.4, “honey” 20.0 → 17.4; no descriptor on record: 45.0 → 39.1.
4 · Perceived character
A modeled reading, not a prediction: more “medicinal” and “sharp” and less “floral” and “honey”, with more of it in the opening.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 26.1 / 73.9 / 0.0; half gone at 4.9 h → 4.1 h.
6 · Odour activity
linalool (3.2 ng/L in air, measured): ×1.74; no ratio with 2-phenylethyl alcohol (water, µg/kg).
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. The modeled dry-down moves: top-class components go from 15.0 to 26.1 per 100 parts, and half the accord is gone at 4.1 h instead of 4.9 h (linalool is top, from its volatility class, modeled).
  2. Components carrying “medicinal” go from 15.0 to 26.1 per 100 parts, because linalool carries it (its descriptors field).
Remove linalool
1 · Composition
linalool 15 → 0 parts (15.0 → 0.0 per 100); the others ×1.18.
2 · Molecules
linalool itself leaves.
3 · Odour space
“medicinal” 15.0 → 0.0, “sharp” 15.0 → 0.0, “floral” 20.0 → 23.5, “honey” 20.0 → 23.5; no descriptor on record: 45.0 → 52.9.
4 · Perceived character
A modeled reading, not a prediction: more “floral” and “honey” and no “medicinal” and “sharp” left, with less of it in the opening.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 0.0 / 100.0 / 0.0; half gone at 4.9 h → 6.0 h.
6 · Odour activity
linalool (3.2 ng/L in air, measured): ×0; no ratio with 2-phenylethyl alcohol (water, µg/kg).
7 · Regulation flags
The accord’s flags do not change.
Why did this change matter?
  1. linalool carried “medicinal” and no other component does on record, so the word leaves the modeled profile.
  2. linalool carried “sharp” and no other component does on record, so the word leaves the modeled profile.
  3. The modeled dry-down moves: top-class components go from 15.0 to 0.0 per 100 parts, and half the accord is gone at 6.0 h instead of 4.9 h (linalool is top, from its volatility class, modeled).
Replace butylphenyl methylpropional with hydroxycitronellal
1 · Composition
butylphenyl methylpropional 20 → 0 parts (20.0 → 0.0 per 100); hydroxycitronellal 40 → 60 parts (40.0 → 60.0 per 100); the others ×1.00.
2 · Molecules
butylphenyl methylpropional itself leaves; hydroxycitronellal itself rises.
3 · Odour space
“lily” 20.0 → 0.0, “sweet” 20.0 → 0.0; no descriptor on record: 45.0 → 65.0.
4 · Perceived character
A modeled reading, not a prediction: no “lily” and “sweet” left.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 15.0 / 85.0 / 0.0; half gone at 4.9 h → 4.9 h.
6 · Odour activity
butylphenyl methylpropional: no threshold on record; hydroxycitronellal: no threshold on record.
7 · Regulation flags
Flags leaving the accord: Regulation (EU) 2021/1902 (substances classified CMR, butylphenyl methylpropional among them); IFRA Standard: p-tert-Butyl-α-methylhydrocinnamic aldehyde (p-BMHCA).
Why did this change matter?
  1. It takes the “Regulation (EU) 2021/1902 (substances classified CMR, butylphenyl methylpropional among them)” flag (a prohibition (Annex II)) off the accord: it names butylphenyl methylpropional and no other component.
  2. It takes the “IFRA Standard: p-tert-Butyl-α-methylhydrocinnamic aldehyde (p-BMHCA)” flag (restriction / prohibition) off the accord: it names butylphenyl methylpropional and no other component.
  3. butylphenyl methylpropional carried “lily” and no other component does on record, so the word leaves the modeled profile.
Replace HICC with hydroxycitronellal
1 · Composition
HICC 5 → 0 parts (5.0 → 0.0 per 100); hydroxycitronellal 40 → 45 parts (40.0 → 45.0 per 100); the others ×1.00.
2 · Molecules
HICC itself leaves; hydroxycitronellal itself rises.
3 · Odour space
no descriptor on record moves; no descriptor on record: 45.0 → 45.0.
4 · Perceived character
A modeled reading: the words on record barely move, and neither does the dry-down.
5 · Dry-down
top / heart / base 15.0 / 85.0 / 0.0 → 15.0 / 85.0 / 0.0; half gone at 4.9 h → 4.9 h.
6 · Odour activity
HICC: no threshold on record; hydroxycitronellal: no threshold on record.
7 · Regulation flags
Flags leaving the accord: Regulation (EU) 2017/1410 (HICC, atranol and chloroatranol prohibited); IFRA Standard: 3 and 4-(4-Hydroxy-4-methylpentyl)-3-cyclohexene-1-carboxaldehyde (HMPCC).
Why did this change matter?
  1. It takes the “Regulation (EU) 2017/1410 (HICC, atranol and chloroatranol prohibited)” flag (a prohibition (Annex II)) off the accord: it names HICC and no other component.
  2. It takes the “IFRA Standard: 3 and 4-(4-Hydroxy-4-methylpentyl)-3-cyclohexene-1-carboxaldehyde (HMPCC)” flag (restriction) off the accord: it names HICC and no other component.

What the model reads, with its evidence

hydroxycitronellal
descriptors: not on record · class: heart (role) · threshold: not on record · regulation: IFRA Standard: Hydroxycitronellal (restriction); Regulation (EC) No 1223/2009 on cosmetic products (a labelling condition (Annex III)) · provenance:
butylphenyl methylpropional
descriptors: sweet, lily
documented

ENE-1 · listed on the record

field: molecule/p-bmhca · descriptors · in this repository: ENE-1

· class: heart (role) · threshold: not on record · regulation: IFRA Standard: p-tert-Butyl-α-methylhydrocinnamic aldehyde (p-BMHCA) (restriction / prohibition); Regulation (EC) No 1223/2009 on cosmetic products (a labelling condition (Annex III)); Regulation (EU) 2021/1902 (substances classified CMR, butylphenyl methylpropional among them) (a prohibition (Annex II)) · provenance: Butylphenyl methylpropional prohibited in EU cosmetics, 2022 (documented)
2-phenylethyl alcohol
descriptors: floral, honey
documented

OT-1 · listed on the record

field: molecule/phenylethyl-alcohol · descriptors · in this repository: OT-1

· class: heart (role) · threshold: 140 µg/kg in water
measured

OT-1 · listed on the record

field: molecule/phenylethyl-alcohol · threshold · in this repository: OT-1

· regulation: none on record · provenance: occurs in rose absolute; occurs in rose otto
linalool
descriptors: sharp, medicinal
house interpretation

the house’s own reading, labelled as such

field: molecule/linalool · descriptors · in this repository: /materials, “Lavandula angustifolia”

· class: top (volatility, modeled) · threshold: 3.2 ng/L in air
measured

OT-2 · listed on the record

field: molecule/linalool · threshold · in this repository: OT-2

· regulation: Regulation (EC) No 1223/2009 on cosmetic products (a labelling condition (Annex III)) · provenance: occurs in bergamot essential oil

the model’s input as data: /data/causal/muguet-teaching-accord.json · the model and its formulas: the site’s build

Sources

  1. OSMO-1 · Osmothèque, Conservatoire international des parfums. Collection Osmothèque, juin 2015 (PDF list: name, house, launch year, concentration, family, sub-family, perfumers). www.osmotheque.fr/wp-content/uploads/2016/04/Col… · read 2026-09-30 · no reuse licence stated; quoted row by row, not filed
  2. ENE-1 · Fan XY, et al. Applications of Ene-Reductases in the Synthesis of Flavors and Fragrances. J Agric Food Chem. 2024;72(33):18305–18320. doi.org/10.1021/acs.jafc.4c02897 · read 2026-09-30 · CC BY-NC-ND 4.0
  3. KLE-1 · Kleoff M, Kiler P, Heretsch P. Synthesis of odorants in flow and their applications in perfumery. Beilstein J Org Chem. 2022;18:754–768. doi.org/10.3762/bjoc.18.76 · read 2026-09-30 · CC BY 4.0

These records are published under CC BY 4.0.

this record as data: /data/accord/muguet-teaching-accord.json · version 1 · updated 2026-09-30 · confidence B