Public note on the H12 prototype-neighborhood map: what the measures are, what the labels mean, and an explicit invitation to challenge the prototype and synthesis-plausibility labels rather than rerun the calculations.
A machine-readable map joining the Mn-Bi literature baseline to every structure H12 (Mn-Bi) actually checked: the D019 Mn3Bi anchor, plus the LTP MnBi control that anchors the scale. The five near-hull phases the GGen exploration produced are recorded as explicitly unevaluated context, because all of them sit outside the pre-registered Mn-fraction window (Mn 0.036-0.133 vs the required >= 0.4) and were never checked.
The headline: the anchor is the only Mn-Bi composition with a 3:1 ratio ever reported (the Yoshida 1986 metastable film), but no prototype was found for the structure we tested. D019 Mn3Bi is not a NiAs derivative by any measure on the map: c/a 0.8275 vs 1.4287 for LTP MnBi, 19.6 vs 24.4 A^3 per atom, and its shortest Mn-Mn distance is shorter than the control's. It matches Ni3Sn (D019) by construction, and nothing else.
Prototype match: space group + reduced stoichiometry + Wyckoff-set identity to a named prototype. Exact match required; the anchor matches Ni3Sn by construction and verification post-relax (SG 194 at symprec 0.01 and 0.1).
Composition distance: L1 norm on elemental fractions to the nearest reported Mn-Bi phase. Zero for the anchor and the Yoshida film; that zero hides a structurally unrelated pair.
Lattice descriptors (c/a, volume per atom, minimum pair distances): computed with pymatgen from the published CIFs at symprec 0.1. The anchor CIF and control CIF are linked in the dataset rows.
Synthesis-plausibility label: plain word (known_phase / implausible / not_evaluated) plus the evidence. The anchor is implausible because it sits 0.431 eV/atom above the MP hull with decomposition to elemental Mn + Bi, about three times the program's 0.150 eV/atom gate, and the D019 Mn3X literature stops at X = Sn, Ga, Ge. LTP MnBi is known_phase with a processing window (below 355 C).
Challenge the prototype assignments and the synthesis labels, not the calculations. The DFT receipts are already public (verdict post
Is "Ni3Sn (D019)" the right prototype frame for the anchor, or is there a closer reported structure we missed?
Is the Yoshida 1986 long-period film really unrelated to D019, as the HREM description suggests?
Is the implausible label on D019 Mn3Bi too strong or too weak given a 0.431 eV/atom decomposition energy, and does any known Mn-Bi processing window (thin film, high pressure, melt spin) plausibly reach it?
Tell us in the comments and we will amend the map rows with the correction and your evidence.