Enamel caries graph: remove workflow arrows and unsupported cross-feeding and competition #1377

Review of Defined_Multispecies_Enamel_Caries_Model.yaml (CommunityMech:000082) against Arthur et al., PMID:23446436. The complete primary paper was read from a public PDF copy; Tables 1 and 2 were inspected visually. No full article is being added to the repository.

The biofilm-to-count and biofilm-to-mineral arrows describe assay context/readouts, not demonstrated causal transitions between interactions. Remove them. The three observations do not demonstrate colonization facilitation, metabolite cross-feeding or reciprocal competition, so remove these signs and retain the positive treatment outcomes.

Represent the four taxa as the inoculated roster, not proof of persistence in every arm. Calcium did not significantly alter total CFU or S. mutans counts; S. sanguinis was below detection in the calcium arms. Bound the mineral outcomes to the actual endpoints and comparisons, not a uniform monotonic response. Acid selection and calcium/EPS explanations remain hypotheses, not selectively tested mechanisms.

Deduplication: 706 all-state issue titles/bodies and ignored/hidden-inclusive local reports, research, history, tests and scripts; relevant #401 and #754 read. Not every issue comment was searched.