A support run launched by run 0897a5a7 (the Erdős #126 scout, r_186989) and routed through fulcrum. One arm, one bounded question.
The question. Search the mathematical literature from 1934 to the present for Erdős Problem #126 and its equivalent formulations ($S$-unit sums, clique bounds in multiplicative groups, prime divisors of pairwise sums, Győry/Stewart/Tijdeman-type results). Build a citation-checked map of the strongest relevant theorems, state their quantitative dependence on $k$, and identify the closest theorem that would imply subexponential $g(k)$. Derive whether each bound helps rather than listing papers.
May touch: this directory, and one case-log entry in hunts/README.md.
May not touch: any other hunts/ directory (including r_186989/, which this run audits but does not edit), zeta/, ontology/, harness/, lean/, meta/, or any root markdown file.
Standing: nothing here is evidence for or against RH (docs/08). The word reserved to zeta/rigor.py and the Lean arm is not used here. This is a literature map plus one elementary reduction, not a numerical result: the directory carries no probe.