What this is. A support run launched by run 0897a5a7 to answer one bounded question: is there a genuinely joint multi-prime pruning argument for Erdős Problem #126? Concretely, can a block of $r$ primes be removed from $S$ while retaining more than $2^{-r+o(r)}$ of $A$, and if not, is there a rigorous obstruction to all coordinatewise-pruning schemes?
What it may touch. hunts/support_d176e366/ and its case-log entry in hunts/README.md. Nothing else. Not zeta/, harness/, lean/, meta/, any other hunts/ directory, or any root markdown file.
What it is not. Not a result about RH. Nothing in this directory is evidence for or against RH (docs/08). Not an attempt at the whole problem: the question is the block lemma and its obstruction, and the deliverable is a rigorous artifact for that lane or an honest "not settled".
Audit obligation. hunts/r_186989/RESULTS.md is input, to be audited rather than trusted.