ayudarg (com-ar + solo): drop the build from 4 COMPOUND pnpm lint && pnpm build hook lines - deferred until either repo is staffed

SPUN OUT OF OPS-221 as the remainder that cannot be dispatched, so it is a row with a named trigger instead of a hold in a PM head. THE FOUR LINES, verbatim as they read 2026-09-12 (read by pmmasterassistant-evolutiva-cc from file content, not from a grep hit count - see the OPS-221 scope correction for why that distinction cost us two false dispatch targets): - ayudarg-com-ar/.gitpush-pre.sh:86 -> cd app && pnpm lint && pnpm build - ayudarg-com-ar/app/.gitpush-pre.sh:2 -> pnpm lint && pnpm build - ayudarg-solo/.gitpush-pre.sh:86 -> cd app && pnpm lint && pnpm build - ayudarg-solo/app/.gitpush-pre.sh:2 -> pnpm lint && pnpm build THE EDIT IS NOT A LINE DELETION AND THAT IS THE WHOLE RISK. All four are COMPOUND. Drop only the trailing "&& pnpm build" and keep "pnpm lint" (and keep the "cd app &&" prefix on the two root hooks). Anyone who deletes the whole line removes lint coverage from that repo - precisely the regression the rejected QUICK_MODE alternative was rejected for on the parent row. The three staffed repos (terra:854, venus:684, pluto:607) were each a STANDALONE pnpm build line; these four are the only compound ones in the fleet, so muscle memory built on the other three is wrong here. WHY DEFERRED RATHER THAN DISPATCHED, and this is the reasoning to argue with rather than the conclusion: both repos have no registered agent on the hub roster, and Elazar rates ayudarg low priority. The tempting move is to have any capable agent make a four-line mechanical edit. It is the wrong move because THE CHANGE RIDES THE NEXT PUSH AND AN UNSTAFFED REPO HAS NO NEXT PUSH. Editing there produces an uncommitted working-tree change, of unknown provenance, in a repo nobody is watching, that sits until whoever eventually staffs it finds a dirty tree they did not make and cannot date. That is strictly worse than the current state, because the current state is honest. The saving is also zero until someone pushes - the build cost this removes is only ever paid by a push that is not happening. TRIGGER, so this does not rot silently: do it as part of the FIRST push by whoever staffs ayudarg-com-ar or ayudarg-solo. Whoever takes either repo gets this row with it. At that point it needs no PM decision, it needs the four-line edit in the same commit as their first real work. ACCEPTANCE: name each of the four files, quote the resulting line, and confirm pnpm lint still runs in each hook. Do not report clean - report which hooks were exercised. A verdict of "removed the build" without the surviving lint line quoted is not acceptance here, because dropping lint is the failure mode this row exists to prevent.