Amend venus closure-authority text to the widened item 6 (coder OR db closes, closure authority is not push authority)

Item 6 of the fleet Push-to-Done was AMENDED by pm-llmmsgsrv-cc within the hour, after coder-mars-cc pushed back that the draft said CODER while Elazar had said "any coder and db agent". Venus had already synced the narrow wording in 811bdf9 (VENUS-381), so venus was carrying text that excluded the db lane. This is not cosmetic in venus specifically: db-venus-cc has NO push authority and routes every artifact through coder-venus-cc, so under coder-closes its work has no one who can write the terminal transition - the exact hole ruling (a) was opened to close, one lane over. Amended in three places: venus-commons.md PM-on-Demand closure paragraph, venus-pm-autonomy.md post-push loop step 2, pm-venus-role.md review flow. All three now say the agent that DID the work closes, coder or db, and carry the new sentence closure authority is not push authority - a handoff commit you cannot push is still one you close, citing the pushers SHA. That sentence is pm-llmmsgsrv-cc own, explicitly NOT Elazar words, resolving a gap his ruling did not address. Fleet-side item 2 + item 6 + evolutiva-pm-commons FINISHED Banner section still NOT landed: they land as ONE change after the pluto/terra objection window, and whoever takes it must /lock each path first since coder-mars-cc and I both hold the right.