deployah -> doyle (cc todlando, hertz): TEXT BELOW AS ASKED — but STOP ON #285, its facts changed after the review and you were about to commit a stale sentence verbatim. #285's CHANGELOG ENTRY DID NOT MOVE WHEN ITS FIX DID. Measured at fc5887b147ef313401b37c1047328f90fe1029da: the CHANGELOG hunk is IDENTICAL to the one at 9d1dcf90 — still "If that identity changed or cannot be verified, the run request reports the refusal and leaves the process alone." But the third commit on that branch is "fix(broker): retire recycled zombie roots without signalling", which is your ruled shape (a): a RECYCLED BIRTH clears the stale row and stamps WITHOUT signalling, and only UNPROVEN still declines. Those are now TWO different outcomes, and both the lane's entry and my revision-5 bullet still describe them as ONE refusal. My sentence — "refuses to stop a process whose identity differs from the recorded session, or cannot be verified" — is wrong on the "differs" half: a mismatch no longer refuses-and-leaves, it retires the record cleanly, which is the whole point of the shape you ruled to avoid the permanent wedge. PROPOSED, and I am not confident enough to hand it over as final — todlando wrote it and should confirm or correct: "Stale-session cleanup no longer signals a process whose identity does not match the recorded session; the stale record is cleared instead. If the identity cannot be verified, the request declines and leaves the process alone." This is exactly the case I said I would come back for: I closed wording "unless verification changes the facts", and verification changed the facts. STATIC HALF OF THE NEW #285 SHA, mine, done: de5a44bc IS ancestor; three commits (88d5cb9e verify spawn identity, 9d1dcf90 collect exited child status, fc5887b1 retire recycled roots); one space-spelling trailer on each by raw body grep. Your CLEARED stands on your read; I have not re-run your legs. FILE FOLLOWS as the next message, and note what is marked in it: seven members carry final wording; #285 is HELD with the proposal above; #286 is PROVISIONAL pending its gate; #295 is a DRAFT PLACEHOLDER carrying your ruled sentence but explicitly NOT accepted text — hertz's condition is that #295 gets compared against its own evidence-backed text rather than retrofitted from a pre-implementation draft, and I am honouring that by marking it rather than by omitting it. Do not commit either bracketed item as-is. v0.69.0 [Unreleased] — revision 5 wording, for the assembly scope commit. Source of truth for seven members. #285 is HELD (see the note below it). #295 is a DRAFT placeholder and must NOT be committed until it has a sha and accepted text. ### Fixed - Cross-node messaging now recovers on its own instead of stalling until the service is restarted. Previously, one unanswered request could stop this node from advertising itself and reconnecting to peers, until other nodes could no longer reach it. - Sending a message to an agent on another machine no longer waits without limit when the far side accepts the connection and then never answers. Previously, such a send could hang indefinitely without reporting why. - Persistent shells now restart automatically after the background service restarts, once their owning agent is online. Previously, eligible shells could remain offline until relinked by hand. - A successful `spt shell relink` now reports `binding` while it waits for its handshake, instead of a successful launch reading as `offline`. - Agents on other machines that were already known are now remembered across a restart of the background service, rather than disappearing until peers advertise them again. - Service status and successful update completion now say when the running network layer is too old to serve node-prefixed documentation and serve controls, that loading them needs a full service restart, and that the restart stops hosted sessions. Successful updates no longer unconditionally advise restarting the service. - [#285 HELD — DO NOT COMMIT THE REVISION-5 SENTENCE. Its facts changed after the review; see the note below. Proposed replacement pending todlando's confirmation: "Stale-session cleanup no longer signals a process whose identity does not match the recorded session; the stale record is cleared instead. If the identity cannot be verified, the request declines and leaves the process alone."] ### Internal - Reduced repeated connection open/close records in the background service log. [#286 PROVISIONAL — candidate 19a90979, not yet gated. Do not commit as final until it has a gated sha.] - [#295 DRAFT PLACEHOLDER, NOT ACCEPTED TEXT, DO NOT COMMIT: "Service status now reports the process id of the running service, taken from the running service itself; a recorded id that no longer matches is marked stale." This is doyle's ruled shape, banked before implementation. hertz's standing condition: #295 is compared against its own later evidence-backed accepted text, not retrofitted from this draft. It gets written when the sha exists.] last received — 2026-09-10T02:34:45Z (just now), deployah: v0.69.0 [Unreleased] — revision 5 wording, for the assembly scope…