Repository navigation
Open
Labels
effort/lLarge: multi-day or open-ended; expect staged check-insLarge: multi-day or open-ended; expect staged check-inskind/featureNew feature or capability requestNew feature or capability requestmeta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels addedpriority/normalOrdinary priority (the default)Ordinary priority (the default)queue/queuedHas a queued task in the Supabase agent queueHas a queued task in the Supabase agent queuesource/humanFiled directly by a humanFiled directly by a humanstatus/parkedDeliberately paused with a recorded resume conditionDeliberately paused with a recorded resume conditiontask/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this itemwork/assignedQueue worker lifecycle stateQueue worker lifecycle state
Description
Implement the approved serve-all design from #5698.
Approved design (Drew, 2026-11-05, on #5698): option (i) — the multicall list. One fleet serve daemon process serving multiple roles, driven by the multicall serve list.
Acceptance criterion (Drew's added requirement, 2026-11-05): the serve list must be able to make the equivalent of all — it must be able to express/produce the equivalent of the complete serve set, not a subset. An implementation that can only represent part of the serve surface does not satisfy this issue.
Design record and discussion: #5698. Related: #5412 (one fleet multicall binary epic), #5699 (ports vs local transports).
Acceptance criteria
fleet serveruns as one daemon process serving multiple roles driven by the multicall serve list, per the approved design on design: one-daemon serve-all — multi-rolefleet servein a single process #5698 (option (i)).- Drew's added requirement holds: the serve list can express/produce the equivalent of ALL of the complete serve set, not a subset — demonstrated by expressing the full current serve surface through the list.
- Existing single-role serve behavior is unchanged for roles migrated onto the list (no dropped roles, ports, or transports — cf. design: ports vs local transports — Unix sockets for same-machine fleet roles, dual-stack with proxy TCP #5699).
Metadata
Metadata
Assignees
Labels
effort/lLarge: multi-day or open-ended; expect staged check-insLarge: multi-day or open-ended; expect staged check-inskind/featureNew feature or capability requestNew feature or capability requestmeta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels addedpriority/normalOrdinary priority (the default)Ordinary priority (the default)queue/queuedHas a queued task in the Supabase agent queueHas a queued task in the Supabase agent queuesource/humanFiled directly by a humanFiled directly by a humanstatus/parkedDeliberately paused with a recorded resume conditionDeliberately paused with a recorded resume conditiontask/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this itemwork/assignedQueue worker lifecycle stateQueue worker lifecycle state
Activity
developing-today commentedon Nov 5, 2026
CODE-1837
dezren39 commentedon Nov 5, 2026
Design blocker cleared. Drew approved option (i), the multicall list design, on #5698 today (2026-11-05), with the added acceptance requirement: the serve list must be able to make the equivalent of all (the complete serve set, not a subset). This issue is the implementation vehicle for that approved design.
dezren39 commentedon Nov 6, 2026
Intake review
fleet servein a single process #5698 (approved by Drew 2026-11-05, option (i) the multicall list); epic: onefleetmulticall binary — search, queue, dispatch, proxy — with capability-selected builds #5412 (onefleetmulticall binary epic); design: ports vs local transports — Unix sockets for same-machine fleet roles, dual-stack with proxy TCP #5699 (ports vs local transports). Background in the tracker: in fleet 0.1.0 the binary's laptop-agent domain is still a registry stub per the house notes — worth re-verifying at pickup.🤖 Fleet monitor intake (repo-fleet-monitor), 2026-11-06.
2 remaining items
dezren39 commentedon Nov 6, 2026
Handed off to opencode claimant per execution-lane assignment (task 11043, issue #6264).
fleet serve(design #5698) #6264 OPEN, no hard-stop labels (held / user/owned / do-not-merge / do-not-auto-edit).dezren39 commentedon Nov 6, 2026
🤖 Agent note · handed off
fleet serve(design #5698) #6264 to the opencode execution lane — assigned_lane=opencode per fleet queue lane; the opencode-claimant launches the bound session.fleet serve(design #5698) #6264dezren39 commentedon Nov 6, 2026
🤖 Agent note · handed off
fleet serve(design #5698) #6264) is assigned_lane=opencode per instruction §1b; handing off to the opencode-claimant — this lane does not work it directly. Revalidated fresh via gh: issue implement: serve-all — multicall list, one-daemon multi-rolefleet serve(design #5698) #6264 OPEN, no hard-stop labels. Attempt 3/6.fleet serve(design #5698) #6264dezren39 commentedon Nov 6, 2026
🤖 Agent note · handed off
fleet serve(design #5698) #6264 — implement serve-all, one-daemon multi-rolefleet serve, design design: one-daemon serve-all — multi-rolefleet servein a single process #5698) claimed by lane code-z and handed off — execution lane assignedopencodebyfleet queue laneper worker instruction §1b; this lane does not work it directly. Issue implement: serve-all — multicall list, one-daemon multi-rolefleet serve(design #5698) #6264 verified OPEN with no held, user/owned, do-not-merge, or do-not-auto-edit labels (fresh gh check 2026-11-06T17:47Z). Attempts 3/6.fleet serve(design #5698) #6264dezren39 commentedon Nov 6, 2026
🤖 Agent note · handed off
opencodefleet serve(design #5698) #6264 (triage). Fresh check: issue implement: serve-all — multicall list, one-daemon multi-rolefleet serve(design #5698) #6264 is OPEN, no hard-stop labels (held / user/owned / do-not-merge / do-not-auto-edit). Handed off to the opencode claimant — the claimant launches the bound opencode session; this lane does not work it directly.fleet serve(design #5698) #6264