Repository navigation
Description
Preserved branch — decision needed
Branch: rescue/3041-631-subagent-ancestor-2026-11-01
Repo: developing-today/opencode
HEAD: edebd4f6b5f941c308cf8bea58c04e0133817c11
What this branch holds
The local line of queue/3041-631-subagent-ancestor as it stood in the laptop fleet worktree ~/git/fleet/3041-631-subagent-ancestor: origin's queue tip 611f6529da plus 16 local commits — a 2026-10-29 merge of origin/main into the queue branch (9bca869ad), the main-line commits that merge brought in, and a rewritten rescue commit on top. Against origin's current queue tip, the tree differs in 29 files (+1086/−794: core model-route / session-step-failure / tool-execute tests, services/www/openapi.json, codemode and route-health cooldown work). This is the variant that existed only on the laptop; origin's own queue/3041-631-subagent-ancestor was never touched.
How it got here
On 2026-11-01 the laptop worktree cleanup rescued this tree's uncommitted work into commit 9efb5a9e35, but that commit had swept in the tree's untracked .playwright-browsers/ download (603 paths, two binaries over GitHub's 100 MB per-file limit), so GitHub's pre-receive hook declined every push (GH001) — tracked as #1203. The commit was rewritten with .playwright-browsers/ excised (tree diff vs the original commit: exactly the 603 browser paths, nothing else) and pushed under this rescue name; the push was verified by ls-remote and the worktree was then removed. The original commit remains preserved in a git bundle on the laptop (~/backups/github-rescue-bundles/opencode-queue-3041-631-subagent-ancestor-9efb5a9e35.bundle). Full story: #1203.
Decision needed
Merge it, cherry-pick from it, or drop it. Nothing else references this branch — no worktree, no open PR — so deleting the branch later loses only what is described above.
(Filed as part of the 2026-11-01 worktree rescue: one issue per preserved branch.)
Acceptance criteria
- A decision — merge, cherry-pick, or drop — is recorded on this issue by the reviewer.
- If merge or cherry-pick: the resulting PR is linked here and the branch's unique commits are accounted for (landed, or explicitly not carried over).
- If drop: the branch is deleted, or explicitly retained with the reason recorded here.
- This issue is closed with the decision outcome; the branch does not stay in limbo.
Activity
developing-today commentedon Nov 1, 2026
Filed after the rescue completed; the full failure-and-rewrite history is in #1203 (closed as resolved-by-rescue).
dezren39 commentedon Nov 1, 2026
Intake review
Related: Origin issue: #631 (from the branch name). The branch is origin's queue tip plus 16 local commits (29 files, +1086/−794 against the queue tip) per the body.
Notes: Preserved-branch decision issue from the 2026-11-01 laptop worktree cleanup; it carried only priority/low at filing. The merge/cherry-pick/drop call is a review decision, hence user/review (Muse reviews first). Nothing else references the branch, per the body — this issue is its only record.
Risks: Dropping without reading the diff loses work that existed only in that worktree; merging stale work wholesale can resurrect superseded approaches.
Suggested approach: Diff the branch against current main, check whether the origin issue's work already landed by another route, then record merge / cherry-pick / drop.
🤖 Agent note · filed
dezren39 commentedon Nov 3, 2026
agent-note:v1
state: reviewed (coverage loop-back)
Coverage pass 2026-11-03, loop-back review: this issue was re-reviewed (drift / overlap / already-handled / stale / blockers / ordering) so every open issue in the four repos is reviewed and in the system. Verdict: still live; in the system as task 7141 (triage, queued, priority 30) in repo-health-opencode-pilot. Labels audited and synced. If it is already handled at current main, its triage closes it with evidence; if oversized, the triage files the split; blockers belong on this issue.
Found via: coverage loop-back 2026-11-03
dezren39 commentedon Nov 3, 2026
Triage verdict (task 7141, lane opencode-a, 2026-11-03): Decision: CHERRY-PICK — this preserved branch, unlike its siblings, holds real work that never landed.
Verified fresh this shift (all via
gh apiat main3e483a8cafd8):rescue/3041-631-subagent-ancestor-2026-11-01atedebd4f6b(matches the HEAD recorded above) and origin'squeue/3041-631-subagent-ancestorat611f6529d.compare main...rescue/3041-631-subagent-ancestor-2026-11-01: diverged, ahead 3 / behind 685. The other commits on the line are old main-line PR merges (ci: basic tier named as the merge requirement + event bots consolidated into repo-sweeper (every 15m) #832, ci: repo-sweeper daily closers are schedule-only (dispatch can never close) #834, feat(codemode): versioned libraries and std v1 pure helpers #812, fix(core): escalate route-health cooldowns on repeated failure #826, chore(client): regenerate generated client #815, fix(www): regenerate services/www/openapi.json from protocol #836) already on main. The delta:packages/app/src/session/requests/model.ts+16 — acreateEffectsyncing the root family (data.session.sync(rootID, { children: true })) whenever a subagent view makes a different session the owner, so the running menu (which walks the owner's whole tree) receives the root's other subagents. Main today syncs the root session + messages on owner change but never syncs children — this is absent, not superseded.packages/app/e2e/regression/subagent-child-navigation.spec.ts+3/−1 — ancestor-path heading expectation correctedchildTitle→taskDescription; main still assertschildTitle.packages/app/e2e/regression/timeline-stream.spec.ts+5/−1 — background-shortcut trigger name corrected to the full"Search callers +2 working"; main (line 1598) still expects"+2 working".Disposition: cherry-pick enqueued as task 8031 ("Fix #1280: cherry-pick the #631 root-family sync from the preserved rescue branch") in
repo-health-opencode-pilot, carrying the full salvage accounting (unique commits611f6529d+edebd4f6b; port the 3-file delta, run both e2e specs, disposition with evidence if the port proves unnecessary). Both branches are explicitly retained until task 8031 lands — they are the only copy of this work — and task 8031 deletes both and notes it here after its PR merges. The branch is not in limbo: its fate is owned by task 8031. Closing this record issue with the decision outcome per its acceptance criteria.