From d00fc9a4a5cb414fe349421df428bdb52171a896 Mon Sep 17 00:00:00 2001 From: Cursor Agent Date: Mon, 9 Nov 2026 02:56:56 +0000 Subject: [PATCH 1/2] fix(orchestrate): spawn a fresh worker for built state New work that needs a checkout, uncommitted diff, or process goes to a fresh worker with that state in the brief. send_message stays a follow-up on the same work. Co-authored-by: Kaitian Xie --- skills/poteto-mode/playbooks/orchestrate.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/skills/poteto-mode/playbooks/orchestrate.md b/skills/poteto-mode/playbooks/orchestrate.md index 4648d60..c3e51f9 100644 --- a/skills/poteto-mode/playbooks/orchestrate.md +++ b/skills/poteto-mode/playbooks/orchestrate.md @@ -9,7 +9,7 @@ Run it with your own tools. Spawn workers with `invoke_subagent`. Keep the progr - **One owner per PR.** The owner is the only writer of its branch and carries the PR from build to merge. Only the owner pushes, rebases, or retargets it. Closing or retargeting a base PR orphans every PR above it, so the owner does that only on your call. You never author or edit code. Conflicted merges, chain rebases, and fixes go to owners. - **A fresh worker for each independent task.** Its kickoff carries everything it needs, including each upstream report it depends on, pasted in full. Workers can't see their siblings. Run workers in an isolated workspace (`Workspace: "branch"`) unless the work needs this machine's own checkout or state (`Workspace: "share"`): a `control-ui` or `control-cli` runtime check, local transcripts under `/brain/`, a simulator, or auth that exists only here. Pick each subagent's model tier for the job, per the Subagents section. - **A sub-coordinator only when one area has many parallel workers.** It owns that area's workers and kickoffs and sends you one rollup, never raw worker reports. Stop at three levels: you, sub-coordinators, and workers. Keep about ten workers in flight per coordinator, and start the next one as each finishes instead of waiting on a whole batch. -- **Resume a worker with `send_message` when new work needs that worker's built state,** such as its checkout, uncommitted changes, or a running process. Message a worker only for a follow-up on the same work. Never message one to check on it, because a message restarts an idle agent. +- **When new work needs that checkout, uncommitted diff, or process, spawn a fresh worker with that state in the brief.** Keep `send_message` only for a follow-up on the same work. Never message one to check on it, because a message restarts an idle agent. - **Each PR gets a verifier that isn't its author,** on a different model tier. The owner merges only on that verifier's pass at the current head. - **Start with one unit.** Take it through kickoff, owner, verifier, and merge before you fan out, so a flaw in the kickoff costs one agent instead of fifty. On near-identical cheap units, the first unit is that test. From 8d1691c77708bf2fcf936ec897c6b00ab59b7446 Mon Sep 17 00:00:00 2001 From: Cursor Agent Date: Mon, 9 Nov 2026 02:59:18 +0000 Subject: [PATCH 2/2] fix(orchestrate): swap only the fork verb Keep the upstream examples and the follow-up sentence. A fresh worker gets the built state in its kickoff. Co-authored-by: Kaitian Xie --- skills/poteto-mode/playbooks/orchestrate.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/skills/poteto-mode/playbooks/orchestrate.md b/skills/poteto-mode/playbooks/orchestrate.md index c3e51f9..5718fd2 100644 --- a/skills/poteto-mode/playbooks/orchestrate.md +++ b/skills/poteto-mode/playbooks/orchestrate.md @@ -9,7 +9,7 @@ Run it with your own tools. Spawn workers with `invoke_subagent`. Keep the progr - **One owner per PR.** The owner is the only writer of its branch and carries the PR from build to merge. Only the owner pushes, rebases, or retargets it. Closing or retargeting a base PR orphans every PR above it, so the owner does that only on your call. You never author or edit code. Conflicted merges, chain rebases, and fixes go to owners. - **A fresh worker for each independent task.** Its kickoff carries everything it needs, including each upstream report it depends on, pasted in full. Workers can't see their siblings. Run workers in an isolated workspace (`Workspace: "branch"`) unless the work needs this machine's own checkout or state (`Workspace: "share"`): a `control-ui` or `control-cli` runtime check, local transcripts under `/brain/`, a simulator, or auth that exists only here. Pick each subagent's model tier for the job, per the Subagents section. - **A sub-coordinator only when one area has many parallel workers.** It owns that area's workers and kickoffs and sends you one rollup, never raw worker reports. Stop at three levels: you, sub-coordinators, and workers. Keep about ten workers in flight per coordinator, and start the next one as each finishes instead of waiting on a whole batch. -- **When new work needs that checkout, uncommitted diff, or process, spawn a fresh worker with that state in the brief.** Keep `send_message` only for a follow-up on the same work. Never message one to check on it, because a message restarts an idle agent. +- **Spawn a fresh worker when new work needs a worker's built state,** such as its checkout, uncommitted changes, or a running process, and put that state in its kickoff. Message a worker only for a follow-up on the same work. Never message one to check on it, because a message restarts an idle agent. - **Each PR gets a verifier that isn't its author,** on a different model tier. The owner merges only on that verifier's pass at the current head. - **Start with one unit.** Take it through kickoff, owner, verifier, and merge before you fan out, so a flaw in the kickoff costs one agent instead of fifty. On near-identical cheap units, the first unit is that test.