Repository navigation
Conversation
… pstack's defaults Sync from udecode/dotai 351c30a, which carries fdc8c20, at the v0.9.88 pin, plus every shared commit since plate's last sync from e23bf33. The owner asked to cut the workflow's overhead while keeping its output, as close to pstack's defaults as possible. Every PR still keeps its plan and page; work that ends in one session without a PR hands back in chat. Offers still go to the ledger, and the hand-back says how many and that /chips cards them; a sweep runs only on /chips and once every task started from this session's cards has landed or been dropped, and that sweep cards the next wave. A change to agent instructions alone gets one panel round, and a failed seat reruns once on its own model, then the run stops. docs/development/agent-skills.md says the same. Who gains or loses what: Claude Code and Codex sessions here stop sweeping at each hand-back and after compaction, and agent-instruction changes get one panel round in both.
Review or Edit in CodeSandboxOpen the branch in Web Editor • VS Code • Insiders |
|
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
The owner asked to cut the agent workflow's overhead while keeping its output, as close to pstack's defaults as possible.
What changed
This branch syncs from udecode/dotai@351c30a, which carries fdc8c20, at the v0.9.88 pin:
/chipscards them. The sweep runs only on/chipsand once every task started from this session's cards has landed or the owner dropped it, and that sweep cards the next wave instead of starting it.docs/development/agent-skills.mdsays the same.It also carries every shared commit since plate's last sync from e23bf33 (compare): per-run plan page files with
page-stop.mjs, cards that open in their own worktree and retitle themselves once landed, the Time budget and Codex worker rules moved intotime.mdandcodex.md,seats.mjsfor panel seats, a stop when a models-sheet model cannot run, and several panel, chips and cleanup fixes. The trunk-delivery commits among them render nothing here.The one-phase-per-session rule for long runs is trunk-only and does not render here. plate-2's verdict playbook now sweeps at a verdict's close; plate gets that with its next workflow sync from plate-2.
Verification
sync-pstack verifypasses here, and dotai's sync-pstack suite passes 317 tests.Who gains or loses what: Claude Code and Codex sessions here stop sweeping at each hand-back and after compaction, and agent-instruction changes get one panel round in both. Work that ends without a PR hands back in chat in both.