Repository navigation
feat(skills): /quest-run and /quest-complete plan the follow-ups they surfaced - #664
Conversation
… surfaced When a run finishes, it runs /quest-plan on each problem it surfaced that no quest or open PR owns, in the repository whose quest tree owns it, instead of only offering to. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…, update descriptions Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…nish up Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…cked to merge Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Design change: at the round 3 checkpoint, the maintainer chose to plan the kept cross-repository follow-ups of every PR picked to merge, not only merged ones, since /quest-complete queues without waiting (rejected: land as is; drop cross-repository follow-ups). The next review restarts with (Written by Claude Opus 5.5) |
…ning in quest-run Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… state Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 27 minutes. View limit details
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. |
There was a problem hiding this comment.
✅ No new issues found.
Reviewed changes: the full diff of the skill text and generated stubs.
assets/skills/run.md: the Loop's finish now runs/quest-planon the run's follow-ups before the summary. "Finish up" plans nothing and lists the follow-ups left to plan. The description is updated.assets/skills/complete.md:/quest-completeplans its follow-ups and writes the summary as the Loop's finish describes. The description is updated.- README and stubs: the README skill table and the
.agentsand.claudestubs match the new descriptions.
claude-sonnet-5-5 | 𝕏
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…anned Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
✅ No new issues found.
Reviewed changes since the prior pullfrog review: skill text and descriptions only.
- Merged origin/main —
complete.mdand the README table now carry #649's waits handling alongside this PR's planning. - Owning checkout and landed picks —
run.mdplans each follow-up in a checkout of the owning repository and keeps landed picks' follow-ups unless the user withdrew the pick. - Summary names plan PRs — the final summary is written after planning so it can name them.
- Descriptions aligned — skill descriptions, README rows, and the four generated stubs all say "follow-ups no quest owns yet".
claude-sonnet-5-5 | 𝕏

When
/quest-runor/quest-completefinishes, it now runs/quest-planon the run's follow-ups instead of ending with an offer to plan them. There are two kinds:Each is planned in the repository whose quest tree owns it, before the final summary, so the summary includes the plan PRs. On "finish up", it plans nothing and lists each follow-up left to plan.
/quest-afkalready defers/quest-plansessions.assets/skills/run.md: the planning line in the Loop's finish, the "finish up" exception, and the description.assets/skills/complete.md:/quest-completestops before the Loop's finish, so it now plans its follow-ups and writes the summary as that finish describes, merged with feat(skills)!: fold /quest-unblock into /quest-complete #649's waits handling. Description updated.quest init.Decisions
The maintainer asked for this on 2026-11-12, after a
/quest-completerun on kixelated/token-max ended by offering to plan its follow-ups./quest-runand/quest-complete(rejected:/quest-completeonly).Design change:comment): ✅ plan the kept cross-repository follow-ups of every PR this run picked to merge, not only merged ones (rejected: land as is; drop cross-repository follow-ups).Public API
The run and complete skills' text changes. There are no CLI or format changes.
Validation
After merging origin/main at 6c97e48:
just checkandjust testpass on df60ab6, includingtest/skills.test.ts's inline size tests (53 pass).just checkandjust testpass on the first head. On the final head,just check,quest check, andtest/skills.test.tswithtest/init.test.ts(109 tests) pass.Follow-ups
None.
Models
Size: S
Implementer: Claude Code, claude-opus-5-5, unknown effort
Implementer: Claude Code, claude-opus-5-5, low effort (iterate, S)
(Written by Claude Opus 5.5)
🤖 Generated with Claude Code
Review
Rewritten by
quest reviewafter each round.Head: df60ab6
Base: origin/main at 9baa1b5
Reviewer: Codex, gpt-6.1-sol, medium effort
Session:
01a1d251-7381-76a0-b1bd-18d33620c32e, round 2Score: 9/10,
The incremental changes consistently clarify which follow-ups require planning and preserve the explicit “finish up” exception, with no actionable defects found by code inspection.Rounds: 3; findings 0: fixed 0, refuted 0, open 0
Findings: none
Reviewer: opencode, opencode-go/deepseek-v4.1-flash, high effort
Session:
ses_e2dae869cffedFaBDQSnYOlcTQ, round 2Score: 10/10,
All three earlier findings are settled: the landed-pick gap is fixed, the outside-the-loop planning is supported by the documented design and existing later-round handling, and the summaries now match line 71; the small documentation change is consistent across the assets, generated stubs, and README.Rounds: 3; findings 6 (medium 3, low 3): fixed 4, refuted 2, open 0
Findings: none
The last round's findings, at ebfe285:
assets/skills/run.md:71:Cross-repository follow-ups are dropped when their pick lands mid-run;assets/skills/run.md:71 now plans the follow-ups of this run's cross-repository picks "landed or not, unless the user withdrew the pick", so a pick that landed mid-run keeps its kept follow-up.assets/skills/run.md:70:Planning remains after the finish condition, outside the loop;run.md:67 and the finish condition at run.md:70 already leave PRs for a later round, and AGENTS.md stops quest work at PR creation, so plan PRs, which add quests and need the user's /quest-merge pick, are correctly left to a later /quest-run or /quest-complete and named by run.md:72.README.md:1870:The user-facing summaries claim more planning than line 71 performs;README.md:1870-1871 and both skill descriptions now say they plan "the follow-ups no quest owns yet", which matches run.md:71's unowned problems and unowned cross-repository follow-ups of this run's picks.