[gmickel]

Why

Two goal-mode stalls are removed in the skill text. The final whole-diff review follows the invocation's --review or the review config, and is skipped with the skip recorded when no backend is set, instead of ending the goal before its final pull request. A goal's child pull requests into its integration branch merge without the review-bot patience window; the final pull request keeps it.

Review focus

Start with goal-mode.md's final pull request bullet. Look hard at land's workflow.md patience paragraph: the waiver must hold only when the caller names the PR as a goal child and the observed base is that goal's goal/<name> branch, so a normal --until=merge PR still waits. Feedback wanted: is the two-condition waiver tight enough, and should the skip outcome strings match the receipts' existing skipped(config: ...) grammar exactly.

Needs you

• unverified: Full suite and ruff: Local run at f44cc1f: 4067 ran, 0 failures; ruff clean. No stored receipt; CI reruns it
• unverified: Live goal run: No --goal run exercised the skip or the waiver; the change is skill text

What changes for a user or operator

A repository with no review backend can now finish a --goal run. Under --until=merge, goal children no longer sit for land.patienceMinutes (10 minutes by default) each; only the final pull request into the default branch waits for bots.

Reading order

2. Final review follows --review and the config [R1, R2, R3]

Check the final review takes --review when passed and skips with a recorded outcome on none or ASK, while a configured backend that fails still stops the run.

Start at plugins/flow-next/skills/flow-next-flow/references/goal-mode.md

3. Child pull requests skip the bot wait [R3]

Check the waiver needs both the goal-child authorization and an observed goal/<name> base, and that the final pull request still waits.

Start at plugins/flow-next/skills/flow-next-land/workflow.md

4. Docs and changelog [R4]

Check the docs say what the skills now do; the Codex mirror is regenerated.

Start at CHANGELOG.md

Safe to skim: 1. Why goals stalled; 6 mechanical files (group 4)

Tradeoffs

Decisions: fn-305 was marked ready because the person invoked flow --auto on it by name. QA skipped (pipeline.qa=off). R2 was confirmed by reading: auto.md forwards --review only to plan, plan-review and work, and goal-mode's final review did not take it; now it does. The waiver lives in land's existing patience rule, not a new config key or flag. The redesign spec fn-306 stays out of this PR. flow-next.dev's goal pages follow with the release's downstream docs pass.

Verified

• Implementation review: SHIP, Codex gpt-6-astra, three reviewers (correctness, contracts, integration), no findings
• Focused tests: test_flow_auto, test_flow_merge_destination, test_codex_flow_dispatch, test_pilot_backlog_mirror_safety, test_land_config: 39 OK
• Codex mirror: ./scripts/sync-codex.sh --check exit 0
• ASK on an unconfigured repo: flowctl review-backend <spec> reads ASK, source none, in a fresh repo

https://claude.ai/code/session_01LbHwMnkD6ovBw6n8cnD8hQ

[Copilot]

🟢 Approval recommended

No active review issues are recorded for the goal-mode final-review and child-PR patience updates.

No open issues 🎉

What changed in this PR

This PR updates Flow-Next’s goal-mode skill instructions to avoid stopping at an unconfigured final review and to remove redundant review-bot waits for child pull requests.

Changes:

• Final whole-diff review honors --review, records skips when review is disabled or unconfigured, and preserves stops for configured-backend failures.
• Child PRs waive bot patience only with goal-child authorization and an observed base matching that goal’s integration branch; final PRs retain normal patience rules.
• Updates runtime documentation, generated Codex mirrors, the changelog, and implementation records.

File	Description
plugins/​flow-next/​skills/​flow-next-land/​workflow.md	Adds the scoped goal-child patience waiver.
plugins/​flow-next/​skills/​flow-next-flow/​references/​tail.md	Defines goal-child authorization passed to land.
plugins/​flow-next/​skills/​flow-next-flow/​references/​goal-mode.md	Updates final-review selection, skip recording, and child handoff.
plugins/​flow-next/​docs/​skills.md	Documents final-review configuration and skipping.
plugins/​flow-next/​docs/​orchestration.md	Documents child versus final PR patience.
plugins/​flow-next/​docs/​flowctl.md	Updates the patience configuration contract.
plugins/​flow-next/​codex/​skills/​flow-next-land/​workflow.md	Mirrors the landing waiver for Codex.
plugins/​flow-next/​codex/​skills/​flow-next-flow/​references/​tail.md	Mirrors goal-child authorization guidance.
plugins/​flow-next/​codex/​skills/​flow-next-flow/​references/​goal-mode.md	Mirrors goal-mode review and handoff changes.
plugins/​flow-next/​codex/​docs/​flow-next/​skills.md	Mirrors the updated skill documentation.
plugins/​flow-next/​codex/​docs/​flow-next/​orchestration.md	Mirrors the patience documentation.
plugins/​flow-next/​codex/​docs/​flow-next/​flowctl.md	Mirrors the configuration contract updates.
CHANGELOG.md	Announces both goal-mode fixes.
.flow/​tasks/​fn-305-goal-runs-finish-without-a-review.1.md	Records implementation results and verification evidence.
.flow/​tasks/​fn-305-goal-runs-finish-without-a-review.1.json	Records completed task metadata.
.flow/​specs/​fn-305-goal-runs-finish-without-a-review.md	Defines acceptance criteria and scope boundaries.
.flow/​specs/​fn-305-goal-runs-finish-without-a-review.json	Records spec completion and review metadata.

_{Review effort: Balanced}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 3 potential issues.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7dfe5ae. Configure here.}