overnight-review-client-delivery

Overnight autonomous B-vs-C parallel insight discovery that surfaces genuinely ah-ha findings from data, with cohort-conditional novelty gate, adaptive tuning, and agent-review-panel loop.

Calibrate review intensity per-PR (Tier 1 two-stage / Tier 2 combined single-agent / Tier 3 bash-only verification) when orchestrating a long chain (10+ PRs) of independent tasks overnight from a single implementation plan. Specializes subagent-driven-development's review step for throughput-tuned chains: prescribes three tiers, a decision rubric, and a concrete bash-verification recipe for low-risk visual-restyle PRs. Sister to subagent-driven-development (master pattern; this specializes the review step) and overnight-multi-issue-implementation (different shape: stacked PRs from issues vs. independent PRs from plan). Companion to large-redesign-parallel-branch-collision-audit (pre-flight before STARTING a chain).

Two sister Claude Code plugins for autonomous overnight workflows — client-delivery polishing + ah-ha insight discovery, both with multi-agent review panels

Fire-ASAP orchestrator pattern for multi-track autonomous overnight workflows on scheduled triggers — a self-rescheduling poll loop that consolidates the moment all tracks report complete, and survives a 12-20h session end. Companion safety pattern.

Pre-flight audit before starting a large multi-PR redesign that rewrites shared files. Catches the failure mode where a long-running parallel feature branch (e.g. client-variant, staging, whitelabel) has unmerged commits touching the same files the redesign is about to rewrite — so they end up stranded with head-on conflicts that can't be cleanly cherry-picked. Sister to pre-merge-client-variant-regression-audit (audits a variant branch BEING merged into main; this audits BEFORE main diverges from a variant) and parallel-pr-scope-overlap-tiebreaker-delta-check (two simultaneous PRs vs. one redesign on main + WIP elsewhere). Companion to overnight-multi-issue-implementation for the file-collision dimension that the existing tracker-id audit doesn't cover.

Companion safety pattern: in an unattended overnight review panel, a reviewer that couldn't see the code reads as a CLEAN one — usually because the code-review subagent lacks a Bash tool, so when told to `gh pr diff`/checkout a PR it returns a BLOCKED report (or reviews main, which predates the PR). Pre-generate per-base diffs + materialize worktrees + hand explicit paths, and treat BLOCKED as not-clean. Overnight specialization of the general skill code-reviewer-subagent-no-bash-blocked-on-pr-diff.

Overnight autonomous workflow that takes a cluster of related GitHub issues (typically a P1 review-panel finding set) and ships them to merged stacked PRs by morning. Builds on subagent-driven-development with overnight-specific discipline: stacked PRs (so PR2 doesn't wait on a human PR1-merge mid-night), pre-flight tracker-id audit (concurrent sessions on main steal IDs), final PR-level code review before proposing merge, review findings preserved as PR comments before squash. Sister to overnight-review-client-delivery and overnight-insight-discovery.