plan-rerun-review-state-sync
Keep iterative plan-review reruns truthful by syncing prompt files, canonical review artifacts, and the plan's own review-summary narrative after each adversarial wave.
Best use case
plan-rerun-review-state-sync is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Keep iterative plan-review reruns truthful by syncing prompt files, canonical review artifacts, and the plan's own review-summary narrative after each adversarial wave.
Teams using plan-rerun-review-state-sync should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/plan-rerun-review-state-sync/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How plan-rerun-review-state-sync Compares
| Feature / Agent | plan-rerun-review-state-sync | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Keep iterative plan-review reruns truthful by syncing prompt files, canonical review artifacts, and the plan's own review-summary narrative after each adversarial wave.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Plan Rerun Review State Sync Use when a plan goes through multiple adversarial review waves and you are revising the same plan file repeatedly. ## Problem Repeated plan reruns create three kinds of drift: 1. old prompt files target outdated plan text 2. canonical review artifacts still reflect older verdicts 3. the plan's own `## Adversarial Review Summary` contradicts the latest draft state This creates false MAJOR findings like: - the plan still says FAIL even after a patch - review artifacts cited in the plan no longer match the current draft - a newer draft is judged using an older provider prompt or older saved artifact text ## Required workflow after every material plan revision 1. Regenerate the provider prompt from the latest on-disk plan. 2. Save raw provider outputs to a new rerun-specific file: - `.planning/quick/review-<issue>-rN-<provider>.out` 3. Overwrite the canonical repo artifact for that provider with the newest substantive verdict: - `scripts/review/results/YYYY-MM-DD-plan-<issue>-<provider>.md` 4. Update the plan's own `## Adversarial Review Summary` immediately after the wave. 5. Distinguish clearly between: - historical review results - current draft state 6. If the current draft has been patched after a failing wave but before the next rerun, the plan summary must say `PENDING RE-REVIEW` rather than continuing to self-declare FAIL for the new draft. ## Strong pattern for the summary section Use this structure: ```md ## Adversarial Review Summary Historical review results for the immediately previous draft revision are recorded below for traceability. They are superseded by the latest patch in this draft and should not be read as the verdict for the current unre-reviewed text. | Provider | Verdict | Key findings | |---|---|---| | Codex | MAJOR | ... | | Codex | MINOR | ... | | Gemini | APPROVE | ... | **Current draft state:** PENDING RE-REVIEW ``` After the next full wave lands, replace that with the actual latest-wave verdicts. ## Additional guardrails - Do not leave already-satisfied work items in the plan as if still pending. - If a plan cites `docs/plans/README.md` update work, verify whether the row already exists before keeping that item. - If a contract/rule is meant to be reusable, avoid tying tests to a single dated artifact when the real rule applies to an artifact class. - When a reviewer complains about stale artifacts or stale branch-state assumptions, treat that as a plan-state-sync failure first, not necessarily a design failure. ## When a plan is not approval-ready If any provider still returns MAJOR: - do not post to GitHub - do not add `status:plan-review` - patch the plan again - regenerate prompt again - rerun review again ## Why this matters Without this sync discipline, iterative plan hardening wastes time because each rerun attacks stale state instead of the actual current draft.
Related Skills
plan-review-prompt-refresh-after-plan-edits
Refresh reviewer prompt files from the latest on-disk plan before every adversarial re-review. Prevents Codex/Gemini from critiquing stale plan text after local edits.
worktree-branch-sync-hygiene
Class-level branch, worktree, dirty-main, stash, sync, and hook hygiene for workspace-hub style multi-repo work.
workspace-hub-overnight-plan-monitor
Monitor and reconcile workspace-hub overnight planning or implementation batches, including process status, result artifacts, issue/commit verification, and controlled failed-lane recovery.
repo-sync
Smart repository synchronization across workspace-hub ecosystem — diagnoses and fixes pull failures (detached HEAD, diverged branches, uncommitted changes)
plan-gated-issue-execution-wave
Execute a multi-issue architecture/planning wave in a plan-gated repo, then safely transition approved issues into implementation with file-based Codex prompts, local approval markers, subprocess monitoring, and cleanup handling for sandbox/hook edge cases.
mixed-ops-vs-repo-fix-plan-boundary
Plan mixed operational-vs-repo remediation issues by proving live-state classification first, then only proposing code changes for confirmed repo-owned failure paths.
wave-based-parallel-plan-execution
Orchestrate phase execution by discovering dependencies, grouping into waves, spawning subagents, and collecting results with optional wave filtering
tdd-verification-and-adversarial-review
Verify pre-written TDD tests pass, conduct adversarial code review on committed diffs, and route findings to existing issues
recover-from-corrupted-git-state
Diagnose and recover from corrupted git states (stale locks, failed rebases, pre-commit hook blocks) during bulk operations
plan-gated-issue-validation-workflow
Systematic validation pattern for plan-approved GitHub issues with pre-existing deliverables
plan-gated-issue-implementation
Workflow for executing pre-approved GitHub issues with mandatory validation checkpoints
multi-role-agent-contract-review-pipeline
Execute a 4-role agent team (Planner/Architect/Reviewer/Integrator) pipeline for self-reviewing knowledge artifacts before delivery