plan-review

Plan Review

Attack a plan's load-bearing assumptions while course correction is still cheap; implementation-plan writes plans and diff-review reviews code.

Workflow

1. Reconstruct intent: Read the full plan, its source requirements, and repository reality, never a summary; state the desired outcome, non-goals, constraints, evidence, and proof required for completion.
2. Steelman first: State the strongest case for the intended outcome before criticizing the approach.
3. Inspect current reality: Verify the source paths, interfaces, dependencies, runtime assumptions, and existing mechanisms the plan replaces or duplicates.
4. Map claims: Extract the decisions and assumptions the plan depends on; flag any requirement without a task, task without a requirement, or success claim without proof.
5. Choose lenses: Name which lenses were applied.
   • Engineering: architecture, data flow, interfaces, invariants, failure handling, security, privacy, performance, compatibility, and maintainability.
   • Delivery: dependency order, vertical slices, test seams, migrations, observability, rollout, rollback, operational ownership, and authority boundaries.
   • Founder and product: use the discovery and specification lenses of product-loop.
6. Challenge the premise: Ask whether the problem is real, the scope is the smallest useful wedge, and a no-build or manual alternative could learn more cheaply.
7. Attack failure modes: Imagine the plan failed through missing value, integration breakage, data loss, abuse, operational burden, migration, adoption, or rollback; trace concrete chains, not categories.
8. Test intended versus planned: Compare documented permissions, journeys, data rules, and operational promises with the actual tasks and verification steps.
9. Rank findings: Score each issue by impact, likelihood, confidence, and cheapness to test; promote only issues that could change the decision or execution order, and prefer five decision-changing findings over a long generic list.
10. Offer remedies: Give the smallest corrective change, cheapest decisive test, and kill or rollback criterion; present numbered alternatives when a real trade-off remains.
11. Issue the verdict: Return GO, GO WITH CHANGES, or NO-GO with the minimum conditions for the next state, then report:
    • Load-bearing assumptions, and findings ranked P0–P3 (scale in diff-review) with evidence and impact.
    • Missing proof, the cheapest tests and kill criteria, and recommended scope changes.
    • Residual risks and owner decisions, plus a revised critical path only when revisions were requested.