What problem does it solve? Plans often reach implementation with unchallenged premises, missing test coverage, incomplete UI states, or poor developer experience, and fixing those gaps in code review costs far more than catching them in the plan. This Skill reviews an already-written plan from up to four expert lenses and surfaces findings as decisions before the plan becomes tickets. ## Core Features & Use Cases - Four review lenses: CEO (premise, ambition, mandatory alternatives, scope modes), Eng (architecture, failure modes, deep test-coverage trace, performance), Design (information architecture, interaction states, AI-slop risk, accessibility), and DevEx (time-to-first-value, error messages, docs, upgrade path). - Orchestrated all mode: Runs applicable lenses in a fixed CEO → Design → Eng → DX order, auto-decides only mechanical findings, escalates taste calls, and synthesizes cross-lens themes into one final gate. - Findings become tasks: Each lens appends a P1/P2/P3 Implementation Tasks list to the plan and returns a CLEARED or NOT CLEARED verdict without rewriting the plan itself. - Use Case: Before converting a spec in docs/plans/ into tickets, run the Eng lens to trace every branch and error path, then run all to get a deduplicated task list and a ship/no-ship verdict. ## Quick Start Ask the agent to review the plan in docs/plans with the void-plan-review skill using all lenses and append the findings as an implementation task list.
