review-plan / default

What problem does it solve?

It helps prevent unsafe or incomplete AI execution by rigorously reviewing an action plan for completeness, testability, reversibility, and faithful handling of provided context references.

Core Features & Use Cases

• Universal plan validation: Enforces required plan sections (Goal, Approach, Tools, Test plan, Success criteria, Rollback, Root-cause scope) plus scope and testability/reversibility requirements.
• Context reference integrity checks: Confirms that captured context_refs are honored across the plan (or explicitly dismissed with reasoning).
• Playbook-other guardrails: For playbook: other, requires an explicit artifact declaration and makes silent omission of context_refs an automatic rejection.

Quick Start

Have the AI review the latest stage plan for an action and either approve it if all required checks pass or return a specific rejection list if anything is missing or inconsistent.