What problem does it solve? When writing, refactoring, or reviewing code, AI assistants often over-engineer solutions, introduce unnecessary abstractions, silently swallow errors with fallbacks, or rewrite stable code just to look more modern. This Skill constrains those decisions with delivery-first engineering principles so changes stay minimal, stable, and verifiable. ## Core Features & Use Cases - Delivery-first decision framework: Prioritizes a stable, working main path over elegant architecture, and reuses existing structures before building new ones. - Boundary and failure discipline: Enforces clear separation between core logic, UI, data, and integration layers, and requires critical problems (missing config, invalid data, broken contracts) to surface early instead of being hidden by silent fallbacks. - Engineering hygiene for Unity and real-time systems: Covers hot-path performance, lifecycle symmetry, caching, nullability, and concise Chinese comments at key logic points. - Use Case: When asking an AI to refactor a Unity gameplay module or review a pull request, this Skill ensures the output explains what changed, why, what the risks are, and how the change was verified. ## Quick Start Use the delivery-first-code-guardrails skill to review this refactoring plan and point out over-engineering, hidden fallbacks, and missing verification steps.
