What problem does it solve? WeChat Mini Program development is full of silent failure modes: setData updates that stutter because of Shadow-tree size rather than payload size, native components whose event handlers never fire, misplaced lazyCodeLoading config that is silently ignored, and CI uploads that fail on robot numbers or key formats. This Skill encodes the platform's actual rules, version gates, and cost model so code is written and reviewed correctly the first time. ## Core Features & Use Cases - Performance diagnosis: Applies the platform-documented update cost model (Shadow-tree node count first, payload size second) to fix stuttering lists, scroll jank, and background-page updates, with measurement via setUpdatePerformanceListener. - Configuration and code review: Catches illegal subpackage nesting, tabBar placement, Skyline/glass-easel prerequisites, native-component event binding and CSS limits, and privacy/authorization flow errors, each tied to the base library version it holds from. - Release pipeline: Builds a miniprogram-ci upload workflow with correct robot numbering, key handling, packNpm, and subpackage size gates, plus miniprogram-simulate unit testing. - Use Case: A feed page stutters past 200 rows and the team plans to diff setData payloads; the Skill rejects that plan, identifies the single 7000-node tree as the real cost, and prescribes splitting rows into self-updating components. ## Quick Start Review my WeChat Mini Program's app.json and feed page for configuration errors and performance problems, listing each finding with its trigger, impact, and fix.
