菜单
Performs code reviews against quality criteria: readability, maintainability, correctness, security, performance, test coverage, naming conventions, and documentation. Use when reviewing pull requests, auditing code quality, or establishing review standards.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Guides on AI/ML system design, LLM application architecture, MLOps pipelines, AI SaaS platform selection (OpenAI, Anthropic, Google Vertex AI, AWS Bedrock, Azure OpenAI, Hugging Face, Cohere, Replicate, Together AI), model serving, feature stores, experiment tracking, evaluation frameworks, RAG architecture, AI agent patterns, responsible AI, cost optimization, and emerging AI/ML trends. Use when designing AI systems, selecting AI providers, building LLM applications, setting up MLOps, or evaluating AI architecture patterns.
AI/LLM integration patterns: Anthropic SDK streaming with TypeScript, RAG architecture (chunking, embedding, vector search with pgvector), tool use / function calling, model selection guide (Haiku 4.5 vs Sonnet 5 vs Opus 4.8), prompt caching, structured output with Zod, token cost management. Use when building AI features, RAG pipelines, AI agents.
Flutter 3.27+ patterns — Riverpod 2 state with codegen, GoRouter 14+ navigation, Impeller rendering (default on all platforms), WASM web, extension types, on-device AI (firebase_ai, MediaPipe), Patrol 3.x testing, platform channels, performance
Expert guidance on Flutter 3.27+ (Impeller default on all platforms, Material 3, WASM web, DevTools, Flame), Dart 3.3+ (records, patterns, sealed classes, extension types, macros), state management (Riverpod 2 with codegen, BLoC/Cubit 9.x, Signals), navigation (GoRouter 14+, auto_route), networking (Dio, retrofit, Ferry/Artemis GraphQL), local storage (Drift, ObjectBox, Isar, Hive CE, Realm), testing (widget_test, golden tests, Patrol 3.x, alchemist, mocktail, bloc_test), on-device AI (firebase_ai, google_generative_ai, MediaPipe Tasks), Flutter for Web (WASM default)/Desktop/Embedded, Mason, very_good_cli, flutter_gen, custom_lint, Flutter Hooks, freezed. Use when building or optimizing Flutter apps.
Designs CI/CD pipelines with build, test, and deploy stages, blue-green/canary/rolling deployments, feature flags, artifact management, and environment promotion strategies. Use when setting up or improving CI/CD pipelines or deployment workflows.
Designs REST, GraphQL, gRPC, tRPC, WebSocket, SSE, and AsyncAPI/event-driven APIs. Covers OpenAPI 3.1, API gateways, GraphQL Federation, BFF pattern, contract testing, API mocking, versioning, rate limiting, webhooks, and HATEOAS maturity model. Use when designing new APIs, reviewing API architecture, or choosing between API styles.
| name | alpha-core:code-review |
| description | Performs code reviews against quality criteria: readability, maintainability, correctness, security, performance, test coverage, naming conventions, and documentation. Use when reviewing pull requests, auditing code quality, or establishing review standards. |
| allowed-tools | Read, Grep, Glob, Bash |
You are a code review specialist. Provide constructive, specific, actionable feedback. Every comment must explain why the change matters -- not just what to change.
[blocker], [suggestion], [nit], [question], [praise]references/review-process.mdreferences/patterns-smells.mdreferences/checklist-lang.mdreferences/checklist-tooling.md