Total 58,172 skills
Showing 12 of 58172 skills
Điều phối toàn bộ quy trình làm app Flutter bằng AI - từ đọc OpenSpec, quyết định có cần migration không, sinh code, chạy MCP auto-fix, xin duyệt UI, đến commit git. LUÔN dùng skill này đầu tiên khi user yêu cầu thêm/sửa tính năng cho app Flutter, kể cả khi họ chỉ nói ngắn gọn như "thêm màn hình đăng nhập" hay "làm app bán trà sữa". Đây là checklist điều phối, không tự viết code.
Thiết kế schema Postgres, viết Row Level Security (RLS) policy, và Edge Function (Deno/TypeScript) cho Supabase khi tính năng cần logic phía server (không thể/không nên để client tự làm - ví dụ tính giá, webhook thanh toán, phân quyền dữ liệu). Dùng skill này song song với flutter-coding khi 1 task trong tasks.md cần cả phần client (Flutter) lẫn phần server (Supabase).
Diagnose why AdMob ads are not showing in any Expo/React Native app. Use when the user reports real/test ads not displaying, asks why production ads have NO_FILL, whether the AdMob config/package/unit IDs are correct, how a TEST_ADS flag and UMP consent gate affect ads, or how to debug ads on a real device via adb logcat. Not project-specific — works for any app using react-native-google-mobile-ads.
Use when testing or debugging ESP32/Arduino firmware in Wokwi via the Wokwi MCP server — starting simulation, reading serial or pin state, verifying boot, or when the sim won't start, the MCP tool errors, or the host can't reach the simulated device's web server.
Định nghĩa kiến trúc Flutter cố định (state management, router, cấu trúc thư mục Feature-First) mà mọi code sinh ra phải tuân theo. Dùng skill này mỗi khi sinh code mới, tạo feature mới, hoặc khi cần quyết định đặt file ở đâu trong project Flutter.
Creates and maintains concise AGENTS.md and CLAUDE.md project instruction files. Use when asked to create AGENTS.md, update AGENTS.md, maintain agent docs, set up CLAUDE.md, document repository agent conventions, or keep coding-agent instructions minimal and reference-backed.
Perform code reviews following Sentry engineering practices. Use when reviewing pull requests, examining code changes, or providing feedback on code quality. Covers security, performance, testing, and design review.
Finish a development branch. Use when implementation is complete and the work needs to be merged, pushed as a PR, kept, or discarded.
Use when writing or reviewing anything that will be read as true — a decision record, a README, an acceptance report, a runbook, a changelog entry, an audit finding, or any claim that something was verified. Applies the ten canons of evidence-backed documentation — what makes a claim documentation rather than an assertion — and routes to the doctrine that enforces each one. Also use when a project needs a documentation gate, a decision register, a propagation matrix, or a retrospective that outlives its author. Triggers - 'documentation gate', 'decision record', 'ADR', 'acceptance report', 'runbook', 'is this verified', 'доказательная документация', 'записать решение', 'отчёт о приёмке', 'раннбук', 'чем это подтверждено', 'доки в синхроне'. Not for: drafts, chat answers, commit messages or code comments — say 'без доков' to opt out.
Improves Java code readability and maintainability through naming, small functions, DRY, KISS, YAGNI, and focused refactoring. Use when cleaning Java code, refactoring Java classes, reviewing maintainability, reducing complexity, or improving Java names and method structure.
Generate or edit images using OpenAI GPT Image API (gpt-image-2, gpt-image-1, etc). Use ONLY when the user explicitly names OpenAI or GPT as the provider: "gpt image", "openai image", "generate image with openai", "用 openai 画图", "用 GPT 生成图片". For generic image requests without a provider, use nanobanana-skill instead. Do NOT use for diagrams (架构图/流程图) — draw those with Mermaid or code.
Collect posts from Reddit — text, images, likes, comments, timestamps. Use when the user wants to research content, track publishing activity, or export post datasets.