zustand
Cách dùng Zustand v5 trong repo này (persist middleware, subscribeWithSelector, UI state only).
Découvrez les skills Données et IA, comparez leurs dépôts et consultez leurs instructions.
Cách dùng Zustand v5 trong repo này (persist middleware, subscribeWithSelector, UI state only).
Chuẩn hoá cách đọc/viết DB model Supabase (migrations + TS types + services) trong repo này.
Tóm tắt context repo Huỳnh Sang Blog (monorepo Next.js 16, Supabase DB-first, Cloudinary media, next-intl i18n, Bun/Biome). Dùng khi cần hiểu kiến trúc repo, tìm đúng chỗ để sửa, hoặc onboarding contributor.
Quy ước dùng Zod trong repo (schemas folder, Vietnamese messages, integration với react-hook-form và server actions).
Next.js 16 App Router patterns for this repo (await params/searchParams, next-intl locale handling, Supabase DB-first via services).
Làm việc với media theo pattern repo (Cloudinary lưu file, Supabase lưu metadata/reference). Dùng khi hiển thị ảnh cover, tối ưu ảnh, hoặc thao tác media trong admin.
Quy ước Tailwind CSS trong repo này (Tailwind v4 CSS-first, tokens, cn(), không hardcode màu).
Scans a filesystem target and produces ds.file_cleanup_snapshot_v1 (no deletions). Input: env.file_cleanup_snapshot_request_v1.
LLM-only skill that accepts a repository structure dump, README, and notes, and returns a standardized markdown snapshot, condensed memory summary, and human checklist. No external HTTP requests or web search.
Use for deterministic changes, evidence-first delivery, or assembling proof kits with clear safety justification.
Coordinates repo snapshot, gate runs, and optional WF cycle operations for a single station execution.
Normalize a single DPP passport input into lab_battery_passport_extended_v1 (happy-path).
Generate a basic cleanup plan (keep/delete/archive/ask) from a filesystem snapshot; no filesystem changes. Input: env.file_cleanup_plan_request_v1, output: ds.file_cleanup_plan_v1.
Assemble a canonical wf_cycle winner_pack from compare artifacts (deterministic copy, replay check).
Use for creating new packs/workflows, adding domains/workers, or adding quality/neg suites with station-compatible scripts.
LLM-only skill that takes a textual description of a new skill and generates a complete minimal file plan for a new MOVA skill (schemas, envelope, manifest, prompt profile, runtime binding, docs, cases).
Summarizes branch readiness versus a base branch and produces artifacts for PR prep.
Given a repo snapshot and a goal description, produce a structured, step-by-step plan of code changes: which files to touch, in what order, with risks and done-criteria. This skill does not modify files and doe...
Given a MOVA skill id, target runtime, and a list of envelopes to implement, generate code skeleton(s) and optional binding JSON files that connect these envelopes to real execution (HTTP, CLI, etc.).
Runs Skill Seeker for env.skill_ingest_run_request_v1 and returns ds.skill_ingest_run_result_v1.
Runs npm validate/test/smoke gates and emits a JSON gate report with log paths.
Generates a static MOVA AI bootstrap pack for a target model/environment (no LLM calls). Input: env.mova_ai_bootstrap_generate_v1, output: ds.mova_ai_bootstrap_pack_v1.
Minimal example skill that turns a natural-language procedure description into a structured list of steps. Serves as the reference pattern for other MOVA skills in this lab.
Helps IDE agents navigate and operate the MOVA Skills Lab: read the registry, plan skill runs and manage episodes.
Le classement utilise les étoiles du dépôt GitHub indexé. Plusieurs skills peuvent partager ce nombre : ce n’est ni un nombre d’installations ni une note de qualité.