LLM & Agents

Multi-agent collaborative decision-making for complex problems. USE WHEN you need multiple specialist perspectives on high-stakes decisions (architecture, prioritization, security, compliance). Orchestrates conversation between custom agent rosters, synthesizes perspectives into actionable decisions. Finds 2-3x more issues than solo agent mode.

A specialized skill for frontend development tasks using the Gemini CLI. It leverages Gemini's multimodal capabilities for generating UI code (React, Tailwind CSS) from descriptions or images (screenshots).

Lightweight search utility for Toss / Apps-in-Toss llms-full.txt docs, caching sources and ranking snippets with keyword + similarity (Korean-aware) scoring.

NONSTOP v2.1 - Unstoppable task executor with ULTRATHINK planning, parallel agents, and perfectionist verification. Full lib integration. Keywords: nonstop, sprint, unstoppable, perfect, parallel, ultrathink

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations. | Sử dụng khi: tạo skill mới, custom skill, skill template.

Validate changes against .agentignore before commit.

Prompt engineering standards and context engineering principles for AI agents based on Anthropic best practices. Covers clarity, structure, progressive discovery, and optimization for signal-to-noise ratio.

This skill should be used when Claude needs to ask the user a question, gather user preferences, clarify requirements, make decisions, or get confirmations. It provides guidance on using the AskUserQuestion tool with well-crafted prefilled options to reduce cognitive load and help users who may feel overwhelmed or unsure how to articulate their thoughts.

Autonomous agent creation skill that generates specialized agent definitions from templates. Use when you need to create new Claude Code agents for specific tasks like code review, deployment automation, testing, documentation, security analysis, or general-purpose research. This skill automates the creation of agent definition files (.md) with proper structure, workflow instructions, and tool access patterns following Miyabi framework standards.

仔细彻底阅读探索用户提供的代码目录/文件，分析拆解出对应的代码片段以供参考，颗粒度细到功能级别，只做片段提取不改变代码内容和页面布局结构，提取到./fragment-extractor目录下，按照指定模板格式生成代码片段模板文件，确保功能完整性和代码片段的实际性。

Bi-directional comparison workflow for config auditing. Compares what agents expect (from templates/standards) against what repositories actually have, presenting differences with remediation options. Use when agents need to audit configs, validate standards compliance, or identify mismatches between expected and actual configurations.

Executes work in a project repo using a startup prompt. Use when user pastes a startup prompt or asks to begin executing a session. This skill is meant for use in project repos, not in vibehub.

Canonical Claude Code skill authoring kit covering agent Skills, sub-agent templates, custom slash commands, orchestration patterns, hooks, memory, settings, and IAM permission rules aligned with official documentation.

Create and manage Claude Code skills following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns, file paths, content patterns), enforcement levels (block, suggest, warn), hook mechanisms (UserPromptSubmit, PreToolUse), session tracking, and the 500-line rule.

Write strategic, behavior-focused tests for core user flows and critical functionality using Vitest for unit tests and Playwright for end-to-end testing. Use this skill when writing test files, testing utilities, or implementing test coverage for critical paths. When working on Vitest test files (.test.ts, .spec.ts), Playwright E2E test suites, test setup and configuration files (vitest.config.ts, test-setup.ts), mock implementations for external dependencies, testing utility functions or test helpers, component test files using React Testing Library, API route test files, or integration tests for Content Collections and data fetching logic.

Manages the lifecycle, configuration, and synchronization of specialized sub-agents (Planner, Implementer, Reviewer, etc.) across various platforms (Claude, Copilot, OpenCode). Use this skill when you need to: (1) Sync sub-agent instructions from master templates, (2) Add new sub-agent definitions, (3) Update agent capabilities or permissions, or (4) Validate that agent instructions are in sync across platforms.

Expert D&D 5th Edition (2014 PHB) character sheet manager and game master assistant. Use when creating characters, leveling up, calculating ability modifiers, managing spells, tracking hit points, updating equipment, applying class features, or answering rules questions about D&D 5e.

Design learning experiences for AI-native software development, integrating the Three Roles Framework (AI as Teacher/Student/Co-Worker), co-learning partnership pedagogy, and "Specs Are the New Syntax" paradigm into programming curriculum. Use this skill when educators need to prepare students for professional AI-driven development workflows, teach effective specification-first collaboration, or balance AI assistance with foundational learning goals. This skill helps create lessons that leverageAI tools appropriately while ensuring students build independent capability, bidirectional learningpatterns, and ethical AI use practices. Aligned with Constitution v4.0.1.

Orchestrates comprehensive audits across multiple specialized auditors for Claude Code customizations and provides guidance on naming, organization, best practices, and troubleshooting. Use when: (1) Auditing - wants complete evaluation, multi-faceted analysis, coordinated audit reports, thorough validation, or asks to audit multiple components; (2) Guidance - asks "what should I name...", "how should I organize...", "best practices for...", troubleshooting issues, understanding evaluation criteria, or needs pre-deployment validation. Automatically determines which auditors to invoke (agent-audit, skill-audit, hook-audit, command-audit, output-style-audit, evaluator, test-runner) based on target type and compiles unified reports with consolidated recommendations.

Master the Arcanean Prompt Language - advanced prompt engineering using mythological frameworks, constraint architecture, and the Centaur Principle for human-AI co-creation