LLM & Agents

여러 AI 에이전트(Claude, Codex)가 동일한 주제에 대해 병렬로 기획을 수행하고, 각 결과를 보여준 후 최종 머지된 기획안을 제시합니다. "3명이 기획해주세요"처럼 에이전트 수를 지정할 수 있으며, Claude와 Codex가 랜덤하게 분배됩니다.

Expert in project scaffolding and lifecycle management for Microsoft 365 Copilot agents using the Agents Toolkit (ATK) CLI. **ALWAYS USE FIRST** when starting new agent projects. Use when (1) Creating/scaffolding new agent projects with 'atk new', (2) Setting up project structure and initial files, (3) Provisioning Azure resources with 'atk provision', (4) Deploying agents with 'atk deploy', (5) Packaging with 'atk package', (6) Publishing with 'atk publish', (7) Sharing agents, (8) Managing environments (.env files), (9) Troubleshooting ATK CLI or deployment issues. This skill handles PROJECT SETUP and TOOLING, not TypeSpec code implementation.

Multi-LLM collaborative brainstorming and planning. Use when user explicitly requests consultation with multiple AI models (ChatGPT, Gemini, other LLMs) before presenting an implementation plan, or asks to "consult the council", "ask other models", or "get perspectives from other AIs". Queries external LLM APIs, synthesizes their perspectives, and presents an adapted implementation plan.

Creates jaw-dropping, production-ready frontend interfaces AND delivers perfectly matched real photos (Unsplash/Pexels direct links) OR flawless custom image-generation prompts for hero images, backgrounds, and illustrations. Zero AI slop, zero fake URLs.

Inlang Paraglide を使用した国際化（i18n）対応を支援します。翻訳メッセージの追加、reactiveMessage の使用、言語切り替え実装、翻訳のテスト、パラメータ付きメッセージなどの国際化関連タスクに使用します。

Dynamic swarm topology reconfiguration, network latency optimization, agent placement, and communication pattern optimization. Use for optimizing swarm structure, reducing communication overhead, and adaptive network design.

Dual-AI code validation using both Claude and Google Gemini to catch 20% more issues. Validates React Native screens for security vulnerabilities, performance anti-patterns, accessibility gaps, and best practices. Use when user says "validate with Gemini", "cross-validate", "AI review", or before deploying code to production.

ENFORCEMENT tool that detects when Skills automation is duplicated in agent definitions, lessons learned, or process docs. Prevents "single source of truth nightmare" by finding bash commands, step-by-step procedures, or process descriptions that replicate Skills. BLOCKING AUTHORITY - workflow cannot complete with violations.

This skill should be used when developing backend services for Manjha using Encore.ts, implementing agents, writing E2E tests, or working with the message-driven architecture. It ensures adherence to TDD practices, type safety, proper logging, and Manjha's development principles.

Access and interact with Large Language Models from the command line using Simon Willison's llm CLI tool. Supports OpenAI, Anthropic, Gemini, Llama, and dozens of other models via plugins. Features include chat sessions, embeddings, structured data extraction with schemas, prompt templates, conversation logging, and tool use. This skill is triggered when the user says things like "run a prompt with llm", "use the llm command", "call an LLM from the command line", "set up llm API keys", "install llm plugins", "create embeddings", or "extract structured data from text".

First-time setup for the claude-foundry plugin (plugin:foundry@claude-foundry)

Customer onboarding specialist

Use when Clawdis needs to test or operate GoHome (Home Assistant clone) via gRPC discovery, metrics, and Grafana.

Generate new Claude Code skills following official best practices. Use when user requests "create a skill" or "make a skill" for automating workflows or tasks.

【WHEN】当用户要新增/修改/评审“治理层文档”（README.md、CLAUDE.md、以及用户显式指定的其他治理层文档）时启用； 【WHEN】当输出提案 vN.md、落地修改治理层文档、或产出 vN-review.md 时启用。 【WHEN NOT】不用于业务功能开发、不用于代码实现细节、不用于非治理层文档（如普通开发笔记/模块设计/接口文档）。

Expert on Shelby Protocol smart contracts on Aptos blockchain. Helps with blob metadata, micropayment channels, auditing system, storage commitments, placement group assignments, Move modules, and on-chain state management. Triggers on keywords Shelby smart contract, Shelby Move, blob metadata, micropayment channel, Shelby auditing, placement group assignment, storage commitment, Aptos contract.

Deconstructs complex problems using the 'Polymath Investor' framework with a 5-step cognitive chain (First Principles → Modular Isolation → Pareto Filtering → Cross-Domain Mapping → Dynamic Zooming). Use when analyzing systems, architectures, businesses, or learning challenges; when seeking high-leverage insights or 'Alpha'; when applying first principles thinking; or when needing cross-disciplinary perspectives to reveal blind spots. Transforms overwhelming complexity into actionable levers.

Manage staff-institution mapping table for vehicle insurance platform. Use when updating mapping files, resolving name conflicts, converting Excel to JSON, or checking mapping coverage. Mentions "update mapping", "staff conflicts", "mapping table", or "institution assignment".

Validate markdown documentation against meta-agentic best practices. Use when validating skills, commands, checking broken links, validating frontmatter, auditing documentation structure, or running compliance checks.

Set up Prometheus for comprehensive metric collection, storage, and monitoring of infrastructure and applications. Use when implementing metrics collection, setting up monitoring infrastructure, or configuring alerting systems.