LLM & Agents

Global frontend aesthetics skill that helps Claude avoid generic "AI slop" UI and make bold, intentional visual decisions while still honoring each project's design-dna, tokens, and architecture.

Generates Nano Banana Pro prompts for profile icons and SNS images. Use when user mentions "プロフィール画像", "アイコン作成", "SNS用画像", or "ヘッダー画像".

Claude Codeフックの詳細仕様、設計原則、設定方法。フック、hook、PreToolUse、Stop時に使用。

멀티 AI 코드 리뷰 오케스트레이터. Claude, Codex, Gemini, Droid를 조율하여 종합적 코드 리뷰 수행. "멀티 AI 리뷰", "코드 리뷰", "종합 리뷰" 요청 시 활성화됩니다.

Claude Harness を使用して Plans.md のタスクを実行。Use when user mentions 実装実行, execute tasks, Harness, 並列実装. Do not use for design or review.

Generate learning science-backed feedback templates for Canvas assignments. Use when educators need feedback templates for grading, want to create rubric comments, need encouraging feedback language aligned with Pillar 3 principles (targeted, encouraging, immediate), or want to set up SpeedGrader comment libraries. Works with canvas-mcp for bulk grading.

Use when building search systems that need both semantic similarity and keyword matching - covers combining vector and BM25 search with Reciprocal Rank Fusion, alpha tuning for search weight control, and optimizing retrieval quality

博客写作辅助。根据主题或灵感生成文章提纲，组织写作思路。当用户想写文章、需要提纲、引用灵感库内容时使用。适用于 output/ 目录。

Writing assistant trained in Ken Klippenstein's investigative journalism style - anti-establishment, transparent, evidence-based, no MSM brain

Improves the RAG (Retrieval-Augmented Generation) chatbot for the Physical AI & Humanoid Robotics textbook with strict grounding, citation requirements, and performance optimization.

PySide6 widget implementation for LoRAIro GUI with Signal/Slot pattern, Direct Widget Communication, and Qt Designer integration best practices

设计和构建任意领域的 AI Agent。适用于以下场景：(1) 用户要求"创建一个 agent"、"构建一个助手"或"设计一个 AI 系统"(2) 想要了解 agent 架构、agentic 模式或自主 AI(3) 需要帮助处理能力、子 agent、规划或 skill 机制(4) 询问 Claude Code、Cursor 或类似 agent 的内部原理(5) 想要为业务、研究、创意或运营任务构建 agent关键词：agent、助手、自主、工作流、工具使用、多步骤、编排

Agentic execution protocols for Claude Code with verification gates, rollback mechanisms, and confidence thresholds. Triggers on any destructive or high-risk operations.

Enforces test-driven development workflow for Claude Code. Implements Red-Green-Refactor cycle with configurable gates. Use when writing code, implementing features, or fixing bugs.

Creating GLSL shaders with interactive parameter exploration. Use this when users request creating shaders, fragment shaders, visual effects, procedural textures, or shader art. Create original shader code that runs in real-time with adjustable parameters.

Audit and optimize website for AI search engines like ChatGPT, Perplexity, Google AI Overviews, and Claude. Use when discussing GEO (Generative Engine Optimization), SEO for AI, llms.txt, AI crawlers, structured data for LLMs, or visibility in AI search results.

Generate and optimize PRDs, Implementation Plans, and Progress Tracking documents optimized as AI artifacts for development agents. Use when creating new feature plans, breaking down long planning docs (>800 lines), or setting up progress tracking. Supports: 1) Create PRD from feature request, 2) Create Implementation Plan from PRD with phase breakdown and subagent assignments, 3) Optimize existing plans by breaking into phase-specific files, 4) Create progress tracking with task assignments. Example: "Create a PRD for user authentication feature" or "Break down the sidebar-polish implementation plan into phase files" or "Create progress tracking for data-layer-fixes PRD".

Reviews pull requests created by GitHub Copilot agents before merging. Triggers on "review this PR", "check PR

Expert assistant for monitoring and optimizing performance in the KR92 Bible Voice project. Use when analyzing query performance, optimizing database indexes, reviewing React Query caching, monitoring AI call costs, or identifying N+1 queries.

Validate skills against agentskills.io specification. Use when adding new skills to the marketplace, reviewing skill PRs, checking skill compliance, or running quality gates on skills. Validates frontmatter fields (name, description, license, compatibility, metadata, allowed-tools), directory naming, line limits, and structure.