Restart Claude Code session to reload new skills, manage context, or start fresh. Use this skill when skills have been added/modified, when needing to compact/clear context, or to reload configuration changes.
Production deployment workflow for agentic systems. Directs to RAG for implementation.
Create new Claude Code skills following SOTA compact design patterns (v3.2)
Initialize long-running agent projects with environment setup, feature list generation, and progress tracking. Use when starting a new implementation project or setting up an existing codebase for Claude-assisted development. Creates init.sh (dev server script), claude-progress.txt (work log), and docs/features.json (feature list with status tracking).
Assist with image analysis, object detection, and visual AI tasks
Return to normal stop behavior. Use PROACTIVELY after completing all tasks in continuous work mode, when work is committed and tests pass, or when reaching natural stopping point. User can invoke with /claude-allow-stop.
Comprehensive performance analysis, bottleneck detection, and optimization recommendations for Claude Flow swarms
创建有效技能的指南。当用户想要创建新技能（或更新现有技能）以通过专业知识、工作流程或工具集成扩展 Claude 功能时，应使用此技能。
ANIMA as limit construction over condensed skill applications. Formalizes prediction markets as belief ANIMAs, structure dishes as condensation media, and impact as equivalence class change. Use for understanding agency at maximum entropy, compositional world modeling, or applying Scholze-Clausen condensed mathematics to AI.
Design, implement, and debug a custom ChatKit backend in Python that powers the ChatKit UI without Agent Builder, using the OpenAI Agents SDK (and optionally Gemini via an OpenAI-compatible endpoint). Use this Skill whenever the user wants to run ChatKit on their own backend, connect it to agents, or integrate ChatKit with a Python web framework (FastAPI, Django, etc.).
Local LangChain AI documentation reference. Use when asked about LangChain, LangGraph, agents, chains, prompts, memory, tools, retrieval, RAG, vector stores, document loaders, or building LLM applications.
Branch the current Claude Code session into a new tmux session using --resume. Creates a parallel conversation that can diverge independently. Use when you want to explore an alternative approach without losing your current context.
Build AI-first applications with RAG pipelines, embeddings, vector databases, agentic workflows, and LLM integration. Master prompt engineering, function calling, streaming responses, and cost optimization for 2025+ AI development. Includes local LLM inference with Ollama for 93% CI cost reduction.
Determines appropriate agent routing based on task type, scale, and agent capabilities
Prompt engineering guidance for Gemini (Google) model. Use when crafting prompts for Gemini to leverage system instructions, multimodal capabilities, ultra-long context, and strong reasoning features.
Stream-JSON chaining for multi-agent pipelines, data transformation, and sequential workflows
Use when beginning a new conversation to work on an open-ended goal, loading context from previous iterations through iteration journals
Universal knowledge storage and retrieval patterns using memory graph
Implement persistent memory patterns for AI agents using AgentDB. Includes session memory, long-term storage, pattern learning, and context management. Use when building stateful agents, chat systems, or intelligent assistants.
Senior MLOps Engineer with 8+ years ML systems experience. Use when integrating LLM APIs (Gemini, OpenAI, Groq), building AI pipelines, managing prompts, setting up model serving, implementing AI cost optimization, or building training data pipelines.
Project-agnostic guidance for continuing work across Claude sessions. Ensures context recovery, task resumption, and state reconciliation.
Multi-repository coordination, synchronization, and architecture management with AI swarm orchestration
