qdrant-retrieval-tool
Generate JSON schemas and Python handlers for Qdrant-based retrieval tools used by the RAG agent to search sitemap-crawled textbook content with URL citations.
Generate JSON schemas and Python handlers for Qdrant-based retrieval tools used by the RAG agent to search sitemap-crawled textbook content with URL citations.
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.
Edit existing images with text prompts using fal.ai Gemini 3 Pro. Use when the user wants to modify, edit, transform, or change an existing image based on a text description. Supports multiple input images for context.
MoAI-ADK's foundational principles - TRUST 5, SPEC-First TDD, delegation patterns, token optimization, progressive disclosure, modular architecture, agent catalog, command reference, and execution rules for building AI-powered development workflows
Interview-based system for creating specialized Claude Code agents. Use when users want to create a Claude Code agent, need help defining agent roles and responsibilities, mention "create an agent" or "build an agent", or want to generate a CLAUDE.md file for agentic coding workflows. Guides through interactive interview to gather requirements and generates complete agent configurations following Claude Code best practices.
Systematic decision tree and epic generation through Socratic discovery
Creates new AI agent skills following the Agent Skills spec. Trigger: When user asks to create a new skill, add agent instructions, or document patterns for AI.
Add speech-to-text transcription widget to web pages using Whisper server. Creates WhisperWidget JS component that records audio, sends to whisper server, and returns transcribed text. Use when adding voice input to web applications.
ユーザーメモリ、プロジェクトメモリ、READMEを再読み込み。「コンテキストを再読み込み」「設定をリロード」「CLAUDE.mdを読み直して」と言われた時、またはclear後にコンテキストを復元したい時に使用
Provides real-time weather forecasts using OpenWeatherMap API. Use when users ask about current or future weather conditions, temperature, precipitation, or weather-dependent planning for any location (e.g., 'What's the weather in Paris tomorrow?', 'Will it rain in Seattle this weekend?', 'Should I bring a jacket to Denver?').
Natural language understanding, intent classification, context management, reference resolution, and conversation history analysis for agentful
Use when implementing different agent types in Microsoft Agent Framework. Triggers: "ChatAgent", "BaseAgent", "WorkflowAgent", "A2A agent", "custom agent class". NOT for: Non-Microsoft agent frameworks or simple single-agent scenarios.
This skill should be used when the user asks to "train a model", "fine-tune", "build NLP model", "create training task", "optimize model performance", "improve accuracy", "what model should I use", or expresses vague training needs like "I want to do sentiment analysis" or "help me with NER". Provides coaching-style guidance to clarify goals, diagnose pain points, and recommend optimal training approaches.
Phase 2 of Ontology Builder Pipeline. AI acts as domain SME to analyze raw inputs, extract entities/workflows/rules, fill knowledge gaps using market expertise. Use after Phase 1 ingestion is complete.
Build and run AI agents using Claude Code CLI. Use when developing autonomous agents, multi-agent systems, CI/CD automation, or scripting Claude for programmatic tasks. Covers authentication, headless mode (-p), JSON output parsing, tool restrictions, subagents, and orchestration patterns.
Build production-grade applications with OpenAI APIs (direct or via Azure OpenAI). Covers structured outputs, function calling, streaming, Assistants API, Responses API, Agents SDK, cost optimization, rate limiting, error handling, and batch processing. IMPORTANT: Before implementing, gather context about the target environment including: - API provider (OpenAI direct vs Azure OpenAI) - Primary use cases (chat, function calling, structured outputs, assistants, agents) - Cost constraints and optimization requirements - Error handling and retry strategy needs
Use when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions
First step for any task. Uses HelixDB to map code relationships, then loads precise context into Repo Prompt. Trigger when starting work, feeling lost, or context seems stale.