LLM & Agents

Use when creating or updating agent evaluation suites. Defines eval structure, rubrics, and validation patterns.

Kailash Kaizen - production-ready AI agent framework with signature-based programming, multi-agent coordination, and enterprise features. Use when asking about 'AI agents', 'agent framework', 'BaseAgent', 'multi-agent systems', 'agent coordination', 'signatures', 'agent signatures', 'RAG agents', 'vision agents', 'audio agents', 'multimodal agents', 'agent prompts', 'prompt optimization', 'chain of thought', 'ReAct pattern', 'Planning agent', 'PEV agent', 'Tree-of-Thoughts', 'pipeline patterns', 'supervisor-worker', 'router pattern', 'ensemble pattern', 'blackboard pattern', 'parallel execution', 'agent-to-agent communication', 'A2A protocol', 'streaming agents', 'agent testing', 'agent memory', 'agentic workflows', 'AgentRegistry', 'OrchestrationRuntime', 'distributed agents', 'agent registry', '100+ agents', 'capability discovery', 'fault tolerance', or 'health monitoring'.

CRITICAL development guide for Microsoft Agent Framework (MAF) in IPA Platform.MUST READ when working with:- Agents (ChatAgent, AgentThread)- Builders (ConcurrentBuilder, GroupChatBuilder, HandoffBuilder, MagenticBuilder)- Workflows (Executor, Edge, WorkflowExecutor)- Multi-agent orchestration- Any code in backend/src/integrations/agent_framework/builders/PREVENTS: Custom implementations instead of wrapping official API.REQUIRES: Import official classes, use Adapter pattern with self._builder = OfficialBuilder()

Node chain testing templates for quality agent. WorkflowBuilder patterns, tier-based testing (simple/complex/full), and assertion patterns.

Provides prompt engineering techniques. Use when working with AGENTS.md, SKILL.md, commands, agents/subagents, or any kind of prompt.

Skill manager for Claude Code. Use when users want to list, search, install, update, remove, or create skills. Activated by requests like "install a skill", "what skills do I have", "find skills for X", "create a new skill", or "update my skills".

Handles MMKV storage operations and data persistence patterns with encryption. Use when implementing data persistence, caching, or user preferences in Fitness Tracker App.

Guide for building applications on Cloudflare's edge platform. Use when implementing serverless functions (Workers), edge databases (D1), storage (R2, KV), real-time apps (Durable Objects), AI features (Workers AI, AI Gateway), static sites (Pages), or any edge computing solutions.

This skill provides comprehensive knowledge for working with OpenAI's Responses API, the unified stateful API for building agentic applications. It should be used when building AI agents that preserve reasoning across turns, integrating MCP servers for external tools, using built-in tools (Code Interpreter, File Search, Web Search, Image Generation), managing stateful conversations, implementing background processing, or migrating from Chat Completions API.Use when building agentic workflows, conversational AI with memory, tools-based applications, RAG systems, data analysis agents, or any application requiring OpenAI's reasoning models with persistent state. Covers both Node.js SDK and Cloudflare Workers implementations.Keywords: responses api, openai responses, stateful openai, openai mcp, code interpreter openai, file search openai, web search openai, image generation openai, reasoning preservation, agentic workflows, conversation state, background mode, chat completions migration, gpt-5, polymorphic o

Conducts comprehensive project specification interviews to clarify architecture, tech stack, design decisions, and implementation details.TRIGGERS - Keywords: interview, specification, spec, requirements, architecture, design decisions, project setup, new project, empty project, unclear requirements, /interview, SPEC.md, project planning, tech decisions, stack decisions, clarify requirements, gather requirements, requirements gathering.TRIGGERS - Phrases: "what should we build", "help me plan", "new project", "start a project", "define the architecture", "unclear what to build", "need to specify", "create a spec", "interview me", "ask me questions", "help me decide", "what technology should".TRIGGERS - Conditions: tech stack confidence LOW, project appears empty, no existing SPEC.md, architecture decisions needed.

Rapidly creates atomic, focused skills optimized with evidence-based prompting, specialist agents, and systematic testing. Each micro-skill does one thing exceptionally well using self-consistency, program-of-thought, and plan-and-solve patterns. Enhanced with agent-creator principles and functionality-audit validation. Perfect for building composable workflow components.

Use PROACTIVELY when building multi-step agentic workflows with validated tool parameters. Implements Anthropic's strict tool use mode for guaranteed schema compliance. Covers tool schema design, multi-tool agent workflows, error handling, testing, and production patterns. Not for data extraction or classification tasks (use json-outputs-implementer instead).

Router de intención para Obsidian en Codex CLI. Selecciona la skill más adecuada (o pide la mínima aclaración), aplica guardrails de lectura real y ejecuta el flujo correcto sin inflar AGENTS.md.

Creates and improves Claude Code skills using best practices. Use when user wants to build, modify, or enhance a skill. Guides through skill architecture, progressive disclosure, workflow design, and testing patterns.

Generate pytest unit tests for Python functions and classes with fixtures, parametrization, and edge cases. This skill should be used when creating test coverage for modules, especially during refactoring to ensure functionality is preserved.

Modular design pattern for external service integrations (CAPTCHA, payment, SMS, file storage, analytics). Use when implementing third-party services, designing service abstractions, or reviewing service architecture. Triggers on /src/services/* files, integration questions, and provider implementations.

Workflow templates for orchestrating multi-agent development processes (feature, security, bugfix)

Quick reference cheatsheets for Kailash SDK patterns, nodes, workflows, and best practices. Use when asking about 'quick tips', 'cheat sheet', 'quick reference', 'common mistakes', 'node selection', 'workflow patterns library', 'cycle patterns', 'production patterns', 'performance optimization', 'monitoring', 'security config', 'multi-tenancy', 'distributed transactions', 'saga pattern', 'custom nodes', 'PythonCode data science', 'ollama integration', 'directoryreader patterns', or 'environment variables'.

Determine package manager, workspace configuration, and available commands in TypeScript/JavaScript projects. Use when agent needs to understand how to run commands in a TS/JS project, including (1) Detecting the correct package manager (npm/yarn/pnpm/bun), (2) Understanding workspace/monorepo setups, (3) Finding available scripts across packages, (4) Determining the correct command syntax for running development scripts

Create D3 questions with drag-and-drop matching interactions. Students drag items (tables, graphs, equations) to categories.