mastralisted

# Mastra Skill ## When to activate - Building a production AI agent with memory, tools, and multi-step workflows - Setting up RAG (Retrieval-Augmented Generation) pipelines - Orchestrating multi-step AI workflows with branching logic and human-in-the-loop - Adding evaluations to test AI output quality - Building analytical CRMs or AI-powered business automation ## When NOT to use - Simple one-shot Claude API calls with no workflow — use the Claude API skill - Chat UIs without agent logic — use the Vercel AI SDK skill - When you need direct Claude Code integration — Claude's native agent loop is better ## Instructions ### Installation ```bash npm install @mastra/core # With specific integrations npm install @mastra/core @mastra/memory @mastra/rag ``` ### Basic agent ```typescript // src/mastra/agents/analyst.ts import { Agent } from '@mastra/core/agent' import { anthropic } from '@ai-sdk/anthropic' export const analystAgent = new Agent({ name: 'Data Analyst', instructions: `You are a data analyst. When given data, you: 1. Identify trends and anomalies 2. Provide actionable recommendations 3. Format output as structured JSON when asked`, model: anthropic('claude-opus-4-7'), }) // Generate a response const response = await analystAgent.generate('Analyze this sales data: ...') console.log(response.text) // Stream a response const stream = await analystAgent.stream('Summarize Q1 performance') for await (const chunk of stream.textStream) { process.stdo