You are probably already using AI to write code. Claude Code, Cursor, Kiro, GitHub Copilot, or some combination. You prompt, you get functions, you move fast. But then you need to build a search application and the workflow fractures. You leave your IDE to read OpenSearch documentation. You switch to the AWS console to provision a domain. You open a terminal to configure indices and pipelines. You write client code to connect everything. Each step requires a different tool, a different mental model, and a different tab.
Your AI agent can generate the code for each piece. What your agent cannot do (until now) is make the architectural decisions, execute the configuration, and verify that the whole system works together. The agent autocompletes functions but has no idea how to actually stand up a search application end to end.
OpenSearch Agent Skills change that. Agent Skills are executable workflows, packaged as SKILL.md files, that give your AI coding agent the domain expertise to build, configure, and deploy OpenSearch applications from natural language. Not documentation. Not code snippets. Actual capabilities your agent can run against your infrastructure.
What Ships Today
The OpenSearch project ships skills covering the workflows where context-switching hurts most: search application development, log and trace investigation, document processing, and cloud deployment to Amazon OpenSearch Service and Serverless.
