OpenSearch Agent Expertise carry built-in intelligence to your agentic IDE

In the present day, we’re launching OpenSearch Agent Expertise, a repository of open, composable abilities that carry built-in intelligence to developer workflows with OpenSearch, straight inside your favourite agentic IDE. By embedding OpenSearch experience into the developer’s present workflow, Agent Expertise cut back setup time, get rid of pointless tool-hopping, and let groups deal with constructing relatively than configuring.

Builders in the present day can go from thought to working prototype in minutes utilizing agentic IDEs like Claude, Cursor, and Kiro. They’ll spin up purposes, generate APIs, and construct end-to-end workflows with a immediate. However whether or not you’re experimenting with a brand new thought, constructing a POC, or operating manufacturing programs, the expertise shortly turns into extra complicated. For instance, bettering relevance in OpenSearch nonetheless requires deep experience in question Area-Particular Language (DSL), rating logic, and hybrid search tuning. Troubleshooting latency or cluster well being points typically means manually piecing collectively indicators from logs, traces, shards, and infrastructure metrics. Even migrations from Elasticsearch or Solr can change into complicated and time-consuming due to schema conversion, compatibility gaps, and efficiency optimization challenges. As AI brokers change into a main interface for constructing and working purposes on OpenSearch, a deeper hole emerges. Translating high-level intent into question DSLs, index configurations, and multi-step workflows nonetheless requires important experience. On the identical time, workflows stay fragmented throughout domains like search, logs, and observability, forcing groups into siloed tooling and disconnected reasoning. The result’s repeated trial-and-error, lack of standardized approaches, and slower time-to-value, regardless of the promise of quicker growth.

What are Agent Expertise?

Agent Expertise, developed by Anthropic, are a light-weight, open format for extending AI agent capabilities with specialised data and workflows. They’re supported by a rising variety of AI instruments and agentic shoppers, together with Kiro, Claude Code, Cursor, VS Code, GitHub Copilot, Codex and others.

At their core, Agent Expertise are pre-built intelligence you possibly can name, lengthen, and reuse. Every ability encapsulates area data, execution logic with multi-step workflows, and steerage with explainability, so that you not solely get outcomes however perceive how they’re achieved. As a substitute of sewing collectively instruments and writing customized logic, you possibly can invoke a ability to deal with a complete job, from evaluation to suggestion to execution.

At launch, OpenSearch Agent Expertise introduces three foundational abilities designed to deal with a number of the commonest and sophisticated developer workflows: Search, Logs, and Solr to OpenSearch Migrations.

Search ability

The Search Ability builds on the muse launched by OpenSearch Launchpad, and brings an agentic, intent-driven expertise to constructing and optimizing search purposes with OpenSearch. Builders can go from a easy requirement or pattern doc to a completely working search software in minutes, whether or not lexical, semantic, hybrid, or agentic, with no

deep OpenSearch experience required.

What it does:

• Interprets pure language necessities or pattern knowledge into search configurations.
• Robotically creates index mappings, ingest pipelines, and ML mannequin integrations.
• Units up key phrase, semantic, and hybrid search capabilities out of the field.

Instance

Construct a semantic search software for product documentation

Output:

• Totally configured OpenSearch index with optimized mappings.
• Built-in embedding fashions and ingest pipeline.
• Working search expertise (API + UI) prepared to check and iterate.

The Search Ability builds on the muse launched by OpenSearch Launchpad, extending the identical capabilities into an agent-native workflow. You possibly can transfer from thought to a production-ready search software in minutes, eliminating guide setup and accelerating each prototyping and deployment in OpenSearch.

Logs ability

The Log Ability analyzes log knowledge and investigates distributed traces straight inside OpenSearch, bringing agentic intelligence to observability workflows. As a substitute of manually crafting PPL queries or piecing collectively hint knowledge throughout providers, builders can categorical their intent and let the ability

deal with the complexity.

What it does:

• Queries and analyzes log knowledge utilizing PPL, together with error patterns, log quantity traits, and anomaly detection.
• Investigates distributed traces, figuring out sluggish spans, error spans, service dependencies, and agent invocations.
• Correlates logs and traces utilizing traceId to floor root causes throughout the complete observability stack.

Instance:

Examine why my service is returning 500s and correlate with latest traces

Output:

• PPL question outcomes surfacing error patterns and log quantity anomalies.
• Hint evaluation figuring out sluggish or failing spans and repair dependencies.
• Correlated view linking log errors to particular hint IDs for quicker root trigger evaluation.

With the Logs Ability, you possibly can transfer from a obscure symptom to a pinpointed root trigger in minutes with no need to grasp PPL syntax or manually navigate hint knowledge.

Solr to OpenSearch migration ability

The Migration Ability streamlines the complicated strategy of migrating from Solr to OpenSearch. Migrations usually contain cluster discovery, compatibility checks, schema translation, knowledge motion, and validation. These steps typically require deep experience and guide coordination. The

Migration ability turns all these steps right into a guided, automated workflow.

What it does:

• Discovers and analyzes supply clusters, together with indices, mappings, and configurations.
• Performs compatibility evaluation and highlights breaking modifications or required transformations.
• Interprets schemas, index settings, and queries into OpenSearch-compatible codecs.

Instance:

How can I migrate from Solr to OpenSearch?

Output:

• Detailed migration plan with compatibility report and required modifications.
• Translated index mappings and configurations prepared for OpenSearch.
• Executed knowledge migration pipeline with progress monitoring.
• Validation report confirming knowledge integrity and question parity between supply and goal.

With the Migration Ability, builders can transfer from a fragmented, high-risk migration course of to a structured, automated workflow. This strategy gives quicker transitions, decreased downtime, and confidence in manufacturing readiness.

The way it works

OpenSearch Agent Expertise are organized as a tree of SKILL.md recordsdata, structured by area class. Relatively than one monolithic ability that hundreds all the pieces, the repo is damaged into centered, independently installable abilities. Every ability is sufficiently small to remain inside a decent context window, however

full sufficient to deal with actual end-to-end workflows.

The highest-level construction at the moment teams abilities into three classes:

• Search: opensearch-launchpad for constructing BM25, semantic, and hybrid search purposes from scratch.
• Observability: log-analytics for PPL-based log querying and error evaluation, and trace-analytics for distributed hint investigation and span evaluation.
• Cloud: aws-setup for deploying to Amazon OpenSearch Service (managed) or Amazon OpenSearch Serverless, with separate manifests for every.

Every ability bundles all the pieces the agent wants: step-by-step workflows, reference docs (like PPL syntax guides and CLI references), and executable scripts that run straight in opposition to your cluster.

Whenever you say “construct a hybrid search app” or “why is my service throwing 500 errors?”, the agent prompts solely the matching ability, follows its directions, and executes the fitting OpenSearch APIs. It returns outcomes alongside clear explanations of what was configured and why. As a result of abilities load on demand, you possibly can have the complete assortment put in with out bloating your agent’s context window.

We’re repeatedly increasing the ability library. Classes like Dashboard and Migration are already on the roadmap, with extra to return because the ecosystem grows.

Getting began

Getting began with OpenSearch Agent Expertise is simple. No MCP server or extras are required. Expertise are put in utilizing npx abilities and work straight together with your present agentic IDE.

Stipulations:

• Python 3.11+ and uv.
• Docker put in and operating.
• AWS credentials configured (non-obligatory, for cloud deployment).

Set up all abilities:

npx abilities add opensearch-project/opensearch-agent-skills

Or set up a selected ability: (e.g. opensearch-launchpad)

npx abilities add opensearch-project/opensearch-agent-skills@opensearch-launchpad --full-depth

npx abilities add opensearch-project/opensearch-agent-skills@log-analytics --full-depth

npx abilities add opensearch-project/opensearch-agent-skills@trace-analytics --full-depth

npx abilities add opensearch-project/opensearch-agent-skills@migration-companion --full-depth

As soon as put in, merely categorical your intent to your agent, for instance, “I need to construct a semantic search app with OpenSearch,” and the agent reads the ability directions and runs the scripts robotically.

Expertise will also be put in to a selected agent (-a claude-code), globally throughout all initiatives (-g), or to all detected brokers (--all). Discover accessible abilities earlier than putting in with --list.

Trying forward

That is just the start. We’re actively increasing the OpenSearch Agent Expertise ecosystem with new capabilities throughout superior relevance tuning, cost-aware efficiency optimization, index lifecycle and schema evolution, and cross-domain workflows that unify search, logs, and analytics.

Over time, we see Agent Expertise turning into a community-driven data layer throughout OpenSearch domains the place fixing a fancy drawback as soon as means everybody advantages. Extra importantly, Agent Expertise mark a basic shift in how builders construct and function with OpenSearch: transferring away from guide, fragmented workflows towards clever, reusable capabilities that information, optimize, and speed up growth at each stage.

Get entangled

OpenSearch Agent Expertise is designed to be an open, evolving ecosystem, and we’re getting began. Right here’s how one can take part:

• Attempt it in your workflow. Set up the abilities in Claude, Cursor, or Kiro and begin interacting with OpenSearch utilizing pure language. Construct new purposes, examine points, or run migrations, and see how far intent-driven workflows can go.
• Construct and lengthen abilities. Agent Expertise are deliberately modular and extensible. Create your personal abilities to encode domain-specific workflows, inner finest practices, or repeatable operational playbooks. Whether or not it’s a customized relevance tuning circulation or a specialised observability pipeline, your contributions can change into reusable intelligence for others.
• Contribute to the ecosystem. We welcome contributions throughout all ranges, from bettering documentation and fixing bugs to including fully new abilities. Should you’ve solved a fancy drawback with OpenSearch, think about turning it right into a ability and contribute to the Git repo.
• Share suggestions and concepts. Tell us what labored, what didn’t, and what capabilities you’d wish to see subsequent, whether or not it’s deeper integrations, new domains, or extra superior automation.
• Be part of the dialog. Interact with the OpenSearch neighborhood by GitHub discussions, neighborhood boards, and dealing teams. Collaborate with others constructing comparable workflows and assist outline the way forward for agent-driven search and observability.

With OpenSearch Agent Expertise, we’re transferring towards a world the place builders don’t solely use instruments however use shared intelligence. If that resonates with you, we’d love so that you can be a part of the journey.

Star and become involved within the OpenSearch Agent Expertise repo. Be part of the dialog on the OpenSearch neighborhood discussion board and join with us within the OpenSearch Slack channel.

Acknowledgments

We wish to lengthen our honest gratitude to the next contributors for his or her useful contributions to this venture Arjun kumar Giri, Sarat Vemulapalli, Chenyang Li, Fen Qin, Janelle Arita, Kaituo Li, Krishna Kondaka, Owais Kazi, Peter Zhu and Zhichao Geng. Your dedication, experience, and collaborative spirit have been instrumental in making this venture profitable. Thanks on your time and contributions.

In regards to the authors