Thanks to everyone who built a skill for the OpenSearch Agent Skills Hackathon. Submissions closed August 17, 2026, and our judges are reviewing every entry now.
Join the Slack channel for updates → #hackathon-agent-skills
What happens next
We know you want to know how you did. Here’s the timeline:
August 18-30, 2026
Judging window
September 1, 2026
Winners announced
Results will be posted on this page, social media and the OpenSearch Slack channel
If you submitted a skill, you don’t need to do anything else. We’ll reach out if we need anything from you.
// your skill, running in the IDE
$ agent skill install ./log-triage-skill
Analyzing OpenSearch cluster context…
✓ Skill loaded: log-triage-skill v0.1 $ agent run "why did latency spike at 2am?"
Querying observability index, correlating traces…
✓ Root cause identified in 3 steps
$ agent skill install ./log-triage-skill
Analyzing OpenSearch cluster context…
✓ Skill loaded: log-triage-skill v0.1 $ agent run "why did latency spike at 2am?"
Querying observability index, correlating traces…
✓ Root cause identified in 3 steps
Didn’t get your submission in?
No hard feelings. Building agent skills for OpenSearch is an ongoing effort, not a one-time event.
Here are a few ways to stay involved:
Browse the OpenSearch Agent Skills repository and see what’s already been built.
View repository
View repository Join the Hackathon interest list. This round was open to U.S. residents only — future events will be offered globally, so make sure you know from day one.
Join
Join Questions about this round? Ask in #hackathon-agent-skills Slack channel.
Want to check the official rules or judging criteria?
Evaluation framework
Judges score on five dimensions
Submissions are evaluated for enterprise-grade value and alignment with OpenSearch Project open source principles, by maintainers, collaborators, and OpenSearch Ambassadors.
Technical innovation – 25%
Does the skill use OpenSearch vector database or agentic capabilities in a novel way? Examples include advanced RAG patterns or multi-step reasoning.
Community creativity and utility – 25%
How innovative, original or resourceful is the use-case or the solution to which it can be applied?
How broadly can this skill be applied? Does it break down data silos or address common pain points for SREs, Search Engineers, or InfoSec Specialists?
Documentation and enablement – 15%
Is the README clear and actionable? Is there a 5-minute video that communicates value to a Tech Lead or CTO?
Vendor neutrality and portability – 15%
Is the skill free of proprietary dependencies? Does it run on any OpenSearch distribution, including self-managed, Docker, and Kubernetes?
Operational excellence – 20%
Is the code production-ready? Does it follow security best practices and deliver low-latency performance?
The judging process
Judges have been selected by the OpenSearch community, consisting of project maintainers, collaborators, and OpenSearch Ambassadors.
