[GSoC 2026] Convert Dubbo capabilities into AI Skills.

[RainYuY]

Pre-check

• I am sure that all the content I provide is in English.

Search before asking

• I had searched in the issues and found no similar feature requirement.

Apache Dubbo Component

Java SDK (apache/dubbo)

Descriptions

https://issues.apache.org/jira/projects/DUBBO/issues/DUBBO-149

https://cwiki.apache.org/confluence/display/COMDEV/GSoC+2026+Ideas+list?show-miniview#GSoC2026Ideaslist-GSoC2026-ConvertDubbocapabilitiesintoAISkills.

Related issues

No response

Are you willing to submit a pull request to fix on your own?

• Yes I am willing to submit a pull request on my own!

Code of Conduct

• I agree to follow this project's Code of Conduct

[mohit7705]

Hi @RainYuY , I was going through the idea of converting Dubbo capabilities into AI Skills and had a few clarifications.

1. For the “AI Skills” representation, is there any preferred standard or format you expect (e.g., JSON-based schema, OpenAPI, or something aligned with MCP/agent frameworks)?

2. Should the skills be limited to documentation-style structured descriptions, or is the expectation to make them directly usable by AI agents (e.g., integration with LangChain or similar frameworks)?

3. Would it be valuable if the project also includes a small prototype demonstrating how an AI agent can utilize these Dubbo skills (for example, generating service configurations or workflows)?

I’m very interested in working on this idea and would love to align early with the expected direction.

Thanks!

[RainYuY]

Hi @RainYuY , I was going through the idea of converting Dubbo capabilities into AI Skills and had a few clarifications.

1. For the “AI Skills” representation, is there any preferred standard or format you expect (e.g., JSON-based schema, OpenAPI, or something aligned with MCP/agent frameworks)?
2. Should the skills be limited to documentation-style structured descriptions, or is the expectation to make them directly usable by AI agents (e.g., integration with LangChain or similar frameworks)?
3. Would it be valuable if the project also includes a small prototype demonstrating how an AI agent can utilize these Dubbo skills (for example, generating service configurations or workflows)?

I’m very interested in working on this idea and would love to align early with the expected direction.

Thanks!

1. No, that's not the standard. We just care more about which approach will be more effective.
2. At least, it should have a documentation-style structure, like Claude's code skills.
3. Yes.
   Specially. This is a Gsoc issue, only can be apply by students. It will stop apply at 3.31. If you quailty these asked. You should follow gsoc guide. You can see it at apache-gsoc website.

[LI123456mo]

Hi @RainYuY , I am interested on working in this project. I have gone through the issue and requirements, thanks

[Sourabh-Majhee]

Hi @RainYuY ,

I have submitted a GSoC 2026 proposal for this idea: “Convert Dubbo Capabilities into AI Skills.” I wanted to briefly align here as well.

My approach focuses on building a structured SKILL.md-based knowledge layer (inspired by Claude Code skills) that covers Dubbo’s core modules, including RPC, Remoting, Registry, Cluster, Config, and Metadata. Each skill is designed to be self-contained and optimized for LLM consumption, with clear architectural explanations, well-defined SPI extension points, and practical guidance for real-world developer scenarios such as usage, debugging, and extension.

In addition, I plan to:

• Generate a module dependency graph (using Maven) to guide skill composition and cross-referencing
• Incorporate benchmark-based validation, comparing the quality of AI-assisted responses (e.g., Copilot/Claude) before and after loading the skills
• Optionally demonstrate a prototype workflow where AI tools utilize these skills to assist with Dubbo configuration or troubleshooting

From your earlier response, I understand that effectiveness and a documentation-style structure are key priorities, which aligns well with this approach.

I would greatly appreciate any feedback on whether this direction aligns with Dubbo’s expectations, or if there are specific modules or areas you would recommend prioritizing.

Thank you.