[FEATURE] Improving Local Development Workflow for BYO Agents #1259
Description
📋 Prerequisites
- I have searched the existing issues to avoid creating a duplicate
- By submitting this issue, you agree to follow our Code of Conduct
📝 Feature Summary
Improving Local Development Workflow for BYO Agents
❓ Problem Statement / Motivation
Hi kagent team,
I’m exploring ways to improve the local development workflow for BYO agents and custom MCP tools. Specifically, I’d like developers to:
- Run their agent/tool locally
- Interact with real cluster services and APIs
From what I understand, local development today may not fully allow realistic cluster testing, but I’m not 100% sure about all edge cases. I’d appreciate any guidance or corrections.
Do you have recommended patterns or best practices for this kind of workflow?
Is there a standard way in the kagent ecosystem to “mirror” or proxy the cluster environment for local development?
Thanks in advance!
💡 Proposed Solution
Introduce a developer workflow that allows BYO agents and custom MCP tools to run locally while safely interacting with a remote kagent cluster. This workflow would:
- Proxy or mirror the cluster environment to local development, including access to:
- Internal services
- Environment variables
- Secrets
- Kubernetes APIs
- Avoid the need for full container rebuilds and redeploys for every code change.
- Allow debugging and iterative development directly in a local IDE.
- Provide example configurations and scripts to make setup simple and repeatable.
This approach would speed up developmen. for anyone extending kagent with BYO agents or tools.
🔄 Alternatives Considered
My assumption is that, as of now, developers building BYO agents typically either: Mock services and APIs locally or Minikube or kind to simulate the environment.
🎯 Affected Service(s)
None
📚 Additional Context
No response
🙋 Are you willing to contribute?
- I am willing to submit a PR for this feature