Description

To improve the effectiveness of AI agents working on this repository, we should add an AGENTS.md file in the root directory. This file will serve as a definitive guide for AI agents (and human contributors), providing context, code standards, and standard commands for development and testing.

We should take inspiration from other Kubeflow repositories that have successfully implemented this, such as trainer, pipelines, and sdk.

Goal

Create an AGENTS.md file that includes the following sections:

1. Who This Is For: Clarify the audience (AI agents, contributors, maintainers).
2. Agent Behavior Policy: Define rules for AI agents (e.g., atomic changes, running verification before committing, not modifying CI without instruction).
3. Repository Map: A high-level tree structure of the docs-agent repository explaining the purpose of key directories (server, server-https, pipelines, manifests, etc.).
4. Environment & Tooling:
   • Primary languages (Python).
   • Key dependencies (Milvus, KServe, Kubeflow Pipelines).
   • Development tools (Helm, kubectl, etc.).
5. Commands:
   • Standard commands for setup, running servers, and testing.
   • Note: Since this repo currently lacks a Makefile, we might want to standardize some of these commands or explicitly list the python and helm commands found in README.md.
6. Development Workflow:
   • Preferred workflows for making changes.
   • Commit message conventions.
7. Core Development Principles:
   • Code quality standards (Python style, typing).
   • Documentation requirements.

Reference Examples

• Kubeflow Trainer AGENTS.md
• Kubeflow Pipelines AGENTS.md
• Kubeflow SDK AGENTS.md

Tasks

• Create AGENTS.md in the root directory.
• Populate it with the sections outlined above, tailored to the docs-agent specifics (RAG, Milvus, KServe context).
• (Optional) Create a Makefile to standardize commands referenced in AGENTS.md, aligning with other Kubeflow repos.