# Lumina Agentic MCP

> Work on the Python multi-agent pipeline and MCP client in agentic_ai. Use when editing files under agentic_ai/, changing agent orchestration, configuration loading, MCP client connectivity, async execution flow, pipeline startup commands, or cloud deployment wrappers for the Python service.

- Skill: `hoangsonww/lumina-agentic-mcp` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add hoangsonww/lumina-agentic-mcp`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hoangsonww/lumina-agentic-mcp/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: hoangsonww (https://skillmd.com/u/hoangsonww)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/hoangsonww/lumina-agentic-mcp

---


# Lumina Agentic Mcp

## Overview

Use this skill for the Python subsystem under `agentic_ai/`. Keep changes consistent with the existing config-driven, async, multi-agent design rather than turning it into a one-off script.

## Load The Right Reference

- Read `references/file-map.md` when you need to locate the pipeline, agent, config, or MCP surface that owns a behavior.
- Read `references/commands.md` before finishing so validation and startup commands match the subsystem.

## Preserve The Existing Architecture

- Keep pipeline behavior flowing through `core/`, `agents/`, and `mcp_client/` rather than duplicating logic.
- Respect async boundaries and existing command entry points in `agentic_ai/__main__.py`.
- Keep the MCP server contract explicit when changing tool names, prompt names, or resource URIs.
- Preserve configuration-driven behavior through files in `agentic_ai/config/`.

## Validate With What The Repo Actually Supports

- Prefer `python -m compileall agentic_ai` for low-friction validation when dependencies or tests are unavailable.
- Use targeted CLI smoke checks such as `python -m agentic_ai visualize` or `python -m agentic_ai run --task "..."` only when the environment is ready.
- Do not assume `pytest` is runnable just because it appears in `requirements.txt`; verify the test target exists first.