# Agentic Pipeline MCP

> Use this skill when working on the Python agentic_ai subsystem or Claude Code MCP integration for this repository, including the LangGraph pipeline, local MCP server, cloud adapters, and the project .mcp.json configuration.

- Skill: `hoangsonww/agentic-pipeline-mcp-2` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add hoangsonww/agentic-pipeline-mcp-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hoangsonww/agentic-pipeline-mcp-2/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/agentic-pipeline-mcp-2

---


# Agentic Pipeline MCP

This skill covers `agentic_ai/` and the repo-local Claude Code MCP setup.

## Working Rules

- Treat `agentic_ai/` as a separate Python project with its own settings and operational assumptions.
- Use the current MCP server implementation as the source of truth for available tools, resources, and prompts.
- Reconcile README claims, Make targets, health checks, and runtime code before saying the subsystem is ready.
- If you wire Claude Code MCP integration, prefer the local Python module entrypoint over hand-wavy docs.

## Validation

- Check `mcp_server/app.py`, `mcp_server/runtime.py`, `agentic_ai/config/settings.py`, and `agentic_ai/requirements.txt` together.
- Confirm whether a requested change belongs in the pipeline, the MCP wrapper, or the deployment adapters.

## References

- For runtime and config details, read `references/pipeline-map.md`.