Skip to content
← Back to work

Mini Agentflow / AI Research Workflow System

A lightweight workflow system for AI-assisted research practice.

A system-oriented project for organizing AI-assisted research workflows, task decomposition, and repeatable analysis routines.

2026 · System designer and developer

AI Workflow/Research Tools/LLM Agents

system diagram / screenshot

Context

Research work often involves many repeated setup, analysis, and documentation steps.

Problem

Without a workflow layer, AI-assisted research can become scattered, hard to audit, and hard to reuse.

Approach

Build a compact system that structures tasks, prompts, artifacts, and iteration loops around research work.

Outcome

Useful as a public signal of research tooling and systems thinking; details can be expanded when links are ready.

Artifacts

  • Workflow notes
  • Prototype code

Notes

Mini Agentflow represents the practical systems side of Xunzhuo’s work: building tools that make research workflows easier to repeat, inspect, and improve.

The first public version describes the project at a high level and avoids implying a polished product before a public release exists.