file-sync-management

Synchronize files between manager and Workers via MinIO with Matrix notifications.

5.3k|652|Updated Feb 21, 2026
One-click install
npx skills add https://github.com/agentscope-ai/HiClaw --skill file-sync-management
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: file-sync-management
Source: https://github.com/agentscope-ai/HiClaw/tree/main/manager/agent/skills/file-sync-management
Command: npx skills add https://github.com/agentscope-ai/HiClaw --skill file-sync-management

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill prevents stale or missing files when coordinating with Workers by ensuring every write and read is synchronized through MinIO, not local storage.

Core Features & Use Cases

  • MinIO push after writes: Immediately uploads files produced in /root/hiclaw-fs/ so Workers can reliably access them.
  • MinIO pull before reads: Pulls the latest Worker outputs into local storage before using results, avoiding outdated local copies.
  • Deterministic notification workflow: Uses Matrix @mention to tell the target Worker to run their file-sync step after MinIO updates.

Use case: you instruct a Worker to generate task artifacts, and when it reports completion you sync the task directory from MinIO, read result.md, and proceed with the next collaboration step.

Quick Start

After a Worker finishes, pull the task directory from MinIO into /root/hiclaw-fs/, then read the generated result.md for the task outcome.

Frequently Asked Questions about file-sync-management

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I synchronize files between manager and worker agents using MinIO?

MinIO file synchronization pushes produced files to object storage after writes and pulls the latest worker outputs before reads. Explicit `mc cp` or `mc mirror` operations with overwrite semantics ensure local `/root/hiclaw-fs/` never assumes stale data.

Why does my multi-agent workflow read outdated worker artifacts from local storage?

Outdated worker artifacts occur when local storage is treated as current. Pulling worker outputs from MinIO before reading results prevents stale files, ensuring every read reflects the latest synchronized task directory data.

What is the best way to notify a worker to run file synchronization after a MinIO update?

The best way to notify a worker after a MinIO update is using a deterministic Matrix @mention workflow. This tells the target worker explicitly to run their file-sync step once the manager updates the shared filesystem.

Can I use `mc mirror` for multi-agent task directory sync instead of copying single files?

Yes, `mc mirror` handles task directory sync by applying overwrite semantics to pull entire worker output directories from MinIO. This strict ordering ensures local storage matches object storage before reading `result.md`.

When do I need explicit overwrite semantics for file synchronization in a shared filesystem?

Explicit overwrite semantics are needed whenever pushing files workers need or pulling worker outputs. Strict ordering guarantees local `/root/hiclaw-fs/` is never assumed current, preventing missing or outdated files during multi-agent collaboration.

Does file synchronization in a multi-agent runtime require pulling from MinIO before every read?

Yes, pulling from MinIO before reads is required to avoid outdated local copies. This file synchronization approach ensures manager agents access the latest worker outputs and task artifacts before proceeding with collaboration steps.