workspace-organizing

Organizes agent workspace files into purpose-based folders with pre-write Vault and memory discovery.

3.6k|1.0k|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/nextlevelbuilder/goclaw --skill workspace-organizing
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: workspace-organizing
Source: https://github.com/nextlevelbuilder/goclaw/tree/main/skills/workspace-organizing
Command: npx skills add https://github.com/nextlevelbuilder/goclaw --skill workspace-organizing

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agent workspaces quickly become cluttered with loose files, duplicates, and inconsistent naming, making it hard for agents and humans to find prior work. This Skill enforces a predictable folder convention and requires semantic discovery (Vault, memory, knowledge graph) before any file write, preventing duplication and lost deliverables.

Core Features & Use Cases

  • Purpose-Based Folder Conventions: Two modes — flat mode (notes/, data/, outputs/, scripts/, archive/) for ad-hoc work and project mode (projects/<slug>/{docs,assets,source,reports,research}/) for named multi-file projects.
  • Pre-Write Discovery: Runs vault_search, memory_search, and knowledge_graph_search before writing files, with correct id routing (doc_id to vault_read, entity_id to knowledge_graph_search, episodic_id to memory_expand).
  • Team & Delegate Workspace Rules: Per-agent namespacing under shared/<agent_key>/, protected shared/_common/ areas, and delegate workspace output publication with SUMMARY.md manifests.
  • Use Case: When a user says "clean up my workspace" or "where did I save that report", the agent lists loose files, searches the Vault for related docs, proposes a move plan, and reorganizes into the convention without deleting anything.

Quick Start

Ask the agent to organize the current workspace and find any existing files related to your topic before creating new ones.

Frequently Asked Questions about workspace-organizing

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

FAQPage Schema
How do I organize an AI agent workspace folder structure?

Use a purpose-based convention: flat mode with notes/, data/, outputs/, scripts/, and archive/ for ad-hoc work, or project mode with projects/<slug>/ containing docs/, assets/, source/, reports/, and research/ for named multi-file projects. Never write files to the workspace root.

How do I prevent duplicate files in a shared agent workspace?

Run vault_search before any file write to check whether a related document already exists across vault docs, memory, and the knowledge graph. If results exist, update or link to the existing file instead of writing a new one.

What is the difference between flat mode and project mode?

Flat mode suits ad-hoc single deliverables using notes/, data/, outputs/, scripts/, and archive/. Project mode suits named multi-file work spanning sessions, using projects/<slug>/ with docs/, assets/, source/, reports/, and research/ subfolders. The two modes never nest.

Can agents write files into a teammate's shared workspace folder?

No. In team scope, every write goes under shared/<agent_key>/ for the current agent, and writing, moving, or deleting files in another agent's namespace is forbidden. Reading other namespaces is allowed, and shared/_common/ is reserved for team-wide artifacts.

Why does vault_read fail on some vault_search results?

vault_search returns three id types that route to different tools: doc_id goes to vault_read, entity_id goes to knowledge_graph_search, and episodic_id goes to memory_expand. Passing an entity_id or episodic_id to vault_read will fail.

When should workspace organizing not be applied?

Skip it for read-only operations, edits inside an existing project tree such as a cloned repository, ephemeral /tmp files, and short-lived files deleted within the same turn. Also do not reorganize a user's personal workspace unless they explicitly ask.