filesystem-context

Store and retrieve agent context using the filesystem as a persistent overflow layer.

Updated Jun 29, 2026
One-click install
npx skills add https://github.com/wangyouan/codex-personal-kit --skill filesystem-context-wangyouan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: filesystem-context
Source: https://github.com/wangyouan/codex-personal-kit/tree/main/skills/filesystem-context
Command: npx skills add https://github.com/wangyouan/codex-personal-kit --skill filesystem-context-wangyouan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill tackles the challenge of context overflow by leveraging the filesystem as a primary overflow layer, allowing agents to store, retrieve, and update an effectively unlimited amount of context.

Core Features & Use Cases

  • Filesystem Overflow Layer: Utilize the filesystem to manage context that doesn't fit within the current window, including tool outputs, intermediate results, and shared state between sub-agents.
  • Dynamic Context Discovery: Focus on on-demand loading of context to enhance token efficiency and reduce irrelevant information.
  • Use Case: Imagine an agent needs to perform a complex analysis of a dataset. The skill can help by persisting intermediate results to the filesystem, enabling the agent to reference them without overwhelming the context window.

Quick Start

Use the filesystem-context skill when your agent requires more context than can be held in the current window.

Frequently Asked Questions about filesystem-context

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

FAQPage Schema
How do I manage agent context overflow when processing large datasets?

Agent context overflow is managed by offloading tool outputs and intermediate results to the filesystem as a persistent storage layer. This enables agents to store and retrieve an effectively unlimited amount of context without overwhelming the current window.

What is the best way to persist plans and intermediate states for multi-turn agent workflows?

Plan persistence and intermediate state retention are best handled by using a file-based system for sub-agent communication. This strategy ensures long-term context retention across multi-turn tasks by writing shared states to disk dynamically.

How do I improve token efficiency during complex agent workflows?

Token efficiency improves through dynamic context discovery, which focuses on on-demand loading of context from the filesystem. This mechanism reduces irrelevant information by fetching stored intermediate results only when the agent requires them.

Can I use the filesystem for sub-agent communication and shared state management?

Yes, the filesystem supports sub-agent communication by acting as a shared state layer. Sub-agents can write intermediate results and tool outputs to files, enabling other agents to discover and retrieve this context dynamically.

When should I use a filesystem-based context management strategy instead of keeping everything in the context window?

A filesystem-based context management strategy is necessary when your agent requires more context than can fit within the current window. It is specifically optimized for multi-turn tasks and complex workflows that demand long-term context retention.