Filesystem Context

Turns out the filesystem is a better context manager than your prompt. This skill shows you how to offload tool outputs, persist plans, and share state between agents using files instead of cramming everything into the context window. The core insight is simple: write once, retrieve selectively. Tool returned 10,000 lines? Write it to scratch, grep for what you need, pull 50 lines back. Long task losing coherence? Persist the plan as YAML and re-read it each turn. The pattern library covers scratchpads, sub-agent handoffs, dynamic skill loading, and self-modifying instructions. It's practical filesystem engineering for context-constrained agents, with clear guidance on when to use files versus actual memory systems.

skills.sh installs: 390
GitHub stars: 848
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add guanyang/antigravity-skills --skill filesystem-context --agent claude-code

Installs into .claude/skills of the current project.

Files

Filesystem-Based Context Engineering

Use the filesystem as the primary overflow layer for agent context because context windows are limited while tasks often require more information than fits in a single window. Files let agents store, retrieve, and update an effectively unlimited amount of context through a single interface.

Prefer dynamic context discovery -- pulling relevant context on demand -- over static inclusion, because static context consumes tokens regardless of relevance and crowds out space for task-specific information.