context-infrastructure-en

Background reading: Why AI Only Says Correct But Useless Things — and How to Push It Out of Its Comfort Zone

This is the complete structure of a context infrastructure system that has been running for a year. Its main value is as a reference implementation: you can see what the system looks like, how data flows, and how memory accumulates.

Core positioning: This is not a plug-and-play tool. It is a blueprint you can study. After cloning, you can immediately experience the difference between "with context" and "without context." But to make AI truly your own, you need to collect your behavioral data from scratch. There is no shortcut.

Quick Start (5 minutes)

Then open rules/USER.md and fill in your basic information. This is the highest-ROI step: AI behavior becomes personalized immediately.

Detailed steps in setup_guide.md.

If you want to extend this into a more complete working system, see docs/SKILL_ECOSYSTEM.md. It lists independently installable public skill repos: web search, Google Docs, Google Maps, email/newsletter, OpenCode, PPTX, social media, payment analytics, home network analysis, and a local process launcher. context-infrastructure-en stays lightweight; full capabilities are installed on demand through separate repos.

Directory Structure

Semantic search is now a standalone public skill repo: semantic-search-skill, no longer under tools/.