X Bookmarks Skill Graph

X Bookmarks Knowledge Graph

A **skill-graph knowledge system** for organizing X (Twitter) bookmarks as an interconnected graph of ideas, not a flat folder of files.

This repository is an experimental testbed comparing **skill-graph architecture** (interconnected markdown files with wikilinks) against traditional **SKILL.md** files for agent knowledge management.

---

Purpose

This project exists to answer one question:

**How does a skill-graph knowledge system compare to SKILL.md files for agent knowledge retrieval and reasoning?**

Traditional agent skills use monolithic `.md` files with structured frontmatter. This repository tests an alternative: a **graph-native** approach where:

• Every bookmark is a node in a knowledge graph
• Connections are explicit via related: links in YAML frontmatter
• Navigation happens through bidirectional links, not hierarchical folders
• Concept MOCs (Maps of Content) serve as entry points for traversal
• Agents traverse the graph by following pointers, not scanning directories

---

Architecture

Core Principles

Principle	Implementation
Graph-native	Every bookmark is a node; related: links form edges
Bidirectional links	If A links to B, B must link back to A
Progressive disclosure	Read frontmatter (15 lines) → follow related: → read full content only when needed
Concept MOCs	Curated entry points for topic-based exploration
YAML frontmatter	Machine-readable metadata for agent traversal

Repository Structure

x-bookmarks/
├── INDEX.md