Enable AI agents to read and edit Obsidian notes
Imagine telling an AI assistant to organize a document and watching it understand the actual file structure to make changes. This repository provides a set of instructions that allow AI agents to understand and manipulate Obsidian's unique format. It enables agents to go beyond simple text handling and utilize Obsidian's link structure and database features.
The Obsidian language agents learn
This repository provides skill files that help agents learn Obsidian's special syntax. Unlike standard Markdown, Obsidian uses unique syntax such as wiki links, embeds, and callouts; these skills guide agents to create and edit these elements correctly. It also covers Obsidian's database feature, Bases, and its visualization tool, JSON Canvas files, allowing agents to manage complex note structures systematically.
Supported agents and installation
These skills follow the standardized Agent Skills specification, making them usable on various agent platforms such as Claude Code, Codex, and Open Code. Installation can be performed simply through a marketplace plugin or by running the npx skills add command. For manual installation, you can copy the skill files to the appropriate folder structure for each platform or clone the repository; for Open Code, you must clone the entire repository for the directory structure to be recognized correctly.
Key skill features
The provided skills cover various tasks in the Obsidian environment.
- Obsidian Markdown Creating and editing Markdown files, including Obsidian-specific syntax like wiki links and callouts.
- Bases database Managing Obsidian Bases files with views, filters, formulas, and summary features.
- JSON Canvas Structural editing of canvas files composed of nodes, edges, and groups.
- CLI interaction Accessing vaults via the Obsidian CLI and supporting plugin development.
- Web content extraction Extracting Markdown from web pages by removing unnecessary elements to save tokens.
Pre-use considerations
This repository is suitable for developers or advanced users who utilize AI agents in an Obsidian environment. Before installing, you should verify that your agent platform supports the Agent Skills specification, and for manual installation, you must strictly follow the specific folder structure of each platform. Specifically, if you are using Open Code, you should not just copy the internal skills/ folder but must clone the entire repository for the skills to be discovered properly. Since the README does not specify concrete performance figures or limitations, it is recommended to test the behavior directly on large vaults.
