Customize Overview
Vyre is designed to be deeply customizable. From rules that guide every agent session to plugins that add new tools — adapt Vyre to fit your team's exact workflow.
Configuration files
Vyre uses a layered configuration system:
| File | Scope | What it does |
|---|---|---|
.vyre/rules/*.md | Repository | Rules for agents working on this project |
.vyre/hooks.json | Repository | Event hooks for CI/CD integration |
.agents/skills/ | Repository | Custom skills for domain-specific tasks |
~/.vyre/config.json | Global | Personal preferences, default model, keybindings |
~/.agents/ | Global | Global skills and rules for all projects |
Commit .vyre/ and .agents/ to your repository. This ensures your entire team benefits from the same configurations without any manual setup.
What's available
Rules
Define persistent coding standards, conventions, and behavioral constraints that every agent session follows.
Skills
Domain-specific instruction sets that activate automatically when the task matches — like a specialized expert for every task type.
Plugins
Extend the IDE itself with VS Code-compatible extensions and native Vyre plugins.
MCP
Connect Agent to external tools via the Model Context Protocol — databases, APIs, Slack, Jira, and more.
Recommended setup
A well-configured Vyre project looks like this:
my-project/
├── .vyre/
│ ├── rules/
│ │ ├── coding-standards.md # TypeScript, naming, comments
│ │ ├── testing.md # Testing conventions
│ │ └── architecture.md # Project structure rules
│ └── hooks.json # Post-agent hooks (tests, format)
└── .agents/
└── skills/
├── api-endpoint/ # How to write API routes here
│ ├── SKILL.md
│ └── examples/
└── react-component/ # How to write components here
├── SKILL.md
└── examples/Start by writing your most important rules — the things you wish every new developer on your team knew on day one. Then add skills for the tasks your team does most often.