Clawfleet
FeaturesPersonal AIFor AgenciesPricingBlog
Sign inDeploy now
Clawfleet

The simplest way to deploy and manage OpenClaw instances. Managed hosting for individuals. Deploy your AI assistant in 60 seconds.

Need multiple instances for your team or clients? Contact us.

Clawfleet is not affiliated with, endorsed by, or connected to the OpenClaw project. OpenClaw is an independent open-source project.

© 2026 Clawfleet. All Rights Reserved. Built with ♥ by Seven Hills Software

@harishganapathi

Product
  • Features
  • Pricing
  • Integrations
  • Changelog
  • Switching hosts?
Resources
  • Personas
  • Use Cases
  • Glossary
  • Blog
About
  • Contact
  • FAQ
Legal
  • Terms of Service
  • Privacy Policy
  • Cookie Policy
  • Refund Policy
Glossary/AGENTS.md

What is AGENTS.md?

The workspace configuration file for OpenClaw's behavior and tools

Definition

AGENTS.md (also sometimes called the workspace config) defines the operational rules for an OpenClaw workspace — what tools are available, what files the agent can access, what commands it can run, and how it should behave in the development environment.

How it works

OpenClaw reads AGENTS.md when initializing a workspace and uses it to understand the scope of what it's allowed to do. This file typically documents available tools, approved commands, file structure conventions, and context about the project the agent is working in. For Clawfleet deployments, this is pre-configured for your specific use case.

Why it matters

AGENTS.md is the safety and scoping layer. It prevents an agent from doing things outside its intended role and gives it enough context to be effective within that role. Well-configured AGENTS.md files result in agents that are both more capable (they know what tools to use) and safer (they know what not to touch).


Run OpenClaw without the setup

Clawfleet manages your OpenClaw instance — AGENTS.md, backups, restarts, and cost tracking — all included. Start for $1.

Deploy for $1 →

Related terms

SOUL.mdThe personality file that defines who your OpenClaw agent isBOOT.mdStartup instructions that run every time OpenClaw initializesOpenClaw HooksEvent-driven triggers that fire on specific agent lifecycle eventsOpenClaw SkillsReusable capabilities you can install into your OpenClaw agent
← Back to Glossary