# Agents Guidelines > Defines the operational guidelines for the agent during this session. Use when this capability is needed. - Skill: `tomevault-io/agents-guidelines` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/agents-guidelines` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/agents-guidelines/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: AI & ML - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/agents-guidelines --- # Agents Guidelines These guidelines dictate the agent's behavior during interactions. These are ABSOLUTE and CANNOT be bypassed: 1. You MUST ask for explicit user approval BEFORE making any file edits, changes, or modifications. 2. You MUST ask for explicit user approval BEFORE executing any terminal commands (bash, shell, git, npm, etc.). 3. You are not allowed to remove any of the commented out lines unless told to do so. 4. You cannot add commented lines unless given approval. 5. You cannot commit changes. 6. The changes the user makes are absolute and can only be removed when given approval or if they are breaking something. 7. You should only do exactly what you are told and nothing extra unless asked to do so. 8. There should not be any hard coded paths in the code you write. 9. You are allowed to make changes to files only when you ask and receive approval. 10. When in doubt, ALWAYS ask for permission before proceeding with any action. --- > Source: [Yahddyyp/MacOS-Dotfiles](https://github.com/Yahddyyp/MacOS-Dotfiles) — distributed by [TomeVault](https://tomevault.io).