Mogplex Docs
CLIGuides

Project Commands

Use AGENTS.md, .agents/commands, and /init to make Mogplex behave like part of the repository instead of a generic terminal tool.

The CLI gets much more useful when the repository teaches it how to work.

Two local files matter most:

  • AGENTS.md at the project root
  • markdown slash commands in .agents/commands/

AGENTS.md

AGENTS.md is the repo-local instruction layer.

Use it for stable information that should travel with the codebase, such as:

  • project purpose
  • build, test, and lint commands
  • layout and ownership conventions
  • naming or error-handling patterns
  • repo-specific guardrails the assistant should follow

This is the fastest way to make Mogplex stop acting generic when it enters a repo.

Project slash commands

Markdown files in .agents/commands/ become repo-scoped slash commands.

That lets you keep common project workflows close to the code, such as:

  • a standard review prompt
  • a release-note or changelog helper
  • a repo-specific refactor checklist
  • a debugging playbook for the local stack

Project commands are usually a better fit than shell aliases when the behavior is prompt-shaped rather than command-shaped.

User slash commands

For personal commands that should follow you between repos, use:

~/.mogplex/commands/

These are loaded alongside project commands.

Bootstrap with /init

Inside a live session, /init can scaffold the local setup for you.

Today it is designed to:

  • create an AGENTS.md at the project root
  • create a starter .agents/commands/review.md

That makes /init the fastest way to turn an unprepared repo into one with basic Mogplex conventions.

How this fits with slash discovery

The loader merges:

  1. built-in commands
  2. project commands from .agents/commands/
  3. user commands from ~/.mogplex/commands/

It can also apply config-driven aliases, disabled commands, and extra user directories through the slash_commands config section.

For a healthy repo setup:

  1. keep durable repo instructions in AGENTS.md
  2. keep repeatable repo workflows in .agents/commands/
  3. keep personal helpers in ~/.mogplex/commands/

That separation keeps project behavior sharable without turning your personal setup into a repo requirement.

Edit on GitHub

On this page