Hermes Agent

Hermes Agent is a tool-using AI agent that can run in terminals, messaging platforms, IDEs, and automation workflows. It is not just a chatbot: it can read and write files, execute shell commands, automate browsers, schedule jobs, search prior sessions, and delegate work to subagents.

The core vocabulary

TermMeaningExamples
CLI commandA terminal command that controls Hermes itselfhermes doctor, hermes model, hermes skills list
Slash commandA command typed inside a running session/model, /skill, /reset, /cron
ToolsetA group of tools that can be enabled or disabled togetherterminal, file, browser, skills
API toolA callable function exposed to the agent in a sessionread_file, terminal, search_files, cronjob
SkillA reusable procedure document for a taskdebugging, GitHub workflows, Obsidian publishing
ProfileAn isolated Hermes runtime with separate config, sessions, skills, and memorypersonal, work, experiment profiles

Why this matters

The words “command,” “tool,” and “skill” sound similar, but they operate at different layers.

  • Commands control Hermes.
  • Tools let Hermes act on the outside world.
  • Skills teach Hermes how to perform a task reliably.

To use Hermes well, choose the model, the toolsets, and the relevant skills together.

Start here

GoalStarting point
Check environment healthhermes doctor, hermes status --all
Change modelshermes model or /model
Inspect toolshermes tools list or /toolsets
Inspect skillshermes skills list or /skills
Check the messaging gatewayhermes gateway status or /platforms
Check scheduled jobshermes cron list or /cron

In this series

  • Commands — terminal and in-session commands.
  • Tools — how the agent acts outside chat.
  • Skills — reusable workflows for specific tasks.