Agent and workflow runtime · Orchestration

Hermes Agent

Understand Hermes as a broader agent and workflow layer that can sit beside coding agents, rather than assuming it is only another terminal code editor.

Made by Nous ResearchSeparate agent guideReviewed 2026-09-26Model ≠ harness ≠ orchestrator

Scope note This is an independent guide to a third-party project or a developer concept. Installation commands, feature boundaries, plans, model availability and security defaults change. Follow the official source links before using it with private code, credentials or automation.

What it is and who makes it

A provider-agnostic agent runtime with CLI, desktop, tools, skills, memory and messaging or gateway surfaces that can coordinate longer-running workflows.

Maker or project: Nous Research. Classification: Agent and workflow runtime; this page treats the surrounding context, tools, permissions and verification as the relevant harness boundary.

A user or event enters through the CLI, desktop, gateway or messaging surface.

A user or event enters through the CLI, desktop, gateway or messaging surface.

Hermes builds context, selects a provider and chooses tools or skills.

Hermes builds context, selects a provider and chooses tools or skills.

It can run bounded actions, persist session information or delegate work depending on configuration.

It can run bounded actions, persist session information or delegate work depending on configuration.

The operator reviews permissions, logs, checkpoints and final evidence.

The operator reviews permissions, logs, checkpoints and final evidence.

Who it may fit

Pick an agent by its control surface and trust boundary, not only by its model label.

Useful scenarios

Persistent or scheduled personal workflows
Messaging and gateway-driven agents
Coordinating tools, skills, memory and other coding agents

Trade-offs and checks

A long-running gateway or messaging bot has a wider attack surface than a one-shot local CLI.
Memory, scheduled jobs and delegation can persist actions beyond the original chat; set limits and review stored data.
When Hermes shells out to another coding agent, you now have an orchestrator plus a worker harness—debug and secure both.

Interactive vs headless and CI

These modes change the approval, identity and observability boundary.

Interactive CLI and desktop sessions suit personal workflows; gateway, messaging, scheduled and delegated workers are headless-style operation. Run those workers with separate profiles, limits, logs and project isolation.

Request-to-verification workflow

Use this mental model when debugging a surprising edit, tool call or cost.

Input

Intent + context

Your prompt, repository instructions, selected files, history and environment become model context.

Action

Tool + permission

The harness proposes a read, edit, shell, web, MCP or other action; policy decides whether it can run.

Feedback

Result + verification

Command output, compiler errors, tests and your review return to the loop before acceptance.

Setup outline

Start with the smallest safe experiment, then follow the current official documentation for the real installation.

Use the official setup wizard and provider documentation.

Start with a local profile and a read-only or low-risk toolset.

Configure approval, terminal isolation, memory and messaging separately.

If delegating to Codex or Claude Code, keep each worker's project and permissions explicit.

Illustrative workflow

# Follow the official installation first
hermes setup
hermes model
hermes tools

# Start the interactive surface
hermes

# Treat gateway, messaging and delegated coding workers
# as separate trust boundaries.

Instructions, skills, MCP and hooks

These surfaces can shape context or add actions; they are part of the trust boundary.

Hermes uses profiles, skills, tools, memory and MCP-related integrations to shape the runtime. A gateway or delegated worker can inherit more capability than a local chat, so review the effective profile and tool list for every surface.

Security and isolation checklist Use a disposable branch or worktree; keep secrets outside the readable path; require approval for writes, shell, network and deployment actions; inspect commands before allowing them; constrain server bindings; review MCP servers, plugins, skills, hooks and instruction files; and run tests independently of the agent's own summary.

Official sources

Use these sources for current installation, capabilities, plans, releases and security notices.

Official Hermes Agent repository ↗Hermes CLI reference ↗Hermes documentation ↗

Reviewed 2026-09-26. This date is a change-tracking note, not a promise that the product remains unchanged.

Related learning

Related agent guides

Want the foundations behind the workflow?

Study AI fundamentals, prompt design, AI-assisted development, agents and MCP without a forced linear course.

Open AI learning AI tools