Source-checked profile · July 26, 2026

Hermes Agent

Skills, security, setup, and practical fit

A stateful open-source agent from Nous Research for teams that need persistent memory, reusable skills, MCP connections, and controlled execution across terminal and messaging surfaces. This profile checks primary sources instead of repeating promotional benchmarks.

Open sourceMIT licenseSelf-hosted

Official docs ↗Security guide ↗GitHub repository ↗

Decision first

Should you use Hermes Agent?

Start with workflow fit and execution risk. Feature count comes second.

Direct answer

Choose Hermes when persistent context and reusable procedures matter more than a lightweight IDE-only assistant.

Strong fit

Self-hosted assistant with persistent memory and reusable skills.

Strong fit

One agent across terminal and supported messaging surfaces.

Limited fit

Inline IDE autocomplete with minimal operational setup.

High risk

Untrusted public access on a host-level local backend.

How the system works

From goal to reusable skill

Hermes combines an agent loop, tools, memory, and deployment backends. The value comes from the whole operating path—not one isolated feature.

01Goal

Start from a task

Terminal or messaging input gives the agent a concrete objective.

02Agent loop

Reason and act

The loop selects tools, executes work, observes results, and continues.

03Memory + skills

Retain and codify

Useful context persists, and successful procedures can become reusable skills.

04Backend + guards

Choose the surface

Run locally, in a container, or remotely with permissions matched to risk.

Risk noteEvery added tool, backend, credential, and messaging channel expands the permission surface. Start with the narrowest configuration that completes the job.

Source delta

What changed in the primary sources

The page records source changes instead of silently preserving stale claims. This keeps comparison copy auditable.

Release: v0.18.2→v0.19.0
GitHub stars: 215.8K→220.7K
Built-in tools: 60+→40+
Messaging: 20+→Named list

Editorial boundary

Claims removed or narrowed

No fixed 20+ total

Use named, sourced messaging integrations instead.

No 60+ tool claim

The current README supports the narrower 40+ wording.

No security absolutes

Documented layers reduce risk; local execution is not risk-free.

Setup + security

Install deliberately. Expand permissions slowly.

Commands are easy. The real setup decision is which backend, credentials, and integrations the agent is allowed to reach.

Install safely

Verify the official source and choose the backend before adding credentials. Read an installer before running remote code.

Unix / macOS

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

Windows PowerShell

iex (irm https://hermes-agent.nousresearch.com/install.ps1)

First-run checklist

✓ Confirm the official source and current release.
✓ Start with one narrow backend.
✓ Add only required credentials.
✓ Test logs, approvals, and the shutdown path.

hermes setup --portal

Provider setup

hermes doctor

Configuration check

hermes

Start after review

Eight documented security layers

These controls reduce exposure; they do not remove the risk of local execution, broad credentials, or misconfigured backends.

01
User authorization
02
Dangerous-command approval
03
File-write safety
04
Container isolation
05
MCP credential filtering
06
Context-file scanning
07
Cross-session isolation
08
Input sanitization

High risk

Local backend + broad filesystem access + reusable credentials is the highest-risk combination. Avoid it by default.

Configuration caveats

• Local mode does not provide a container boundary.
• Approval mode can increase or reduce autonomy and risk.
• Forward only the environment variables a task requires.
• Multi-user messaging needs explicit authorization controls.

Practical use

Three bounded workflow patterns

These are implementation patterns, not measured performance claims. Test them with your own models, permissions, and evaluation set.

Research with a source ledger

Give the agent a bounded question, require URLs and retrieval dates, and human-review any conclusion that changes money, access, or production state.

Repository maintenance in isolation

Run the agent in a container or disposable branch, provide a narrow test command, and require a diff plus test evidence before merge.

Private messaging assistant

Allowlist users, begin with read-only tools, keep approval enabled, and add one integration at a time.

What we verified

Evidence this page can support

• Official installation commands and setup flow.
• Current feature wording, surfaces, and backend count.
• Eight documented security layers and backend caveats.
• Repository license, dated stars, and current release tag.

What remains unproved

Claims this page does not make

• A universal speed, quality, or token-efficiency advantage.
• That every security layer is active in every configuration.
• A fixed all-in operating cost for every user.
• That GitHub popularity predicts success on your workload.

Continue with the right depth

Related Hermes Agent paths

Quickstart →A focused first-run path.Skills guide →How reusable skills fit the workflow.Hermes vs OpenClaw →Compare operating models and fit.Best skill stacks →Design a bounded tool and skill set.Security guide →Review permissions and execution risk.Verify a skill →Check a skill before adding it.

Common questions

Hermes Agent FAQ

What is Hermes Agent?+

Hermes Agent is an open-source agent from Nous Research. It combines a terminal agent loop with persistent memory, reusable skills, built-in tools, MCP support, messaging integrations, and multiple execution backends.

Is Hermes Agent free?+

The repository is MIT-licensed. Running it can still involve model-provider, hosting, messaging, search, or other tool costs, so there is no universal all-in price.

Who maintains Hermes Agent?+

Nous Research maintains the official Hermes Agent repository and documentation.

Does Hermes Agent create skills?+

Yes. The official documentation describes a learning loop that can create and improve reusable skills. Its skills are compatible with the agentskills.io specification.

Does Hermes Agent support MCP?+

Yes. Hermes Agent supports Model Context Protocol servers in addition to its built-in tool set. MCP credential filtering is one of its documented security layers.

Can Hermes Agent run on Windows?+

Yes. The official documentation provides a native PowerShell installer for Windows. WSL2 is also included in the Unix-style installation path.

Is Hermes Agent safe to run?+

It includes eight documented security layers, but risk depends on configuration and execution backend. Local execution does not provide container isolation, approval can be disabled, and forwarded environment variables may expose secrets.

Does Hermes Agent need Docker?+

No. It can use a local backend, while Docker and other container or remote backends provide a stronger execution boundary. Choose the backend according to the trust level of the tasks and data involved.

Is Hermes Agent an IDE autocomplete tool?+

No. It is a broader autonomous agent environment. Developers who only want inline code completion may find a dedicated IDE assistant simpler.

Primary-source ledger

Editorial profile · no ranking guarantee · last checked 2026-07-26

Official docsSetup, tools, platforms, backends, skills, and MCP. ↗Security guideSecurity layers, approvals, authorization, and backend caveats. ↗GitHub repositoryLicense and dated popularity snapshot. ↗Current releaseHermes Agent v0.19.0 · tag v2026.7.20. ↗

Claims were rechecked against the primary sources above on July 26, 2026. Product behavior can change after publication.