Skip to content
Back to skills

Choose Hermes Agent Workflow

ASecurity

Route Hermes Agent work to the correct operator, extension-development, messaging-gateway, Nous services, or Socket compatibility workflow. Use when a request spans surfaces or its extension boundary is unclear.

  • 7 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 5, 2026
ai-agentspythongoshellapibackenddocumentation

Works with

  • terminal
  • cli
  • api
  • mcp

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned September 5, 2026

npx -y skills add gaelic-ghost/socket --skill choose-hermes-agent-workflow --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Choose Hermes Agent Workflow?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Choose Hermes Agent Workflow
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/gaelic-ghost-choose-hermes-agent-workflow/badge)](https://www.skillsdirectory.com/skills/gaelic-ghost-choose-hermes-agent-workflow)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: choose-hermes-agent-workflow
description: Route Hermes Agent work to the correct operator, extension-development, messaging-gateway, Nous services, or Socket compatibility workflow. Use when a request spans surfaces or its extension boundary is unclear.
metadata:
  hermes:
    category: agent-portability
    tags: [hermes, routing, extensions, gateway, nous]
---

# Choose Hermes Agent Workflow

Classify the requested outcome before opening broad documentation or changing configuration.

## Route the Request

| Requested outcome | Use |
| --- | --- |
| Install, update, configure, run, secure, troubleshoot, or use Hermes | `operate-hermes-agent` |
| Create skills, plugins, providers, tools, hooks, desktop/dashboard extensions, or programmatic integrations | `build-hermes-agent-extensions` |
| Run Hermes inside Zed, Xcode, or another ACP client | `operate-acp-agent-integration` plus the client-owning workflow |
| Let Hermes discover, call, or be called by an independently operated peer agent | `operate-a2a-agent-integration` plus gateway operation for inbound service lifecycle |
| Decide between ACP, MCP, TUI gateway, API, terminal, or native host integration | `choose-agent-integration-protocol` |
| Run a messaging gateway, API server, webhook intake, messaging platform, or long-lived remote agent | `operate-hermes-agent-gateway` |
| Use Nous Portal, Tool Gateway, Nous Chat, subscription proxy, or Hermes Cloud | `use-nous-research-services` |
| Export Socket skills, translate Socket MCP, or classify Codex-to-Hermes portability | `hermes-agent-compatibility` |

Use more than one workflow when the request genuinely crosses boundaries. Keep one workflow in charge of each write surface.

## Keep the Three Gateway Terms Separate

- `Hermes messaging gateway`: the long-running Hermes process that connects messaging platforms, webhooks, and the API server.
- `Nous Tool Gateway`: Nous-hosted backends for web, media, browser, and optional cloud terminal tools.
- `TUI gateway`: the JSON-RPC protocol used by Hermes TUI and custom host integrations.
- `ACP`: the editor-to-agent protocol used when Zed, Xcode, or another client launches and renders Hermes through `hermes acp`.
- `A2A`: the peer-agent protocol used when Hermes exchanges discovered messages
  or tasks with independently operated agents.

Never use an unqualified “gateway” in a decision or implementation note when more than one meaning is possible.
Do not describe ACP as a gateway or as a general agent-to-agent protocol.

## Select the Smallest Extension Surface

Prefer, in order:

1. A skill for instructions plus existing tools or commands.
2. MCP for an external tool server.
3. Config-driven backends or shell hooks when Hermes already defines that surface.
4. A standalone Python plugin for registered runtime behavior.
5. A specialized provider, platform, memory, context, browser, search, media, secret, desktop, or dashboard extension only when that exact subsystem owns the behavior.
6. A Hermes core contribution only when the behavior belongs in the upstream product rather than a user, project, or third-party extension.

## Source and State Rules

- Refresh the official Hermes docs for claims that can drift; the project moves quickly.
- Inspect the installed Hermes version and active profile before diagnosing local behavior.
- Keep `~/.hermes/` user state, project `.hermes/` state, standalone plugin source, and the upstream Hermes repository distinct.
- Do not mutate Portal accounts, gateway services, messaging accounts, cloud deployments, or user-home configuration unless the user requested that action.
- Do not describe a Codex or Claude plugin manifest as a Hermes plugin.

## Output Contract

Return:

1. `workflow`: the selected skill or combination.
2. `surface`: the concrete Hermes runtime, configuration, hosted service, or source tree involved.
3. `reason`: why that surface owns the request.
4. `state`: local, project, hosted, or upstream state that may change.
5. `verification`: the narrowest proof needed.

Read [references/workflow-map.md](references/workflow-map.md) when the request crosses multiple Hermes subsystems or uses an ambiguous “gateway,” “plugin,” “cloud,” or “provider” label.

Files in this skill

  • SKILL.md4.2 KB
  • agents/openai.yaml284 B
  • references/workflow-map.md2.9 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.