Archived ClawShop Setup: Hermes Agent and OpenClaw

What "framework setup" means

Status note (updated August 25, 2026): ClawShop has been consolidated and its standalone setup offer is no longer sold. This article preserves proposed scope rather than advertising a current service or support commitment. Exact framework support, third-party licensing, and deployment work require a separate current engagement.

Framework setup is the work between "we have a model" and "the assistant or skill is usable."

That work is not glamorous, but it is where agent projects usually stall: environment setup, secrets handling, tool wiring, schedules, prompts, validation, logs, and handoff documentation.

The former ClawShop framework-setup lane targeted buyers who did not need a new fine-tune. Today, open framework documentation is the starting point; managed business operations route to Icarus, and model behavior work routes to Logos.

Hermes Agent setup

The proposal treated Hermes Agent as a candidate runtime for assistant-shaped workflows involving tools, reusable skills, memory, scheduled automation, and messaging gateways. Hermes Agent is a software runtime; it should not be conflated with the separate Nous Research Hermes model family.

A proposed Hermes Agent setup could have included:

Chat templates and system instructions matched to the skill workflow
Tool schemas and function-call validation
Routing for tools or documented messaging channels selected during setup
Memory and retrieval configuration when the agent needs project context
Escalation behavior for low-confidence or high-risk requests
A runbook so the team knows how to restart, update, and audit the setup

Using a Hermes-family model was not a requirement for using Hermes Agent, and installing Hermes Agent would not itself fine-tune a model.

OpenClaw setup

OpenClaw describes itself as a personal AI assistant that runs on a user's devices, with its Gateway acting as the control plane for channels, workspace access, and skills. It is not a fine-tuning or model-training stack. The former ClawShop proposal considered setup help around that assistant runtime.

That proposed setup path was aimed at teams that:

The buyer wants an open stack rather than a hosted chatbot.
The workflow needs local or private model support.
The team wants scripts, configs, and handoff docs they can keep.
The team might separately commission model work through Logos, without implying that OpenClaw performs that training.

NemoClaw and IronClaw references

NemoClaw- and IronClaw-branded fork services were planning ideas, not delivered ClawShop products. DDG's internal scope notes explicitly deferred branded forks because they would add an ongoing upstream-maintenance burden without demonstrated customer demand. This archive therefore makes no claim that DDG shipped, supported, or hardened either service.

Sensitive, regulated, contractually restricted, or air-gapped work requires a separately documented security and legal scope. A product or framework name does not establish compliance.

What about ClawHub?

Some buyers searching for "ClawHub" are looking for an OpenClaw/Hermes skill registry, marketplace, or agent framework setup path. DDG does not sell a product named ClawHub and no longer sells ClawShop as a packaged setup service.

That distinction is intentional. ClawHub is an ecosystem-adjacent search term; ClawShop is an archived DDG offer.

For technical users, the open skill ecosystem is the discovery path. The Hermes skills documentation describes the reusable-skill surface and points to its current discovery resources. DDG's managed path begins at Icarus when those skills need customer channels, approvals, reporting, and recurring operations.

The Skills Hub links above are the current discovery target; this article does not promise future ClawShop listings.

When Logos is still needed

Framework setup does not magically teach a model a domain.

Use Logos when the model needs durable new behavior:

Brand voice from approved examples
Domain-specific responses
Better classification or routing behavior
Tool-call examples that match your schema
Policy answers grounded in internal documents

Use framework setup only when the model behavior is already good enough and the missing piece is deployment.

What the former setup scope intended to deliver

A framework setup engagement, had the archived scope been offered, would have aimed to leave the buyer with:

A selected framework installed and pinned
Configuration matched to the deployment target
Tool wiring for the first workflow
Secrets/access model documented
A repeatable runbook
Handoff notes for the team that will operate it

That is the boring part of agents. It is also the part that turns a model demo into a useful skill.