How to use the OpenClaw Marketing Pages agent skill

What is an agent skill and what does this one do?

An agent skill is a folder containing a SKILL.md file with instructions that an AI agent can load when a request matches the skill's description. This one is named openclaw-marketing-pages and is licensed MIT.

When installed, the agent follows the skill's workflow for OpenClaw public pages. It first reads references/page-patterns.md, inspects the existing header, footer, and sections of the site, and reuses local patterns before adding new ones. It aims to show the real product, brand, place, or offer in the first viewport rather than a generic template.

• Reads the project's reference patterns before composing a page
• Reuses the existing header, footer, sections, and page primitives
• Uses the canonical tokens and typography contract
• Checks navigation, content hierarchy, media, light/dark themes, and responsive behaviour in a browser
• Keeps SEO metadata, canonical URLs, and semantic heading order intact

Check the requirements before you install

The skill is specific to the OpenClaw site. The rules and token contract assume that site's structure, so it is not useful for projects outside OpenClaw.

You need a local checkout of the OpenClaw site repository that includes its header, footer, and tokens. For the validation step you need a browser; the skill files do not name a specific tool. The skill does not run scripts and does not use the network.

• A local checkout of the OpenClaw site repository with its header, footer, and tokens
• A browser for the validation step

Install the skill from the ZIP download

Download the ZIP of the skill from the agent-skills directory page. Then install it in one of two ways.

For personal use across projects, unzip it into ~/.claude/skills/. To limit it to one project, unzip it into that project's .claude/skills/ folder. Either way the skill folder should contain SKILL.md and references/page-patterns.md.

In the Claude apps you can instead upload the ZIP under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/
• Project install: .claude/skills/ inside your project
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Trigger the skill with a natural request

You do not need a special command. Describe the public page you want, and mention OpenClaw so the request matches the skill's description. The skill's description covers websites, landing pages, content pages, ecosystem pages, and campaigns, and topics such as page structure, navigation, responsive layout, media, SEO presentation, and public-site visual polish.

Be concrete about the page type and the next action. The skill asks the agent to identify the next action for each decision area, so a clear goal produces a more useful page.

• "Create an OpenClaw ecosystem page for the community projects, with a section for each contributor project."
• "Update the OpenClaw landing page for the spring campaign so the product screenshot shows in the first viewport."
• "Check the navigation and mobile layout on the OpenClaw content page I changed."

Review what the agent should deliver

The skill sets clear expectations for the output. The first viewport should name the literal brand or product as the primary heading, with supporting copy that says what it is, who it helps, and the next action. Real product or community imagery is preferred over stock media.

Sections should be unframed full-width bands with constrained inner content. Cards are reserved for genuinely repeated items. Headings stay proportionate, and a hint of the next section should remain visible at common desktop and mobile heights.

Ask the agent to report its browser checks. The skill requires validation of navigation, hierarchy, media, light and dark themes, and responsive behaviour, and checks at least one narrow mobile viewport and one desktop viewport.

Safety notes: scripts and network

The skill does not run scripts and does not use the network, according to its listed requirements. It consists of instructions and reference documentation that guide how the agent edits your local OpenClaw site files.

Because the agent will edit files in your checkout, review its changes with normal version control before committing. The browser validation step is the one part that depends on a tool you provide.

• Runs scripts: no
• Uses network: no
• Review file edits in your repository before committing

Tips for better results

Point the agent at the page you want to change and the existing section it should reuse. The skill tells the agent to reuse local patterns before adding another composition, so naming a similar existing page helps it match the site.

If you want a quick generic template, this skill is the wrong tool, because it explicitly discourages generic landing-page templates and feature tours.

• Name an existing OpenClaw page to use as a reference
• Ask for the real product or community image rather than stock media
• Ask for the light and dark theme and mobile checks in the same request