page tuner
Visual human-in-the-loop editor for polishing and restyling single-file HTML via MCP.
Free
Works with the AI Agents you already use
page tuner
Example session with this skill installed
Open /users/dev/projects/landing/index.html in the tuner. The spacing on the hero section feels off and I want to fix the colors myself.
- Read your context and instructions
- Compiled the page tuner
- Generated the UI component
I have started the visual editor for your page. You can now modify styles, reorder elements, or lasso specific areas to give me instructions.
Editor URL: http://localhost:8000/edit/users/dev/projects/landing/index.html
Editor Session Started
| Component | Status | Path |
|---|---|---|
| Target File | Active | /users/dev/projects/landing/index.html |
| Local Server | Running | http://localhost:8000 |
| Change Journal | Pending | index.html.changes.json |
| Selection Data | Pending | index.html.selection.json |
What you can do now:
- Click the link above to open the visual editor.
- Adjust the colors and spacing for the hero section directly in the UI.
- If you want me to add a new section or perform a complex change, lasso the area, write a note, and click
Save to Agent.
4. Return here and let me know when you are ready for me to sync the changes to your source code.
Next steps
- Open the local editor URL in your browser.
- Perform your manual style adjustments.
- Type "I'm done" or "Sync changes" here once you have saved your edits in the tuner.
index.html
TSX · React component
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop fighting with LLMs over minor UI tweaks and spacing issues. Page-tuner gives you a human-in-the-loop workflow for single-file HTML pages by launching a local visual editor where you can hand-tweak styles, and then it reads your intent to apply source-level code changes.
What it does
- Visual editing starts a local editor for restyling, reordering, and deleting elements by hand.
- Intent capture reads lassoed areas and user notes from the editor to understand complex requests.
- Source-level updates applies style, text, and layout changes directly to the HTML file without full regeneration.
- Smart sync maintains a two-way loop between your manual edits and agent-generated improvements.
- HTML cleaning strips internal tracking IDs and empty attributes once the session is finalized.
How it works
- Start the editor by providing an absolute path to a single-file static HTML page.
- Hand-tweak the UI in the browser to adjust colors, padding, or element order.
- Lasso problem areas and provide a brief note for the agent to handle complex additions or changes.
- Agent review pulls your hand-edit journal and selection data to update the source code.
- Finalize the session to clean the markup and shut down the local server.
Frameworks & tools
This skill requires Python 3.9+ and the page-tuner MCP server. It is designed specifically for single-file static HTML where styles, scripts, and images are inlined.
Why this beats prompting it yourself
Writing prompts to fix "a little more padding on the second card" is inefficient and often fails. This tool lets you fix what is easy by hand and delegate what is hard to the agent, ensuring the code matches exactly what you see on screen.
Use cases
- Refining generated landing pages to match a specific brand aesthetic.
- Restructuring pricing tables or checkout flows through visual drag-and-drop.
- Polishing typography and spacing in complex static marketing pages.
- Adding new sections to existing pages while maintaining class-name consistency.
Known limitations
Supports single-file static HTML only. It will not work with bundlers, React/Vue components, or multi-file projects. Requires local Python environment to run the MCP server.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Requires Python 3.9+ and page-tuner MCP server. Single-file static HTML only. No support for bundlers or frameworks like React/Next.js.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 16 days ago
- Free with an Agensi account
- Software
Python