$ loading_
Review agent-built HTML with anchored comments and push fixes.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "Terrarium" MCP server from askskill: Run: claude mcp add --transport http 'io-github-nelsonsantryhkd-terrarium' 'https://www.tryterrarium.com/api/mcp'
Use Terrarium to publish this agent-built HTML page for human visual review.
A publishable artifact is created for the team to inspect the page visually.
Pull anchored comments from this HTML artifact and summarize the issues to fix.
Anchored comments are returned so each issue can be fixed precisely.
Apply the review feedback to this HTML and push the fixed version back.
A fixed HTML version is produced and pushed back for iteration.
Front-end developers or designers can publish an agent-generated HTML page and gather human visual feedback. It fits cases where layout, copy, or visual details need quick review.
When anchored comments already exist, you can pull the feedback and push fixes accordingly. It works well for collaborative iteration on HTML artifacts.
It supports human visual review for agent-built HTML: publish artifacts, pull anchored comments, and push fixes.
The core flow is: publish the HTML artifact, collect anchored comments, then fix and push the changes back. Only these three steps are明确ed in the provided material.
The provided info only says it targets human review for HTML artifacts; installation, key, or runtime requirements are not given. See the source repository for details.
Build complex multi-component web artifacts with React, Tailwind, and shadcn/ui.
Provides structured aquarium and habitat data for AI analysis and app development.
Create, edit, preview, and export multi-format HTML content with a local AI agent.
Turn AI images and videos into playable game-ready art assets.
Search curated generative art algorithms by style, intent, and code snippets.
Let AI agents get structured judgments from real humans.