- Home
- Skills
- Frontend & Web Apps
- LanhuFlow — Lanhu Design & PRD Context
Works with the AI Agents you already use
LanhuFlow — Lanhu Design & PRD Context
Turn Lanhu (蓝湖) design and PRD links into focused development context: tokens, layout, layers, asset metadata and product requirements, using your connected LanhuFlow MCP server.
Free
LanhuFlow — Lanhu Design & PRD Context
Example session with this skill installed
Illustrative example with fictional project data: Use the Lanhu project link I provide, select the Checkout artboard, and give me the design tokens, layout and component context for implementing it in our existing Vue app. Keep the response focused and identify requirements that still need a PRD.
- Read your context and instructions
- Compiled the lanhuflow
Illustrative handoff: Checkout
Example only. All artboard names, IDs, dimensions and values below are fictional. This shows the handoff format; it is not a recording of a live Lanhu session or a promise that every project returns these fields.
Selected scope
| Item | Example value |
|---|---|
| Artboard | Checkout |
| Design identifier | fictional-checkout-id |
| Layout coordinates | 375 × 812 |
| Requested outputs | tokens, layout, layers |
| Layer depth | 4 |
Example design context
| Property | Illustrative value | Implementation use |
|---|---|---|
| Background | #F8FAFC | Map to the project's surface token |
| Primary action | #2563EB | Reuse the existing primary-button token |
| Main text | 16px / 24px | Map to the body typography scale |
| Card radius | 12px | Compare with the current Card primitive |
| Outer spacing | 16px | Apply within the existing page container |
Suggested component boundaries: CheckoutPage, OrderSummary and PaymentMethod. These boundaries are implementation choices derived from the hypothetical layout, not component APIs supplied by Lanhu.
Requirements to confirm
The selected design alone does not establish payment endpoints, validation rules, empty/error states or mobile breakpoints. If those matter to the requested implementation, select the relevant PRD pages and keep any remaining behavior as an explicit assumption.
Handoff
Use the actual returned design dimensions and token values in the existing Vue project. Preserve available asset identifiers/download URLs, verify a file exists before reporting it saved, and list any failed outputs. Report a rendered comparison only after it has actually run.
Connects securely to your tools. The creator never sees your data.
About this skill
Design and product context from Lanhu
A free companion skill for the open-source LanhuFlow MCP server. Give your agent a Lanhu (蓝湖) project, artboard, share link or product-document link, and it follows the right retrieval workflow for your task.
What it helps your agent do
- List artboards first, then select the relevant design by name, UUID or index.
- Read design tokens, layout summaries and bounded layer trees without loading every image into context.
- Request generated HTML, a preview image or original asset metadata when the task needs it.
- Resolve share/invite links and read selected PRD or Axure pages.
- Distinguish design coordinates from preview-image pixels, handle duplicate names and report partial extraction accurately.
- Use the retrieved facts for implementation planning, frontend work or a design review in your existing project.
How it works
- Connect the separate LanhuFlow MCP server to your agent and configure your Lanhu login locally.
- Share a Lanhu link and describe the artboard, page or development task you need.
- The skill guides the agent to retrieve a scoped set of evidence and return a usable handoff, including missing information and assumptions.
Requirements
A local MCP client with stdio support, Node.js 20+, the lanhu-flow-mcp npm package, and a Lanhu account with access to the selected content. Loading this skill through Agensi and connecting the LanhuFlow server are separate steps. Setup guidance is included.
Your Cookie stays in your local server configuration. PRD analysis downloads document HTML/resources locally; asset-file downloads and frontend edits depend on your requested task and the host's tools.
Scope and limitations
This is a workflow skill, not a bundled MCP server or hosted Lanhu account. Extraction depends on the available source data and permissions. Generated HTML is a reference for implementation; complete extraction and pixel-perfect results are not guaranteed. It does not edit the source Lanhu design.
The package contains SKILL.md, focused setup/tool references and an MIT licence. It is maintained against LanhuFlow MCP 1.0.0. This is an independent third-party project, not an official Lanhu product.
Source and setup: https://github.com/bangjier/lanhu-flow
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Frontend & Web Apps.
Requires a separate local LanhuFlow MCP connection and valid Lanhu access. This skill does not include the server or credentials. Extraction may be partial; generated HTML is an implementation reference. PRD analysis downloads resources locally. Visual parity and end-to-end compatibility across every agent have not been independently tested.
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 19 days ago
- Free with an Agensi account
- Accounts
LanhuappLanhu account
- Software
- LanhuFlow MCP
Node.js 20+npx
Use lanhu-flow-mcp 1.0.0 or a schema-compatible version in an MCP client with local stdio support. The Lanhu account needs project access and LANHU_COOKIE set locally.
Permissions
Allowed Hosts
File Scopes
The package contains instructions and references, with no executable scripts. Network access and LANHU_COOKIE are used by the separately configured local MCP server to access Lanhu. The agent does not need to read or print the credential. PRD analysis downloads HTML/resources under LANHU_DATA_DIR (default data/ts). Terminal access is for requested setup; project file reads, asset downloads and code edits depend on the user's task.