← Back to Skills Marketplace
118
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install xdesign
Description
设计流程引擎，将想法到可交付物的全流程压缩为对话；当用户需要创建/设计/迭代视觉产物如幻灯片、原型、动画、UI设计、落地页或设计系统时使用
Usage Guidance
This skill appears to do what it says (produce HTML prototypes and integrate with repos/files), but it asks the agent to read broad context (entire codebases, cross-project files, GitHub repos) and its docs contain conflicting guidance about copying existing product UIs. Before installing or using it:
- Be prepared for OAuth prompts (GitHub) and file-access requests; limit scopes if possible.
- Avoid uploading proprietary or copyrighted materials unless you have the right to reproduce them; the skill's docs both prohibit and implicitly encourage copying popular UIs.
- If the agent will output standalone HTML, inspect generated artifacts for scripts that call external endpoints (e.g., window.claude.complete) before sharing or hosting them, since those calls run in viewers' browsers and could leak data or use their model quota.
- Ask the skill author to clarify the exact tools it will call at runtime and what permissions they require, and request a clear statement about how exported artifacts behave (network calls, external integrations).
If you want help drafting specific questions to ask the author (permissions needed, GitHub scopes, artifact runtime behavior), tell me and I can produce them.
Capability Analysis
Type: OpenClaw Skill
Name: xdesign
Version: 1.0.0
The 'xdesign' skill bundle is a comprehensive design engine for creating UI/UX prototypes, presentations, and landing pages. It utilizes a structured 'Perception-Planning-Action-Feedback' loop and includes advanced features like GitHub integration, cross-project file access, and automated verification agents (fork_verifier_agent). While the skill has broad capabilities (e.g., web fetching via web_fetch and file system access), these are clearly aligned with its stated purpose of gathering design context and generating artifacts. The instructions in SKILL.md include standard defensive guardrails to protect internal logic and respect intellectual property, with no evidence of malicious intent, data exfiltration, or unauthorized persistence.
Capability Tags
Capability Assessment
Purpose & Capability
The name/description (design flow / HTML prototypes) align with instructions to read files, produce HTML, and integrate with GitHub and export tools. Minor inconsistency: SKILL.md's guardrail says 'Never recreate copyrighted designs', but references/workflow-guide.md explicitly encourages providing screenshots of Stripe/Notion/Linear to get 'same-tier UI' — this conflicts with the stated copyright guardrail and could encourage IP reuse.
Instruction Scope
The instructions repeatedly tell the agent to 'gather ALL available context' and to read attached files, design system files, entire codebases, cross-project files, and GitHub repositories. That is plausible for a design-to-code skill, but it is broad (cross-project reads, GitHub imports, run_script + readFileBinary for Office files). Also technical-specs shows generated HTML may call window.claude.complete from the viewer without an API key — exported artifacts executing remote model calls in a user's browser increases the surface for unintended data exposure. These behaviors are within a design skill's domain but are high‑impact and deserve explicit user consent and scope limitation.
Install Mechanism
Instruction-only skill with no install spec and no code files — lowest installation risk. Nothing is downloaded or written at install time according to metadata.
Credentials
The skill requests no environment variables or credentials in metadata. However, the runtime instructions assume authorization flows (connect_github, github_import_files) and use a variety of platform tools (gen_pptx, read_pdf, register_assets, etc.). These are proportional to a design skill but will require elevated file and GitHub access at runtime; users should expect OAuth/permission prompts.
Persistence & Privilege
always:false and normal autonomous invocation settings. The skill does reference forking a verifier agent and using platform tools, but it does not request permanent 'always' inclusion or claim system-wide configuration changes.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install xdesign - After installation, invoke the skill by name or use
/xdesign - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
x-design 1.0.0
- Initial release of the x-design skill: an end-to-end design workflow engine that turns ideas into deliverables via conversation.
- Supports creating, designing, and iterating on visual artifacts (slides, prototypes, UI, landing pages, design systems) using HTML.
- Implements a structured PPAF (Perception → Planning → Action → Feedback) design loop for iterative design.
- Emphasizes extracting structured requirements, building from existing context, and always beginning with a design system and wireframe phase.
- Includes robust safety and quality guardrails, including design reasoning explanations, strict IP protection, and a detailed self-check checklist before output.
Metadata
Frequently Asked Questions
What is xdesign?
设计流程引擎，将想法到可交付物的全流程压缩为对话；当用户需要创建/设计/迭代视觉产物如幻灯片、原型、动画、UI设计、落地页或设计系统时使用. It is an AI Agent Skill for Claude Code / OpenClaw, with 118 downloads so far.
How do I install xdesign?
Run "/install xdesign" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is xdesign free?
Yes, xdesign is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does xdesign support?
xdesign is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created xdesign?
It is built and maintained by qomob (@qomob); the current version is v1.0.0.
More Skills