PRD to Prototype
v1.0.0从产品需求到可交互原型的完整工作流。当用户表达产品想法（如'我想做一个...'、'帮我设计...'）时触发。支持：1) 零提问直出PRD 2) 平台选择确认 3) 生成Awwwards级别的高保真HTML/Tailwind原型（移动端/PC端）。端到端产品设计流程。
⭐ 0· 109·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (PRD → prototype) match the instructions: it writes a PRD to /workspace/docs/prd.md, then generates HTML/Tailwind files in /workspace/prototype/. It does not request unrelated binaries, credentials, or config paths.
Instruction Scope
Instructions are explicit and constrained to the /workspace directory and to generating PRD + prototype. However, the skill enforces a strict 'no clarification questions' policy and mandates the exact GenUI confirmation form, which is a behavioral design choice (not a security mismatch) but can lead to incorrect assumptions or low-quality PRDs because the agent is forbidden from clarifying ambiguous user input. The instructions also require fetching external assets (Unsplash, Tailwind CDN, FontAwesome CDN) when rendering previews.
Install Mechanism
No install spec and no code files — instruction-only skill. No downloads or archive extraction to disk; lowest-risk install posture.
Credentials
Skill requires no environment variables, credentials, or special config paths (proportionate). It does reference external CDNs and Unsplash image URLs for runtime previews; this implies network access and use of third-party resources, which can have privacy/referrer implications but are reasonable for a web prototype generator.
Persistence & Privilege
Does not request always:true, does not modify other skills or system settings. It writes to /workspace (expected for a code/artifact generator) and may call a deploy tool if available (this is conditional and not specified as privileged).
Assessment
This skill appears to be what it says: it will create a PRD at /workspace/docs/prd.md and HTML/Tailwind files under /workspace/prototype/. Before enabling or running it: 1) Be aware it will not ask any clarifying questions — review the generated PRD carefully before asking it to build the prototype to avoid incorrect assumptions. 2) The preview pages load external resources (Tailwind, FontAwesome, Unsplash); if you care about privacy or leakage of prototype content, run it in a sandboxed workspace or replace remote image/CDN references with local assets. 3) Because the skill can write files and (optionally) call deployment tools, verify generated code and any deploy action before publishing. If you want tighter control, require the user to explicitly confirm the PRD (rather than auto-follow the 'no questions' rule) or disable autonomous invocation until you have reviewed outputs.Like a lobster shell, security has layers — review code before you run it.
latestvk971mdwqm7h5yd2stgfp3vqfsn83nhw2
License
MIT-0
Free to use, modify, and redistribute. No attribution required.