web-page-workflow

Ordinary Web Page Processing Specification (Simplified)

This skill adopts progressive disclosure: first determine the user's needs, then open only the relevant documents.

Quick Triage

• Overall Asset Extraction Process (Orchestration Entry):skills/web-page-workflow/asset-extraction.md
• Theme Generation Rules:skills/web-page-workflow/theme-generation.md
• Document Generation Rules:skills/web-page-workflow/doc-generation.md
• Data Generation Rules:skills/web-page-workflow/data-generation.md
• Prototype Restoration/Optimization/Visual Replication:skills/web-page-workflow/prototype-restoration.md

General Prerequisites (required for any scenario)

1. Use extract-page-data skill (located in skills/extract-page-data/) to obtain screenshots, themes, links, and other data:

   node skills/extract-page-data/scripts/extract.mjs <URL> --all
   

   If high-precision cloning is needed, use clone-page skill (located in skills/clone-page/）。
2. Obtain web page link: if the user has not provided one, ask for the URL.

General Constraints

• Use scripts in Axhub Skills by default; use Firecrawl only when scripts fail, link discovery is insufficient, or the user explicitly requests it.
• If scraping is incomplete, you may add a wait parameter (--wait / Firecrawl waitFor）。
• Promptly inform the user of the tool currently in use, and ask for user confirmation first when uncertain.

Reference Resources

• theme-guide.md
• development-guide.md