ClaudeSuperPower

scrape-analyze-page

Skill

Extract structured data (all available fields with values) from a page saved locally as an HTML file, optionally following a schema. Use this skill only to process already downloaded files. Do not invoke when the user provides a URL. When invoking, pass the user's full request verbatim as args — do

Install

git clone https://github.com/zytedata/claude-skills.git ~/.claude/skills/scrape-analyze-page

What is scrape-analyze-page?

Extract structured data (all available fields with values) from a page saved locally as an HTML file, optionally following a schema. Use this skill only to process already downloaded files. Do not invoke when the user provides a URL. When invoking, pass the user's full request verbatim as args — do not pre-parse file paths and don't rephrase it.

What this can do

Capabilities declared in this component's own frontmatter — not inferred.

Run shell commands

Declares Bash

Create and modify files

Declares Write or Edit

~87 tokens of context used while enabled, before you invoke anything

All declared tools (4)
BashReadSkillWrite

Documentation

README · ~4 min read

You are extracting structured data from a page. Given saved HTML, identify all available fields and extract their values.

Read ${CLAUDE_SKILL_DIR}/../scrape/references/python-environments.md.

Input

This is the user prompt: $ARGUMENTS. You need to extract the following information from it:

  1. Path to the saved HTML file, e.g. product1.html. This is what you need to analyze. Don't proceed if it's not provided.
  2. Path to output file, e.g. product1.json. When provided, this is where you will save the structured analysis.
  3. Path to data-type spec.json. When provided, guides extraction using schema field names, descriptions, and examples.
  4. Whether to strictly extract only the fields listed in the schema, if the schema was provided. When asked for strict extraction, extract only schema fields — no extras.
  5. --list-mode: when present, the page is a listing page. Extract ALL repeated item instances (e.g., every product card) as an array rather than a single-item object.

Process

Reviews

Log in to leave a review.

No reviews yet — be the first.

Explore related

Other things in this space — across every part of the ecosystem, not just skills.