ingest-web

What problem does it solve?

Web articles and blog posts often contain cluttered HTML, missing metadata, remote image links, or are difficult to import cleanly into a markdown-based wiki; this Skill automates extracting readable content and packaging it for ingestion into a vault.

Core Features & Use Cases

• Readable extraction: Fetches a URL and extracts the article title, author, published date, and main body while stripping navigation, sidebars, and ads.
• HTML-to-markdown conversion: Preserves headings, lists, blockquotes, links, code blocks, and image references when converting to clean markdown.
• Image and asset handling: Downloads referenced images into vault/raw/assets, replaces remote URLs with local paths, and records images-downloaded in the file frontmatter.
• Metadata-first output: Writes a YAML frontmatter including source-url, title, author, date-fetched, and images-downloaded, saving results to raw/<descriptive-slug>.md for downstream wiki workflows.
• Fallback guidance: Notes when extraction is likely to fail (heavy JS/SPAs) and recommends using a browser clipper for better fidelity.

Quick Start

Ingest the article at https://example.com/article into vault my-research to extract content, download images, and save a markdown file in raw/.