AGENT SKILL

scraping-html-to-markdown

Use when the user wants a single page rendered as clean Markdown plus structured metadata. Covers `kreuzcrawl scrape <url>`, JSON vs Markdown output, what metadata is returned, and how to handle JS-heavy pages.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 1.6k
Forks
491

01 前言 Front Matter

name
scraping-html-to-markdown
description
Use when the user wants a single page rendered as clean Markdown plus structured metadata. Covers `kreuzcrawl scrape <url>`, JSON vs Markdown output, what metadata is returned, and how to handle JS-heavy pages.

03 SKILL.md 正文

Scraping HTML to Markdown

kreuzcrawl scrape <url> is the right tool when the user has a single page in mind. It returns Markdown plus a full structured payload (metadata, links, images, JSON-LD, HTTP response info).

Quick recipe

kreuzcrawl scrape https://example.com/article --format markdown

JSON form (default) when downstream needs metadata:

kreuzcrawl scrape https://example.com/article --format json

Flag surface

Flag Default Purpose
--format json json or markdown.
--timeout 30000 Per-request timeout in ms.
--proxy — HTTP, HTTPS, or SOCKS5 proxy URL.
--user-agent — Override request UA.
--respect-robots-txt off Honour robots.txt.
--browser-mode auto auto, always, never — see headless-fallback skill.
--browser-endpoint — External CDP ws:// URL.
--config — Inline JSON or @file.json for full CrawlConfig.

Output shape

Markdown mode

Prints the rendered Markdown only. Use when piping to a file the user will read, or when the result becomes LLM context downstream.

JSON mode

Top-level PageResult with:

  • url, final_url (after redirects), status_code.
  • markdown: { content, fit_content, warnings } — fit_content is a pruned LLM-optimised variant.
  • metadata: Open Graph, Twitter Card, Dublin Core, article tags, JSON-LD, headings (H1–H6), feeds, favicons, hreflang.
  • links: arrays for Internal, External, Anchor, and Document.
  • images: <img>, <picture>, srcset, og:image.
  • tables: structured table data preserved separately from Markdown.
  • response: HTTP headers, content type, charset, body size.

Read result.markdown.content for the Markdown string when scripting.

Common pitfalls

Empty or stub content

Static fetch returned a JS shell. Symptoms in JSON output:

  • markdown.content is short or only contains nav/footer chrome.
  • markdown.warnings mentions JS-render-required.
  • metadata.headings is empty when the page clearly has headings.

Re-run with --browser-mode always and see the headless-fallback skill.

WAF block

Auto mode detects 8 WAF vendors and retries through headless Chrome automatically. If you forced --browser-mode never, the WAF response will fall through. Check response.status_code — 403/406/503 with WAF headers (server: cloudflare, x-amz-cf-id, etc.) is the giveaway.

Robots.txt blocking the fetch

If --respect-robots-txt is set and the path is disallowed, the scrape returns an error rather than partial content. Drop the flag only on hosts you own or have authorisation for.

Wrong charset

Most pages declare UTF-8. Pages that lie about their charset surface as mojibake in markdown.content. Override via --config '{"force_encoding":"latin-1"}' or similar.

Examples

Scrape an article for downstream LLM context

kreuzcrawl scrape https://blog.example.com/post-123 --format markdown \
  > /tmp/article.md

Scrape with proxy and custom UA

kreuzcrawl scrape https://example.com \
  --proxy http://proxy.internal:3128 \
  --user-agent "kreuzcrawl (research@example.com)" \
  --format json

Extract just the OG metadata

kreuzcrawl scrape https://example.com --format json \
  | jq '.metadata | {title: .og.title, description: .og.description, image: .og.image}'

When to reach for crawl or interact instead

  • The user wants the whole site, not one page → crawling-a-site skill.
  • The user needs to click, type, or scroll before extracting → use kreuzcrawl interact with the action list.
  • The user only wants the list of URLs → kreuzcrawl map.

04 同仓库更多 Skills

hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-review-with-annotations

Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-find-and-synthesize-context

Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-update-docs

Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-manage-records

Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-manage-interactive-html

Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-collaborate-threads

Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.

★ 1.6k