From panews
Fetches public PANews web pages (homepage, articles, columns) as Markdown with YAML metadata via HTTP GET. Supports zh, zh-hant, en, ja, ko locales for single-page reads.
npx claudepluginhub panewslab/skills --plugin panewsThis skill uses the workspace's default tool permissions.
This skill is for reading rendered PANews web pages as Markdown, including the homepage, article pages, and column pages. Use it when the user wants page content in a Markdown-friendly format with page metadata, rather than structured API fields or filtered search results.
Routes website URLs through markdown.new for clean Markdown views of docs, blogs, changelogs, and GitHub issues to extract notes or summarize.
Fetches any URL as clean Markdown via proxy cascade (r.jina.ai, defuddle.md, agent-fetch). Handles WeChat articles with Playwright and Feishu/Lark docs with API credentials.
Converts webpage URLs to Markdown files using standard, AI-optimized (with YAML frontmatter, summaries, key points), or dual modes. Saves locally for archiving web documents or AI context processing.
Share bugs, ideas, or general feedback.
This skill is for reading rendered PANews web pages as Markdown, including the homepage, article pages, and column pages. Use it when the user wants page content in a Markdown-friendly format with page metadata, rather than structured API fields or filtered search results.
It is best suited for single-page retrieval from PANews URLs. The skill should preserve the rendered page structure and metadata, and it should not be used for broad crawling, creator workflows, or API-style data discovery.
Read www.panewslab.com pages as Markdown via Accept: text/markdown. Responses include a YAML frontmatter block with page metadata (title, description, image).
Runtime behavior: this skill performs direct HTTP GET requests to https://www.panewslab.com with Accept: text/markdown and returns the rendered Markdown response. It does not require local scripts, creator credentials, or broad site crawling.
| Locale | Prefix |
|---|---|
| Simplified Chinese | /zh |
| Traditional Chinese | /zh-hant |
| English | /en |
| Japanese | /ja |
| Korean | /ko |
PANews Web Progress:
- [ ] Step 1: Confirm this is a website-page task, not an API task
- [ ] Step 2: Choose the locale prefix
- [ ] Step 3: Fetch with Accept: text/markdown
- [ ] Step 4: Preserve frontmatter metadata in the response
1. Start from https://www.panewslab.com
2. Build a page URL using one of the supported locale prefixes:
/zh, /zh-hant, /en, /ja, /ko
3. If the caller provides a PANews path without a locale prefix, prepend the prefix from `--lang`, or default to `/zh`
4. Send an HTTP GET request with:
Accept: text/markdown
5. Return the Markdown body as-is, including YAML frontmatter metadata
6. If the response is 404, report the page as unavailable
Accept: text/markdownpanews if the user asks for structured search or filterable API datahttps://www.panewslab.com/en
https://www.panewslab.com/en/ARTICLE_ID
https://www.panewslab.com/zh-hant/columns/COLUMN_ID
404, report it as unavailable rather than trying to synthesize content from API endpoints--lang or default to zh