
Read URL or PDF
Fetch a web page or PDF and return a source-grounded summary, quotes, or clean Markdown.
- Repository
- tw93/Waza
- Stars
- 7,493
- Licence
- MIT
- Files
- 7 files
What it does
When you give the agent a web link or PDF, this skill tells it how to retrieve the page and extract its readable text, then answer your request from that content. Plain requests get a short summary with the source; requests to save or convert get clean Markdown. It also explains failures such as paywalls or login pages instead of guessing. It can optionally use a DeepSeek search when the built-in web search is weak, and it routes Feishu/Lark, WeChat, and GitHub links to specialised methods.
Great for
- People who want a quick grounded summary of an article or report link
- Researchers who need quotes or citations from a web page or PDF
- Users who want a web page saved as Markdown for notes
Skip it if
- Authenticated, internal, or sensitive URLs: the skill's proxy tiers send URLs to third-party services, and the skill itself forbids that for such links
- Local text files already in the repo, which the skill says it should not handle
- Paywalled or JavaScript-heavy pages where extraction fails; the skill stops and warns rather than fabricating content
Try asking
- “看这个链接 https://example.com/article”
- “Save this PDF as Markdown in my notes folder: https://example.com/report.pdf”
- “Quote the paragraph about pricing from this page with its source”
- “Search with DeepSeek for recent Chinese sources on this topic”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 7 original files, the MIT licence and an attribution note.
read.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip read.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip read.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Use plain fetching for public pages; only add --use-proxy for JS-heavy or paywalled pages you are comfortable sharing with third parties
- Install readability-lxml and html2text for noticeably cleaner output
- Read the stderr fetch-tier line when a fetch fails; it names the tier and reason
What's inside
- references/read-methods.md
- scripts/deepseek.py
- scripts/fetch_feishu.py
- scripts/fetch_local.py
- scripts/fetch_weixin.py
- scripts/fetch.sh
- SKILL.md
Redistributed unmodified from tw93/Waza under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.