Web Clipper

Save any web page as a clean Markdown note with metadata, organized in Obsidian format, using Jina Reader API without browser extensions.

Security Scan

The skill's code, instructions, and requests are consistent with a web-to-Markdown clipper that uses the Jina Reader API and writes clippings to a local Obsidian-style folder; nothing requested or installed is disproportionate to that purpose.

Purpose & Capability

Name/description (Obsidian Markdown clipping via Jina Reader) match the included script and SKILL.md: the script fetches a URL via Jina, converts to Markdown, and writes a file into a local vault. No unrelated credentials or system-wide services are requested.

Instruction Scope

SKILL.md directs execution of the bundled Python script with a URL and optional flags. The script only fetches remote http(s) URLs, converts content, and writes files into a user-scoped vault path. It does not read arbitrary system files, other skill configs, or send data to unexpected endpoints (it posts to r.jina.ai as documented).

Install Mechanism

This is instruction-only with a small bundled script; there is no install/download of third-party archives. The only runtime dependency is the widely used 'requests' Python package (documented).

Credentials

Registry metadata lists no required env vars, but SKILL.md and the script optionally use JINA_API_KEY and OPENCLAW_VAULT loaded from ~/.openclaw/.env. The script explicitly limits loaded keys to JINA_API_KEY and OPENCLAW_VAULT, which is proportionate, but the registry could explicitly declare the optional JINA_API_KEY for clarity.

Persistence & Privilege

Skill is not marked always:true and does not modify global agent settings. It writes clippings into a user-scoped directory (default ~/.openclaw/obsidian-cache) and only reads a small, documented ~/.openclaw/.env file for two allowed keys.

Assessment

This skill appears to do what it says: it will fetch remote http(s) pages you ask it to and save Markdown files under a directory in your home (~/.openclaw/obsidian-cache by default). Before installing, consider: (1) review the bundled script (already included) to confirm the save path is acceptable; (2) if you provide a JINA_API_KEY, store it in ~/.openclaw/.env and protect that file's permissions—the script only loads JINA_API_KEY and OPENCLAW_VAULT by design; (3) be aware that invoking the skill causes the agent to make outbound requests to r.jina.ai for each URL you save, so only ask it to fetch URLs you trust; (4) ensure the environment has Python and the 'requests' package; and (5) if you allow autonomous agent actions, remember the agent can execute skills you enable—this skill will fetch and save any URL the agent is instructed to clip, which is expected behavior but worth keeping in mind.

Save any web page as a clean, readable Markdown note in your Obsidian vault. Powered by Jina Reader API for content extraction — no browser extension required. Works from inside Claude.

When to Use This Skill

• You want to save an article for later reading in Obsidian
• You found a useful page and want it archived as Markdown
• You want to add tags and organize clips into subfolders
• You need to clip technical docs, blog posts, or news articles
• You want to avoid browser extension bloat — just tell Claude to save it

What This Skill Does

1. Fetches the page via Jina Reader API — strips ads, nav, and noise
2. Converts to Markdown — clean readable format with frontmatter
3. Saves to your vault — default: ~/.openclaw/obsidian-cache/clippings/
4. Adds metadata — title, source URL, date saved, and your tags
5. Confirms the filename — Claude tells you exactly where it was saved

Requirements

• Python 3.7+
• requests library: pip install requests
• JINA_API_KEY — optional. Works without a key (free tier). Set for higher rate limits:

echo 'JINA_API_KEY=jina_xxxxxxxxxxxx' >> ~/.openclaw/.env

Get a free key at jina.ai

Note: The script only accepts http:// and https:// URLs. Local file paths are not supported.

How to Use

Basic — Save a URL

save https://example.com/article

保存这个 https://v2ex.com/t/123456

With Tags

save https://example.com/article --tags "ai,research"

Save to a Subfolder

save https://example.com/article --folder clippings/tech

Full Options

python3 ~/.openclaw/skills/web-clipper/scripts/save_web_page.py \
  --url "https://example.com/article" \
  --folder "clippings/tech" \
  --tags "ai,tools" \
  --vault "/path/to/your/obsidian/vault"

Example

User: save https://www.v2ex.com/t/1197958

Claude executes:

python3 ~/.openclaw/skills/web-clipper/scripts/save_web_page.py --url "https://www.v2ex.com/t/1197958"

Output file (~/.openclaw/obsidian-cache/clippings/2026-01-15 V2EX Thread Title.md):

---
title: "V2EX Thread Title"
source: "https://www.v2ex.com/t/1197958"
saved: "2026-01-15 14:32"
tags: []
---

# V2EX Thread Title

> Source: https://www.v2ex.com/t/1197958
> Saved: 2026-01-15 14:32

---

[Clean article content in Markdown...]

Claude confirms: "已保存为 2026-01-15 V2EX Thread Title.md ✅"

Supported Content

Content Type	Support
Blog posts / articles	✅
Technical documentation	✅
News articles	✅
V2EX / Reddit threads	✅
GitHub READMEs	✅
PDF files (remote)	✅
JS-heavy SPAs	⚠️ Jina handles most; complex apps may need PinchTab
Login-required pages	❌

File Structure

web-clipper/
├── SKILL.md                    ← This file (Claude reads this)
└── scripts/
    └── save_web_page.py        ← Main clipping script

Output Format

Saved files include YAML frontmatter for Obsidian compatibility:

---
title: "Article Title"
source: "https://original-url.com"
saved: "2026-01-15 14:32"
tags: [ai, research]
---

Notes

• Filenames are auto-sanitized and dated: YYYY-MM-DD Title.md
• Duplicate filenames get a counter suffix: 2026-01-15 Title (2).md
• Default vault: ~/.openclaw/obsidian-cache/clippings/
• The JINA_API_KEY free tier supports ~1000 requests/month