Claude Code Skill: Scrapling
A Claude Code skill that picks a Scrapling Fetcher per site, handles cookie logins and Cloudflare, and writes scraping scripts.
- 478 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Bypassing Cloudflare and WAF protection can breach site terms and data access law.
- The skill handles cookies and login sessions: a leaked cookie file gives access to an account.
- Third-party page content enters the agent context and can carry prompt injection.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/claude-code-skill-scraplingDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Bypassing Cloudflare and WAF protection can breach site terms and data access law.
- The skill handles cookies and login sessions: a leaked cookie file gives access to an account.
- Third-party page content enters the agent context and can carry prompt injection.
This is third-party code. Review the repository files before installing.
What it does
The skill tells an agent which Scrapling class to use: `Fetcher`, `StealthyFetcher`, `DynamicFetcher`, `FetcherSession` or `Selector`. For simple extraction the agent first tries the `scrapling extract` CLI with `--ai-targeted`, and handles harder cases with a Python script from a template. The repository holds a decision tree, templates for static pages, form login, parsing ready HTML and Cloudflare handling through `StealthyFetcher` on Camoufox, a cookie vault template, a site pattern library and an error guide. `references/security.md` sets the guardrails: authorization, prompt injection and redaction.
Who it is for. Developers and analysts who need data from a site, including pages behind protection or a login, without learning the Scrapling classes.
Good fit when
- A page serves content through JavaScript and a plain request returns empty HTML.
- The data sits behind a login form and the cookie session must persist across pages.
- A site answers with a Cloudflare block and needs StealthyFetcher.
- You have saved HTML and want links or fields out of it without network access.
Not a fit when
- Site terms or law forbid automated data collection.
- The site has an official API that solves the task more simply.
- Cookies and tokens cannot be kept in local files.
Example request
Collect headings and prices from the store catalog page, try `scrapling extract` with `--ai-targeted` first, and save the result as Markdown.Limitations
The skill does not bundle the library: first install `scrapling[fetchers]` and browser dependencies with `scrapling install`. Cloudflare and WAF handling relies on `StealthyFetcher`, and results depend on the site and the Camoufox version. Real cookies belong in a local `references/cookie-vault.local.md` that must never be committed. The agent reads page content, which can carry prompt injection. The last repository commit is from June 2026.
How to disable. Delete `~/.claude/skills/scrapling` or the project copy `.claude/skills/scrapling`. Clean local cookie files separately.
FAQ
How do I install the skill?
First install Scrapling with `pip install "scrapling[fetchers]"` and run `scrapling install`. Then copy the repository into `~/.claude/skills/scrapling` or a project's `.claude/skills/scrapling`.
Which sample requests does the README give?
Pull the title and body of a blog, collect prices from a shop page, handle Cloudflare, parse ready HTML.
Where should cookies be kept?
Only in `references/cookie-vault.local.md`. The `cookie-vault.md` template stays empty and real values are never entered into it.
What is the `--ai-targeted` flag for?
It prepares `scrapling extract` output for an agent. Details are in `references/security.md`.
Are there built-in script templates?
Yes, `templates/` holds `basic_fetch.py`, `parse_only.py`, `session_login.py` and `stealth_cloudflare.py`.
Discussion
No questions or reviews yet.
Related
SkillsAround 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
A community collection of Apify skills: ready routes to Actors for leads, jobs, prices, ads, Reddit and Yandex search
Official Apify skills for agents: web scraping via ready-made Actors, plus building, actorizing and integrating Actors
Official ScrapeAtlas terminal client: an agent or script pulls public data from 49 platforms with one command and gets plain JSON
A Claude Code agent for protein design: target research, antibody and nanobody generation, screening and candidate ranking.
A skill retrieves Carrel evidence, reads source passages and images and helps agents cite original files in answers.
README in short
The README is bilingual, English and Chinese: features, a two-step install, structure, sample prompts and cookie storage rules.