Cloakbrowser Skill
Claude skill that wraps CloakBrowser stealth Chromium — fetch, screenshot, PDF, and run JS through anti-bot detection (Cloudflare, reCAPTCHA, FingerprintJS) from any Claude agent.
Install
npx skills add OctavianTocan/cloakbrowser-skillclaude mcp add cloakbrowser-skill -- npx skills add OctavianTocan/cloakbrowser-skill{
"mcpServers": {
"cloakbrowser-skill": {
"command": "npx",
"args": [
"skills",
"add",
"OctavianTocan/cloakbrowser-skill"
]
}
}
}Install command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →cloakbrowser-skill
A Claude skill that wraps CloakBrowser — stealth Chromium that passes most bot-detection layers (Cloudflare Turnstile, reCAPTCHA v3, FingerprintJS, ShieldSquare, BrowserScan, generic 403/429 walls).
Drop it into your Claude agent and it can fetch, screenshot, PDF, or run JavaScript against sites that block normal headless browsers — no hand-rolled Playwright code per task.
Install — one command
Uses the vercel-labs/skills installer. It auto-detects every Claude agent / coding harness you have installed and copies the skill into each one's skills folder:
npx skills add OctavianTocan/cloakbrowser-skill
Then run the post-install setup once per host (installs CloakBrowser itself,
downloads the patched Chromium, and pins a stable Python wrapper at
/usr/local/bin/cloakbrowser-python):
~/.claude/skills/cloakbrowser/install.sh # or wherever the installer dropped it
That's it. The skill is now usable by any agent on this machine.
Manual install (no skills CLI)
git clone https://github.com/OctavianTocan/cloakbrowser-skill.git \
~/.claude/skills/cloakbrowser
~/.claude/skills/cloakbrowser/install.sh
The installer is idempotent — re-run it on every new host or after upgrading CloakBrowser.
What's in here
cloakbrowser-skill/
├── SKILL.md # frontmatter + agent instructions
├── install.sh # post-add one-time setup
├── scripts/
│ ├── fetch.py # HTML / visible text, optional CSS selector
│ ├── screenshot.py # PNG (full page, element, or custom viewport)
│ ├── pdf.py # render page → PDF
│ ├── eval-js.py # run a JS expression, get JSON
│ ├── check-stealth.py # confirm stealth fingerprint passes 4 core tells
│ └── serve.py + serve.sh # long-lived CDP endpoint for shared sessions
└── references/
└── stealth-knobs.md # when to use humanize / proxies / geoip / profiles
The agent reads SKILL.md, picks the right script for the task, shells out.
No heredoc Python per call.
Usage at a glance
SK=~/.claude/skills/cloakbrowser/scripts
$SK/fetch.py https://example.com # full HTML
$SK/fetch.py https://example.com --text --selector h1 # scoped text
$SK/screenshot.py https://example.com out.png --full
$SK/pdf.py https://example.com out.pdf --landscape
$SK/eval-js.py https://news.example "Array.from(document.querySelectorAll('h2')).map(h=>h.innerText)"
$SK/check-stealth.py # exit 0 if stealth passes
$SK/serve.sh --port 9222 # long-lived CDP for shared sessions
All scripts take --humanize, --proxy URL, --geoip, --wait STATE,
--timeout MS where they make sense. See references/stealth-knobs.md for the
escalation order when a site still blocks.
For multi-step interactive flows (login, click, conditional waits) the agent
writes inline Python against the Playwright-shaped API — see "Inline Python"
in SKILL.md.
Why a skill rather than just uv tool install cloakbrowser?
CloakBrowser is a library, not a CLI. Without this skill, every agent run reinvents the same Playwright boilerplate as a heredoc. The skill gives the agent:
Similar plugs
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.
Production-grade engineering skills for AI coding agents.
The 1st Collaborative AI Design Agent Workspace. Local-first and open source. Create prototypes, websites, apps, dashboards, presentations, images, and videos with your team…