A browser-based automation skill for X (Twitter) that enables posting tweets, threads, and media without expensive API fees.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install x-browser
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install x-browser using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
x-browser is a high-performance automation skill designed to interact with X (Twitter) through the HTTP Browser API. Unlike traditional methods that rely on the increasingly expensive and restrictive official API, this tool mimics human browser behavior to perform actions. It is a core component of the Openclaw Skills ecosystem for developers who need to programmatically manage social media presence.
By leveraging sequence-based actions such as clicking, typing, and waiting for specific DOM selectors, x-browser provides a resilient way to handle posts, replies, and complex threads. The skill operates via a headless browser environment, ensuring that your automation remains functional even as official API specifications change, while significantly reducing operational costs for social media bots and agents.
To start using this skill, configure your environment variables and use the sequence API.
export MOLTBOT_URL="https://your-worker.workers.dev"
export CDP_SECRET="your-secret"
You can then trigger a tweet via a simple curl command:
curl -X POST "${MOLTBOT_URL}/browser/sequence?secret=${CDP_SECRET}" \
-H "Content-Type: application/json" \
-d '{
"url": "https://x.com/compose/tweet",
"actions": [
{"type": "waitForSelector", "selector": "[data-testid=\"tweetTextarea_0\"]"},
{"type": "type", "selector": "[data-testid=\"tweetTextarea_0\"]", "text": "Hello from Openclaw Skills!"},
{"type": "wait", "ms": 1000},
{"type": "click", "selector": "[data-testid=\"tweetButton\"]"},
{"type": "wait", "ms": 3000},
{"type": "screenshot"}
]
}'
The skill organizes its operations using a JSON-based action manifest. Below is the primary metadata and action structure used within Openclaw Skills:
| Field | Type | Description |
|---|---|---|
url |
string | The target X.com page (e.g., /home, /compose/tweet) |
type |
string | The action type: waitForSelector, type, click, wait, or screenshot |
selector |
string | The CSS selector or data-testid used to find UI elements |
text |
string | The content to be typed into text areas or input fields |
ms |
number | Duration in milliseconds for wait actions |
Loading
A robust integration for posting text, images, and replies to X/Twitter using official API v2 and OAuth 1.0a.

A comprehensive Linux-based diagnostic tool for analyzing Wi-Fi signal strength, channel congestion, and network throughput.

A high-speed web search skill that provides summarized information from multiple sources including technical docs and news.

A robust validation framework that enforces successful tests, builds, and type checks before an AI agent can declare a task complete.

A comprehensive AI-driven pipeline for automating faceless YouTube channel creation, from niche research and script writing to video production and SEO-optimized uploads.

A sophisticated tool designed to bridge OpenClaw with Docker-based Xunlei services for automated, intelligent magnet link downloading.








































