An advanced automation tool for driving ChatGPT Web conversations via local browser profiles with persistent state.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install chatgpt-skill
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 chatgpt-skill using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The ChatGPT Web Skill provides a robust framework for automating interactions with the ChatGPT web interface. By leveraging the power of Openclaw Skills, developers can bypass API limitations to perform one-shot prompts, manage multi-turn dialogues, and maintain persistent login states through a local browser profile. This skill is particularly useful for workflows requiring the specific behavior of the ChatGPT web UI, including specialized models or extended thinking features that might not be available via standard API endpoints.
It ensures that all data, including cookies and conversation metadata, remains stored locally within the skill directory for maximum privacy and control. This makes it an essential tool for those who need to integrate ChatGPT's latest web features into their automated pipelines while maintaining a consistent user session.
First, verify the authentication status of your local browser profile within the Openclaw Skills environment:
python3 scripts/run.py auth_manager.py status
If login is required, use the setup command to open a visible browser and complete the authentication manually:
python3 scripts/run.py auth_manager.py setup
Once authenticated, you can send your first prompt:
python3 scripts/run.py ask_chatgpt.py --question "Hello, how can you help me today?"
The skill organizes its data within the data/ directory to ensure all state is persistent and localized:
| File/Folder | Description |
|---|---|
auth_info.json |
Metadata regarding authentication and last validation status. |
browser_state/ |
Persistent browser profile including cookies and storage state. |
conversations.json |
Index of discovered and opened conversation metadata. |
sessions.json |
Active and historical persistent session metadata. |
session_runtime/ |
Daemon sockets, PIDs, and runtime artifacts for active sessions. |
screenshots/ |
Debugging screenshots captured during execution or verification failures. |
Loading
A high-performance search and web data extraction toolkit designed to give your AI agent real-time internet access.

A sophisticated Model Context Protocol integration that connects Meegle's visual agile management tools directly to your AI development workflow.

Clawchest is a secure banking and storage infrastructure for AI agents to deposit files, JSON data, and encrypted secrets.

An entertainment-focused Bitcoin solo mining skill that uses idle CPU power for a lottery-style chance at block rewards.

A specialized tool for permanent, decentralized website and file deployment on the MegaETH mainnet using SSTORE2 technology.

A developer tool for deploying permanent, immutable websites on the MegaETH blockchain using SSTORE2 bytecode storage.








































