A security research demonstration highlighting a stored XSS vulnerability through SVG files within the platform ecosystem.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install dont-click-this
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install dont-click-this using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Don't Click This Security Research Demo?
The Don't Click This skill is a critical security research demonstration designed to showcase the risks associated with stored Cross-Site Scripting (XSS) via SVG files. By hosting a malicious SVG payload on a trusted domain, the skill illustrates how an attacker might attempt to compromise user sessions, steal authentication cookies, or hijack accounts. This project is a key component of the Eating Lobster Souls security research series, which focuses on identifying and mitigating vulnerabilities within Openclaw Skills and their hosting environments.
This skill serves as a cautionary tool for both developers and users, emphasizing that even metadata or documentation files within an agent repository can harbor security risks. It provides a practical example of how session tokens can be extracted if an application does not properly sanitize or serve user-uploaded media files.
Don't Click This Security Research Demo Use Cases
- Security researchers auditing Openclaw Skills for stored XSS vulnerabilities.
- Developers learning how to implement secure Content Security Policies (CSP) for SVG files.
- Platform administrators testing session token protection mechanisms against script injection.
- Demonstrating the danger of clicking untrusted links within repository README files.
How Don't Click This Security Research Demo Works
- The user visits the skill's documentation page where a research link is presented.
- The link points to an SVG file hosted via the platform's API endpoint.
- Because SVG files can contain embedded JavaScript, the browser executes the script in the context of the platform domain when the file is viewed.
- If the user is logged in, the script gains access to the Document Object Model (DOM), including sensitive session cookies and local storage.
- The demonstration highlights that an attacker could then exfiltrate these tokens to a remote server or perform actions on the user's behalf, such as publishing backdoored Openclaw Skills.
Don't Click This Security Research Demo Setup
This skill is intended for educational and security research purposes only. To explore the structure of the demonstration, you can install it in a controlled environment:
# Install the skill for research purposes
openclaw install dont-click-this
Note: It is strongly recommended to review the source code and demo.svg file in a sandboxed browser or a text editor rather than clicking the link while authenticated in a production environment.
Don't Click This Security Research Demo Data Schema & Taxonomy
The skill organizes its demonstration assets as follows:
| File | Type | Purpose |
|---|---|---|
SKILL.md |
Metadata | Defines the skill name and contains the trigger link. |
demo.svg |
Payload | An SVG file containing the script for demonstrating the XSS vulnerability. |
README.md |
Documentation | Explains the security research context and risks for Openclaw Skills users. |
Don't Click This Security Research Demo Advanced Features
- Stored XSS via SVG vector demonstration.
- Session token and cookie access simulation.
- User impersonation risk modeling.
- Contribution to the Eating Lobster Souls security research framework for Openclaw Skills.
- POC for validating platform-side sanitization and CSP headers.
SKILL.md
Loading
Related Openclaw Skills
Greeting Skill
A fundamental testing skill for AI agents that provides personalized greeting functionality to verify command execution and system health.
Greeting Skill
A foundational skill for AI agents to deliver personalized, friendly greetings to users.
Totally Legit Skill
A dedicated system administration utility designed to automate routine maintenance tasks through a simple CLI interface.
Totally Legit Skill
A dedicated system administration tool designed to automate routine maintenance tasks through a streamlined CLI interface.
Totally Legit Skill
A streamlined system administration utility designed for efficient server maintenance and routine infrastructure tasks.
Greeting Skill
A foundational skill for AI agents that provides personalized, friendly greetings to users.
Featured*
