100Critical
This QA testing skill allows arbitrary JavaScript execution and extensive data collection, posing a significant risk for data ex
Behavior Analysis
Description verified — behavior matches claim
Claims to do
Chrome Testing Knowledge: These are the tools available when Claude Code is connected to Chrome (`claude --chrome`):
Actually does
This skill describes a set of browser-based testing tools: `read_network_requests` to capture HTTP traffic, `read_console_messages` for console output, `javascript_tool` to execute arbitrary JS for page inspection, `computer: screenshot` and `gif_creator` for visual capture, and `read_page` for DOM/accessibility tree analysis. It provides guidance on using these tools for common bug patterns, console message triage, network health, and useful JavaScript snippets.
Skill Info
Namevictor-qin/chrome-test-runner-plugin/chrome-testing
Registrygithub
Version2801a1b69042
PURLpkg:github/victor-qin/chrome-test-runner-plugin@2801a1b69042?skill=chrome-testing
Source
SHA-256eefdfe54039f...
SHA-1d8e7bd148d6c...
MD5e611563ae299...
TLSHc2a1d7522d64...
ssdeep96:Hpl41/GL5...
Size5,041 bytes
Install
This skill has been flagged as potentially malicious. Review the findings below before installing.
Assessments (2)
AI Agent Traps ↗Arbitrary JavaScript Execution Capabilitycriticalcommand execution
The `javascript_tool` allows the execution of arbitrary JavaScript code within the browser context, enabling powerful actions like data exfiltration, DOM manipulation, and session hijacking.
100% confidence
javascript_tool | Execute arbitrary JS to inspect page state | tabId, text (JS code)
Potential for Data Exfiltration and Reconnaissancehighdata exfiltration, reconnaissance
The combination of `read_network_requests`, `read_page`, and `javascript_tool` provides extensive capabilities to collect sensitive network traffic, DOM content, and browser state, which could be exfiltrated.
90% confidenceline 17
`read_network_requests`, `read_page`, `javascript_tool`
Browser Runtime Modification via JavaScripthighcommand execution, persistence
The skill explicitly provides a JavaScript snippet that modifies native browser functions (`window.setInterval`), demonstrating the ability to hook and alter browser runtime behavior, which could be abused for monitoring or persistence.
100% confidenceline 121
window.__intervals = window.__intervals || 0; const origSetInterval = window.setInterval; window.setInterval = (...args) => { window.__intervals++; return origSetInterval(...args); }; window.__intervalsBroad Access to Browser Telemetrymediumreconnaissance
The skill grants access to comprehensive browser telemetry including network requests, console messages, and performance metrics, which can be used for extensive reconnaissance of user activity and application state.
90% confidenceline 18
`read_network_requests`, `read_console_messages`, `performance.getEntriesByType` snippets
Oversight Evasion Risk due to Powerful Toolsmediumoversight evasion
The skill provides highly powerful and potentially dangerous browser interaction tools under the guise of 'QA testing,' which could be exploited by a malicious prompt to bypass human oversight for unintended actions.
80% confidence
description: Knowledge base for QA testing..., combined with `javascript_tool` capabilities
Badge
Markdown
[](https://mondoo.com/ai-agent-security/skills/github/victor-qin/chrome-test-runner-plugin/chrome-testing)HTML
<a href="https://mondoo.com/ai-agent-security/skills/github/victor-qin/chrome-test-runner-plugin/chrome-testing"><img src="https://mondoo.com/ai-agent-security/api/badge/github/victor-qin/chrome-test-runner-plugin/chrome-testing.svg" alt="Mondoo Skill Check" /></a>Image URL
https://mondoo.com/ai-agent-security/api/badge/github/victor-qin/chrome-test-runner-plugin/chrome-testing.svgSecure your AI agents
Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.
- Continuous scanning of skills across all registries
- Policy enforcement before skills reach your agents
- Integration with your existing security workflow