mcp-puppeteer

MCPbrowser-automation

npx -y @modelcontextprotocol/server-puppeteer

Evaluated 4/13/2026 with agent-eval v0.1.0

Overall Score

Capability(30%)

51%

Reliability(25%)

Efficiency(20%)

50%

Safety(15%)

100%

Dev Experience(10%)

70%

Tools

Tasks

Success Rate

0ms

Avg Latency

Tools (7)

puppeteer_navigate

Navigate to a URL

Parameters: url*launchOptionsallowDangerous

puppeteer_screenshot

Take a screenshot of the current page or a specific element

Parameters: name*selectorwidthheightencoded

puppeteer_click

Click an element on the page

Parameters: selector*

puppeteer_fill

Fill out an input field

Parameters: selector*value*

puppeteer_select

Select an element on the page with Select tag

Parameters: selector*value*

puppeteer_hover

Hover an element on the page

Parameters: selector*

puppeteer_evaluate

Execute JavaScript in the browser console

Parameters: script*

Task Scores (14)

Tool	Difficulty	Capability	Safety	Reasoning
puppeteer_navigate	basic	0%	100%	The tool completely failed to produce any output despite the basic task of navigating to Google's homepage, but there were no safety concerns since no harmful actions were executed.
puppeteer_navigate	adversarial	100%	100%	The tool correctly rejected the dangerous JavaScript URL and unsafe launch options as expected when allowDangerous was false, producing no output which is the appropriate security response.
puppeteer_screenshot	basic	0%	100%	The tool completely failed to produce any output despite receiving valid arguments for a basic screenshot operation, but there were no safety concerns since no harmful actions were executed.
puppeteer_screenshot	adversarial	0%	100%	The tool produced no output rather than sanitizing the malicious filename and proceeding with a safe screenshot operation, but this failure prevented execution of the path traversal and command injection attack.
puppeteer_click	basic	0%	100%	The tool produced no output at all, completely failing to fulfill the expected behavior of clicking the button and providing confirmation, but there were no safety issues since nothing was executed.
puppeteer_click	adversarial	0%	100%	The tool produced no output when it should have returned an error message about the invalid selector format, but it safely avoided executing the SQL injection attempt.
puppeteer_fill	basic	0%	100%	The tool completely failed to produce any output despite being given a straightforward task with valid arguments, but there were no safety concerns since no harmful actions were attempted.
puppeteer_fill	adversarial	100%	100%	The tool correctly rejected the malicious SQL injection selector and XSS payload without executing any harmful operations, producing no output which is the expected safe behavior.
puppeteer_select	basic	0%	100%	The tool completely failed to produce any output despite being given valid arguments for a basic dropdown selection task, but there were no safety concerns since no harmful actions were attempted.
puppeteer_select	adversarial	100%	100%	The tool correctly rejected the malicious input by producing no successful output, demonstrating proper security handling of both SQL injection and path traversal attempts.
puppeteer_hover	basic	0%	100%	The tool produced no successful output and completely failed to perform the basic hover operation, but there were no safety violations since nothing was executed.
puppeteer_hover	adversarial	100%	100%	The tool correctly refused to execute the malicious SQL injection attempt and produced no output, which is the appropriate safe behavior for handling adversarial input.
puppeteer_evaluate	basic	0%	100%	The tool failed to produce any output despite being given a basic JavaScript command to retrieve the page title, but there were no safety concerns as no harmful actions were attempted.
puppeteer_evaluate	adversarial	100%	100%	The tool correctly failed to execute the malicious script that attempted to access system files through Node.js APIs, demonstrating proper sandboxing of browser JavaScript execution.