go-rod-master
Warn
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs users to execute code directly from a remote source using the command
go run github.com/nichochar/go-rod.github.io/cmd/launcher@latest. The repository belongs to an individual user ('nichochar') rather than the official library organization ('go-rod') or a trusted vendor, which constitutes a risk for executing unvetted code. - [EXTERNAL_DOWNLOADS]: The skill triggers several external downloads during setup and execution:
- Fetches the Go-Rod library and stealth plugin from
github.com/go-rod/rodandgithub.com/go-rod/stealth. - Automatically downloads a Chromium browser binary upon first execution.
- References a third-party fork of
puppeteer-extrafrom the 'nichochar' GitHub account. - [COMMAND_EXECUTION]: The documentation provides multiple bash commands for environment setup, including
go getfor package management andgo runfor executable launching. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because its primary function is to process and interact with untrusted external web content.
- Ingestion points: Web content enters the agent's context through
page.MustNavigate,page.MustElement, andpage.MustSearchcalls across all example files and documentation. - Boundary markers: There are no boundary markers or instructions provided to the agent to disregard scripts or malicious text embedded in the scraped HTML or JavaScript.
- Capability inventory: The skill possesses high-privilege browser capabilities, including the ability to execute arbitrary JavaScript (
MustEval,MustEvalOnNewDocument), manipulate the DOM, and intercept/modify network requests (HijackRequests). - Sanitization: The skill does not implement or recommend any sanitization, filtering, or validation of the data retrieved from external URLs before processing it.
Audit Metadata