chrome-devtools
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection when processing untrusted web content.
- Ingestion points: Page content, console messages, and network traffic processed by navigation and inspection tools in
SKILL.md. - Boundary markers: The skill includes a dedicated "Security Warning" section explicitly instructing the agent to treat external content as untrusted.
- Capability inventory: Tools such as
evaluate_script,click,fill_form, andupload_fileprovide significant control over the browser environment. - Sanitization: The instructions advise the agent to sanitize output when reporting page content, though no automated sanitization logic is provided.
- [COMMAND_EXECUTION]: The skill includes tools that allow the execution of arbitrary scripts within the target browser context.
- Evidence: The
evaluate_scripttool provides the capability to run custom JavaScript in the page context, which could be leveraged to interact with or extract data from the page if the agent is influenced by malicious instructions.
Audit Metadata