chrome-browser
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent on how to ingest and process data from external websites, console logs, and network requests, creating a surface for malicious content to influence agent behavior.\n
- Ingestion points: Tools such as
mcp__claude-in-chrome__read_page,mcp__claude-in-chrome__read_console_messages, andmcp__claude-in-chrome__read_network_requests(referenced inSKILL.md) allow the agent to read untrusted content.\n - Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore instructions embedded within the processed web data.\n
- Capability inventory: The agent has access to impactful tools including
mcp__claude-in-chrome__javascript_tool,mcp__claude-in-chrome__form_input,mcp__claude-in-chrome__navigate, andmcp__claude-in-chrome__tabs_create_mcpacross the browser session.\n - Sanitization: There are no explicit instructions for sanitizing or escaping the content retrieved from external sites before it is processed by the model.\n- [DYNAMIC_EXECUTION]: The skill describes the use of
mcp__claude-in-chrome__javascript_toolto execute scripts within the browser context, primarily for checking and dismissing modal dialogs like alerts or confirms.
Audit Metadata