The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external websites, creating a potential vector for indirect instructions.\n- Ingestion points: Untrusted data enters the agent context via mcp__chrome-devtools__get_console_message after executing snippets on external web pages.\n- Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions that might be embedded in the browser console output.\n- Capability inventory: The skill utilizes mcp__chrome-devtools__evaluate_script and mcp__chrome-devtools__navigate_page within the chrome-devtools MCP server to interact with web content.\n- Sanitization: There are no documented procedures for sanitizing or filtering the content retrieved from the console before the agent interprets it.

webperf