The Agent Skills Directory

[SAFE]: The skill performs legitimate visual regression testing. It uses the browser-use MCP to interact with web pages and local files for the purpose of visual validation without engaging in malicious activities.
[COMMAND_EXECUTION]: The instructions include a standard cp command to move screenshots into a local regression history folder (~/.claude/benchmarks/). This operation is consistent with the skill's purpose and does not involve privilege escalation or suspicious shell commands.
[DATA_EXFILTRATION]: Network activity is restricted to navigating a web browser to specific developer-provided URLs (local or remote) to analyze UI components. There is no evidence of harvesting credentials or sending sensitive user data to unauthorized external servers.

visual-verdict