browsing-styles
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill scripts use dynamic module loading from computed paths. They attempt to resolve and import internal utility files by searching across directories and reading agent configuration metadata.\n- [COMMAND_EXECUTION]: The skill performs automated local command execution to open the default web browser and to run image optimization tasks using shell commands.\n- [DATA_EXFILTRATION]: The skill accesses agent-internal configuration. It reads the
installed_plugins.jsonfile to locate its own installation directory, exposing plugin environment metadata to the skill.\n- [EXTERNAL_DOWNLOADS]: The skill makes requests to an external AI service. It uses the Gemini API to generate reference images for artistic styles based on predefined prompts.
Audit Metadata