browser
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local CLI commands for service management, including
bdata zones --jsonto fetch configuration andbdata browser opento initialize sessions. It also suggests usinggh secret setto securely store connection strings in GitHub Actions secrets.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API endpoints (e.g., customer IDs, zone names, and passwords) and interpolates this data into generated code snippets for Playwright, Puppeteer, and Selenium. This creates a surface for indirect prompt injection, though the risk is minimized as the data source is the vendor's own authenticated API and the skill includes basic validation for inputs like country codes.\n- [DATA_EXFILTRATION]: The skill instructions involve reading local sensitive files such ascredentials.jsonfrom the Bright Data CLI configuration directories. This information is used to authenticate requests to the official Bright Data API and is handled in-process, with explicit warnings provided to the agent against logging, printing, or otherwise exposing these credentials.
Audit Metadata