ai-search-visibility
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads the Bruin CLI installation script from the official vendor domain (getbruin.com). This is a standard setup step for the tool.
- [REMOTE_CODE_EXECUTION]: The downloaded installation script is executed directly via the shell (sh). While this is a common installation pattern for CLI tools, it constitutes remote code execution from a vendor-controlled source.
- [COMMAND_EXECUTION]: The skill executes multiple local commands including bruin (for initialization, validation, and running pipelines), git (for repository management), and host-specific commands like claude mcp add to configure the environment.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could contain malicious instructions. 1. Ingestion points: Reads Search Console queries and crawler logs via S3, GCS, or BigQuery as described in SKILL.md and analysis.md. 2. Boundary markers: None identified in the instructions for isolating untrusted data. 3. Capability inventory: Uses bruin run and bruin validate which execute data transformation logic. 4. Sanitization: No specific sanitization or filtering of the ingested log or search data is mentioned.
Audit Metadata