gcs-security-assessment
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses Python's
subprocess.runto executegcloudandcurlfor authentication and REST API communication with Google Cloud services. The implementation avoidsshell=True, utilizes list-based arguments, and includes input validation inscripts/validation.pyto prevent command injection. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from GCP resource metadata (bucket names, labels, tags) and BigQuery query results. This constitutes an indirect prompt injection surface; however, the skill mitigates this by enforcing strict structural templates for findings and instructing the agent to ground all analysis in verified telemetry rather than autonomous reasoning.
- [SAFE]: The skill is authored by a trusted vendor and all network communications are directed to official, well-known Google API endpoints (
*.googleapis.com). Security best practices are followed, such as writing sensitive headers to temporary files with restricted permissions (0600) to avoid exposure in process logs.
Audit Metadata