google-agents-cli
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains standard developer documentation and code patterns for an agent development kit. No malicious code, obfuscation, or unauthorized data access patterns were detected.\n- [COMMAND_EXECUTION]: The skill describes a CLI tool (
agents-cli) that handles project scaffolding, deployment to Google Cloud, and evaluation runs. These operations are expected behaviors for the described development utility.\n- [INDIRECT_PROMPT_INJECTION]: The skill includes components for data ingestion and RAG workflows that process external content. Mandatory Evidence Chain:\n - Ingestion points: Cloud storage buckets and local directories accessed via
agents-cli data-ingestion.\n - Boundary markers: None explicitly mentioned in the code samples.\n
- Capability inventory: Cloud service provisioning, project scaffolding (file writes), and agent execution.\n
- Sanitization: The
CalculatorToolprovides a safe execution pattern usingast.parsefor mathematical operations instead of unsafeeval()functions.
Audit Metadata