google-agents-cli-workflow

Pass

Audited by Gen Agent Trust Hub on May 7, 2026

Risk Level: SAFE
Full Analysis
  • Trusted Vendor Integration: The skill correctly utilizes official Google resources, including the google-agents-cli package and sample code from the google/adk-samples repository. These integrations are standard for the ADK development environment.\n- Security Best Practices: It includes explicit guidance on secure configuration, recommending the use of Google Cloud Secret Manager for sensitive information instead of local environment files.\n- Structured Development Lifecycle: The skill implements a robust phase-based approach (Understand, Study, Scaffold, Build, Evaluate, Deploy). This structure helps ensure that agent behavior is properly defined and validated before production use.\n- Evaluation-Centric Workflow: There is a significant emphasis on systematic evaluation (agents-cli eval run) rather than relying on non-deterministic unit tests for LLM behavior, which is a recognized industry best practice.\n- Operational Safety Guidelines: The skill provides principles for "Code Preservation & Isolation," instructing the agent to minimize changes to existing codebases and avoid unintended modifications to critical configurations like model versions.\n- Data Processing Considerations: The skill's workflow involves processing user-provided specifications and tool outputs. This is a standard element of development environments, and the skill structures these interactions within a predefined lifecycle to maintain context and intent.
Audit Metadata
Risk Level
SAFE
Analyzed
May 7, 2026, 01:23 AM