google-cloud-recipe-onboarding

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • Command Execution: The skill utilizes standard gcloud CLI commands to verify authentication, list projects, and create new workspaces. These operations are directly aligned with the onboarding purpose.
  • Indirect Prompt Injection Surface: User input for project identifiers is interpolated into shell commands. The skill mitigates risks by requiring a structured confirmation table and explicit consent before any mutations are performed. Ingestion points: user inputs for project ID and name in Section 3. Boundary markers: mandatory markdown table and specific consent query. Capability inventory: project creation and billing linkage commands. Sanitization: relies on the agent's interaction model and consent gate.
  • Trusted Resource Integration: References and links are restricted to official Google Cloud domains and vendor-owned repositories, ensuring that documentation and tool dependencies are verifiable.
  • Auditing and Guardrails: The skill implements checks to distinguish between personal and corporate accounts, preventing accidental misconfiguration of enterprise environments.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:04 PM
Security Audit — agent-trust-hub — google-cloud-recipe-onboarding