agent-platform-alert-configuration

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • Command Execution for Cloud Discovery: The skill uses subprocess.run to call official CLI tools like gcloud and bq. These commands are used to discover agent runtimes, verify telemetry status, and check for existing monitoring configurations within the user's authorized environment. The use of list-based arguments and the absence of shell execution mitigates standard command injection risks.
  • Interaction with Trusted GCP APIs: Multiple scripts perform network requests to official Google Cloud API endpoints (e.g., logging.googleapis.com, monitoring.googleapis.com) to retrieve metadata. These operations use standard Google Auth libraries to handle OAuth2 tokens securely.
  • Standard Package Dependencies: The skill identifies necessary Python dependencies in a requirements.txt file, which includes well-known libraries such as google-cloud-aiplatform and requests. These are standard tools for interacting with cloud infrastructure.
  • Instructional Safety Tiers: The SKILL.md file defines a clear 'Safety & Confirmation Tier' system. It distinguishes between read-only discovery (Tier R) and actions that incur costs or create resources (Tier B), enforcing a mandatory 'Explicit User Confirmation' step for the latter.
  • Configuration Validation: The skill includes a dedicated linting utility (lint_syntax.py) that uses regular expressions to validate the syntax of generated PromQL and Terraform HCL configurations, helping to ensure the integrity of produced infrastructure-as-code files.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 07:15 AM
Security Audit — agent-trust-hub — agent-platform-alert-configuration