cloud-monitoring-alert

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODECOMMAND_EXECUTION
Full Analysis
  • [SAFE] (SAFE): No malicious code, scripts, or patterns detected within the skill definition.
  • [NO_CODE] (SAFE): The skill is purely informational and does not bundle any executable binaries or scripts.
  • [COMMAND_EXECUTION] (LOW): The skill requests broad access to 'Bash(gcloud:*)'. While appropriate for a GCP administration skill, users should audit generated commands before execution to prevent unintended cloud resource modifications.
  • [Indirect Prompt Injection] (LOW): * Ingestion points: User requests and data related to GCP alerts. * Boundary markers: Absent. * Capability inventory: Read, Write, Edit, Bash(gcloud:*). * Sanitization: Not specified. The skill is susceptible to instructions embedded in processed alert data if used in an automated pipeline.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:15 PM