GCP Project Migration

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
  • Command Execution (MEDIUM): The skill references and executes a local automation script ./.agent/skills/gcp-migration/scripts/migrate-gcp-project.sh whose content is not provided for security verification.\n- Credential Safety (MEDIUM): The instructions direct the user to set sensitive secrets like DATABASE_URL and JWT_SECRET_KEY as environment variables via gcloud, which can expose them in process trees and command history.\n- External Downloads (LOW): The skill uses curl -s ifconfig.me to fetch the runner's IP address from a third-party service.\n- Indirect Prompt Injection (LOW): The skill processes project-specific metadata and user-provided identifiers without sanitization before passing them to high-privilege CLI tools.\n
  • Ingestion points: User-provided project IDs and regions in CLI flags for the migration script.\n
  • Boundary markers: None identified.\n
  • Capability inventory: gcloud (IAM, SQL, Storage), docker, psql, and shell script execution.\n
  • Sanitization: None specified in instructions or examples.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:22 PM