blue-green-deployment

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Several scripts in the references/ directory interpolate user-defined shell variables directly into commands without proper sanitization or quoting. For example, in deploy-blue-green.sh, the $VERSION variable is used in a kubectl set image command, and in validate-schema.sh, the $SCHEMA_FILE variable is used directly. This pattern allows for command injection if an attacker can control these input variables.
  • [INDIRECT_PROMPT_INJECTION]: The skill exposes an attack surface for indirect prompt injection by processing untrusted input variables in scripts that perform high-privilege operations.
  • Ingestion points: Shell script arguments in references/blue-green-with-load-balancer.md, references/blue-green-rollback-script.md, and scripts/validate-schema.sh (e.g., $ENVIRONMENT, $VERSION, $SCHEMA_FILE).
  • Boundary markers: None. The scripts do not implement input validation, schema enforcement, or boundary delimiters for the processed variables.
  • Capability inventory: The skill uses kubectl (set image, patch, exec, port-forward, rollout) and aws CLI (modify-listener, describe-target-groups), providing significant control over infrastructure.
  • Sanitization: Absent. Input variables are interpolated directly into shell strings without escaping or validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 10:44 AM
Security Audit — agent-trust-hub — blue-green-deployment