ci-cd-best-practices

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Data Exposure & Exfiltration] (SAFE): The skill correctly demonstrates the use of environment variables (e.g., $DEPLOY_KEY) for secret management rather than hardcoding sensitive information.
  • [Unverifiable Dependencies & Remote Code Execution] (SAFE): No remote scripts are downloaded or executed. Example commands like npm ci are standard for the described use case and are presented as static documentation.
  • [Prompt Injection] (SAFE): The content consists of instructional guidance for an expert persona without attempting to override safety filters or bypass system instructions.
  • [Privilege Escalation] (SAFE): The skill includes security-positive examples, such as using chmod 600 to restrict access to deployment keys, rather than attempting to escalate privileges.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:02 PM