gitlab-ci-patterns
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The Multi-Environment Deployment examples include the use of
kubectlwith the--insecure-skip-tls-verify=trueflag. Disabling TLS verification is a significant security risk as it allows for man-in-the-middle attacks against the Kubernetes API server connections.\n- [PROMPT_INJECTION]: The skill operates on untrusted user inputs (goals and constraints) to generate CI/CD configurations, establishing an indirect prompt injection surface.\n - Ingestion points: User-defined goals, constraints, and inputs mentioned in the instructions for SKILL.md.\n
- Boundary markers: None identified.\n
- Capability inventory: Ability to generate and execute shell commands and trigger dynamic CI/CD pipelines.\n
- Sanitization: No input validation or sanitization routines are specified.
Audit Metadata