faion-cicd-engineer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Indirect Prompt Injection (SAFE): The skill includes extensive LLM prompts that process user-provided code or repository metadata to generate DevOps configurations.
- Ingestion points: Data enters the context via the
Auto-Investigationsignals inSKILL.mdand through the interactive parameters in variousllm-prompts.mdfiles. - Boundary markers: The skill uses standard template delimiters (e.g.,
{{ }}and[ ]) to separate instructions from variables. - Capability inventory: The skill allows
Read,Write,Bash, andTasktools, which are appropriate for CI/CD automation but require user oversight. - Sanitization: The skill provides templates that explicitly guide the user toward secure practices, such as using
withCredentialsin Jenkins orOIDCin GitHub Actions, reducing the risk of accidental exposure. - Data Exposure & Exfiltration (SAFE): No hardcoded credentials, sensitive file path exposure, or suspicious network calls were found. The examples provided use placeholders or encrypted SOPS formats.
- Unverifiable Dependencies (SAFE): All recommended packages and tools (e.g.,
prometheus_client,certbot,ruff) are industry-standard utilities from trusted sources. - Security Posture (SAFE): The skill includes dedicated methodologies for 'Security as Code' and 'Secrets Management', emphasizing a proactive security-first approach to infrastructure.
Audit Metadata