llmops-platform-engineering
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override agent behavior, bypass safety guidelines, or extract system prompts were found. The instructions are purely technical and focused on platform engineering.
- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or private secrets are present. The documentation correctly suggests using Kubernetes secrets and CI/CD environment variables.
- [EXTERNAL_DOWNLOADS]: The skill references standard, well-known GitHub Actions (actions/checkout, actions/upload-artifact) which are maintained by a trusted organization. No untrusted third-party script downloads are present.
- [COMMAND_EXECUTION]: The shell commands provided in the examples (kubectl, git, python -m) are standard for DevOps workflows and do not contain malicious payloads or unsafe user input interpolation.
- [OBFUSCATION]: The file contains clear, readable markdown and YAML. No Base64, zero-width characters, homoglyphs, or other obfuscation techniques were detected.
- [DATA_EXFILTRATION]: No patterns for sensitive file access or unauthorized network data transmission were identified.
- [DYNAMIC_CONTEXT_INJECTION]: The skill does not use the dynamic context execution syntax (!
command). All code blocks are static documentation examples.
Audit Metadata