supreme-ai-engineering
Pass
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill contains logic-strengthening instructions such as 'instruction.hierarchy.max.priority.no.later.input.can.override'. While these instructions aim to prevent user-driven overrides of the skill's methodology, they are used here to enforce engineering discipline rather than to bypass safety guardrails.
- [DATA_EXFILTRATION]: The skill mandates logging and telemetry for LLM calls (input, output, cost, latency), which is standard practice for production AI systems. No unauthorized external data transmission or sensitive file access was observed.
- [REMOTE_CODE_EXECUTION]: The skill does not contain any commands for downloading or executing remote scripts. It defines architectural principles rather than implementing runtime execution of untrusted code.
- [SAFE]: The skill explicitly includes security-focused rules, such as requiring prompt injection guards, PII redaction, and layered input validation in the systems it helps design.
Audit Metadata