agent-spec
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a structured framework for designing AI agents, prioritizing safety, explicit authority, and human approval gates for irreversible actions. It encourages the implementation of guardrails, cost budgets, and escalation triggers in agent designs. No technical vulnerabilities such as command execution, data exfiltration, or obfuscation were identified.
Audit Metadata