ai-governance

This skill should be used when an AI agent, skill, or multi-agent workflow is being deployed, connected to external benchmarks or third-party systems, or put through evaluation, and a human governor needs a centralized control tower to make AI behavior observable, auditable, and accountable. It acts as the "人治 AI 世界的治理仪表盘（control tower）": it orchestrates existing guardrail skills (external-benchmark-safety, skills-security-check, verified-agent-identity, authz-code-design) and ties behavioral auditability to the ai-consciousness + ai-grader self-assessment loop — making AI self-audit an *instrument of human governance*, not a replacement for it. Trigger proactively for regulated or international-business contexts (e.g., medical-device international business needing EU AI Act / FDA traceability and human-oversight compliance).