ai-risk-management

What problem does it solve?

This Skill helps you assess and govern AI/ML risk across the model lifecycle, including governance, fairness and bias evaluation, robustness, transparency, accountability, third-party model risk, monitoring for drift, and AI incident response—using the NIST AI Risk Management Framework (AI RMF 1.0) as the organizing structure.

Core Features & Use Cases

• NIST AI RMF 1.0 end-to-end workflow: Guides you through GOV (govern), MAP (map context), MEASURE (evaluate), and MANAGE (treat risks) for AI systems.
• Practical MEASURE methods: Covers accuracy/performance, fairness/bias metrics, robustness and distribution shift testing, explainability/transparency, privacy leakage risk, and security posture summaries (with prompt-injection handled as the adjacent security slice).
• Regulatory alignment guidance: Provides high-level mappings and action items for governance expectations across the EU AI Act and US enforcement landscape.
• Governance artifacts and operationalization: Produces inventory structures, risk registers, management plans, and posture checklists suitable for internal review gates and documentation cadence.

Quick Start

Ask the AI agent to produce a NIST AI RMF 1.0 risk assessment for your production AI system named "Customer Support Assistant," including a system inventory, MEASURE findings for fairness and robustness, and a 90-day MANAGE plan based on the highest-risk gaps.