Show HN: Crosswalk mapping AI-agent design controls to NIST, ISO 42001, OWASP
A new tool called Crosswalk maps AI-agent design controls to standards including NIST AI RMF, ISO 42001, and OWASP, helping developers align agent safety and governance with established frameworks.
Background
- **Agent Kits** is a startup (agent-kits.com) that sells compliance tooling for teams building AI "agents" — autonomous systems that can act on behalf of a user, e.g., book flights, edit code, or handle customer support.
- **NIST AI Risk Management Framework (NIST AI RMF)**, **ISO/IEC 42001** (the international AI management-system standard), and **OWASP Top 10 for LLM Applications** are the three dominant emerging compliance frameworks. Companies deploying agentic AI increasingly need to show auditors they follow one or more of these standards.
- This "Crosswalk" is a mapping document that takes the design controls Agent Kits itself uses (e.g., "limit tool access per agent," "log all agent actions") and shows which specific clause in each standard each control satisfies.
- The motivation: agentic AI introduces new risks (an agent can take irreversible actions autonomously), and regulators globally are tightening rules. A crosswalk helps engineering teams translate vague regulatory language into concrete implementation steps, and helps compliance officers prove coverage during audits.