Work

Selected Work

AI security standards authorship and CIS AI Benchmarks Community leadership as part of the CIS Benchmark Development Team.

Published Benchmarks

Public Author

CIS MCP Server Benchmark v1.0.0

Center for Internet Security · Released September 16, 2026

Authored the first CIS Benchmark for Model Context Protocol servers: 55 prescriptive, vendor-neutral recommendations across 10 security domains, covering governance and versioning, transport and connectivity, authentication and authorization, client (host) and server configuration, data protection and privacy, observability and audit, supply chain security, isolation and execution safety, and resource limits and caching. Each recommendation ships with rationale, audit procedures, and remediation guidance.

What it demonstrates

End-to-end ownership of a published security standard, from scoping MCP-specific risk through consensus review to auditable, implementable recommendations that organizations can configure and verify today.

MCP Security Security Benchmarks Secure Configuration Audit & Remediation Standards

Published Guides

Public Principal Co-author

CIS Controls v8.1 Model Context Protocol (MCP) Companion Guide

Center for Internet Security · April 2026

Translated MCP-specific risks into practical guidance mapped to all 18 Controls. Covers MCP clients, servers, tools, resources, prompts, authorization, least privilege, auditability, confused deputy risk, tool boundaries, third-party server governance, and secure configuration considerations. Co-authored with Shreyans Mehta (Cequence).

What it demonstrates

Andrew's ability to translate a fast-moving AI protocol into practical security guidance aligned with an established control framework, combining protocol-level technical depth with clear, actionable implementation guidance.

MCP Security Controls Mapping Authorization Secure Configuration Tool Boundaries

Public Collaborator

CIS Controls v8.1 AI and LLM Companion Guide

Center for Internet Security · April 2026

Collaborated on the AI and Large Language Models Companion Guide, contributing to practical guidance on LLM security, model-layer risk, inference behavior, data handling, monitoring, and control mapping for enterprise AI environments.

What it demonstrates

Andrew's ability to contribute meaningfully to practical AI security guidance for enterprise AI systems, bridging model-layer technical risk with organizational controls.

LLM Security Prompt Injection Data Handling Controls Mapping Model Risk

Public Collaborator

CIS Controls v8.1 AI Agents Companion Guide

Center for Internet Security · April 2026

Collaborated on the AI Agents Companion Guide, contributing to guidance on agent autonomy, tool invocation, memory, planning, human approval gates, monitoring, and secure integration of agentic AI workflows into enterprise environments.

What it demonstrates

Andrew's ability to connect agentic AI risk to practical security controls, making emerging multi-agent and autonomous AI risk patterns actionable for security practitioners.

Agentic AI Human Oversight Tool Invocation Controls Mapping Multi-Agent

Authorship note: Andrew Dannenberger is the author of the CIS MCP Server Benchmark v1.0.0 and Principal Co-author of the CIS Controls v8.1 MCP Companion Guide, alongside Shreyans Mehta (Cequence). The AI/LLM and AI Agents Companion Guides are collaborative Center for Internet Security (CIS) publications; Andrew contributed as a collaborator. Authorship credits reflect publicly available acknowledgment pages. This personal website does not represent CIS.

Active Initiatives

Ongoing Leader, CIS AI Benchmarks Community / CIS Benchmark Development Team

CIS AI Benchmarks Community

Center for Internet Security · Ongoing

Officially leads the CIS AI Benchmarks Community as a member of the CIS Benchmark Development Team. Owns scope definition, community coordination, consensus feedback cycles, and the publication roadmap for AI security standards, working with practitioners to turn AI security risk into prescriptive, auditable CIS Benchmark recommendations.

What it demonstrates

Standards leadership: building AI security Benchmarks end to end, from scoping emerging AI risk through community consensus to published, verifiable recommendations organizations can configure and audit.

Security Benchmarks AI Security Community Standards Leadership

Ongoing Hands-on MCP Implementation Experience

Platform MCP Concepts

Ongoing

Hands-on work developing MCP server concepts for a security platform and working through secure integration patterns, governance, and supportability with technical stakeholders. That implementation experience grounds the MCP security recommendations in how real deployments behave.

What it demonstrates

How implementation experience feeds standards work: recommendations shaped by real integration constraints are easier for teams to configure, verify, and audit.

MCP Secure Integration Implementation Standards Input

Public Ongoing

Personal website: This is a personal professional portfolio. This website does not represent Center for Internet Security (CIS) or any employer.