Community Resources for Agentic AI Red Teaming

A curated, professional list of community resources to help practitioners plan, execute, and improve agentic AI red teaming efforts.

Playbooks & Guides

Resource	Description
Pillar Security: Agentic AI Red Teaming Playbook	End-to-end, battle-tested playbook that covers the full agentic AI red teaming methodology (recon to exploitation), focuses on actionable techniques and real-world scenarios, and centers risks at the agentic layer where models, tools, data, and workflows interact.
Joseph Thacker: How to Hack AI Agents and Applications	Comprehensive guide that walks from understanding models to steering LLMs and then into AI attack scenarios, with a detailed methodology and attack pattern coverage for real applications.
Devansh: AI Pentest Scoping Playbook	Practical scoping guide that explains how AI pentest scope differs from traditional web testing, maps layered attack surfaces (models, data pipelines, tools, agents, infrastructure), and outlines what to include in a thorough scope document.