emmanuelgjr/AI-RedTeam-Framework

# AI Red Team Framework [](https://airedteam.dev) [](LICENSE) [](CONTENT-LICENSE) [](https://airedteam.dev/attacks) The practitioner playbook + interactive tooling for planning, scoping, executing, and reporting AI red team engagements at regulated enterprises. [**Try the Engagement Planner →**](https://airedteam.dev/planner) ## What's inside (v0.1) - 15-chapter playbook from charter to maturity model - **Engagement Planner** — answer 12 questions, download a professionally drafted Word Rules of Engagement - 25 attack patterns mapped to OWASP LLM Top 10, OWASP Agentic Top 10, MITRE ATLAS, NIST AI RMF - 15 tool reviews — open source, commercial, AI-augmented BAS — with honest limitations - Compliance crosswalk: OSFI E-21, NIST AI RMF, ISO/IEC 42001 ## Why this exists By 2027, every Schedule I bank, large insurer, and major healthcare system in North America will have an AI red team function — or be told by a regulator to build one. Almost no public guidance exists on *how* to actually stand one up at a regulated enterprise. This framework fills that gap. ## Who built this **Emmanuel Guilherme Jr.** — Senior Global IT Auditor, Co-lead of the OWASP GenAI Data Security Initiative, Candidate Expert on Canada's ISO/IEC JTC 1/SC 42 mirror committee. Built an in-house Red Team framework end-to-end at a Canadian Schedule I bank, including AI-augmented offensive tooling (NodeZero, Cymulate). ## Ethics This is a defensive resource for authorized testing. We do not host weaponized payloads or zero-days. See [ETHICS.md](ETHICS.md). ## Local development npm ci npm run dev # http://localhost:4321 npm run validate npm run test npm run build ## Use it. Cite it. Guilherme Jr., E. (2026). AI Red Team Framework (Version 0.1.0). https://airedteam.dev ## Licensing - **Code:** [MIT](LICENSE) - **Content** (patterns, playbook, templates): [CC-BY 4.0](CONTENT-LICENSE) ## Related work in this portfolio - [AI-Controls-Catalog](https://github.com/emmanuelgjr/AI-Controls-Catalog) — Audit AI - [AI-RedTeam-Framework](https://github.com/emmanuelgjr/AI-RedTeam-Framework) — Attack AI *(this repo)* - [AI-Governance-Toolkit](https://github.com/emmanuelgjr/AI-Governance-Toolkit) — Govern AI - [Shadow-AI-Defense](https://github.com/emmanuelgjr/Shadow-AI-Defense) — Defend against rogue AI

标签：自动化攻击