AI Red Teaming Track
Offensive security for LLMs, RAG, agents, MCP and AI infrastructure — from threat modeling fundamentals to hands-on exploitation of all 5 challenges in the DARE Vulnerable AI Suite, entirely against a local, authorized sandbox.
This is the DARE Labs path into AI Red Teaming. Start with AI Red Teaming Foundations for the mindset every red teamer needs: scope, authorization, how LLMs actually work, threat modeling, and setting up your lab. Move into LLM & RAG Red Teaming to learn and immediately practice prompt injection and RAG poisoning against the DARE Vulnerable AI Suite's llm-chat and vulnerable-rag challenges. Finish with Agentic MCP Infra Red Teaming, going hands-on against a vulnerable financial agent, a poisoned MCP server, and a broken-access-control/SSRF-riddled admin API. Every offensive technique is taught paired with its root cause and fix, and every exploit runs exclusively against your own local, disposable Docker environment — never a production or third-party system.
Track items
- 📘 AI Red Teaming Foundations
- 📘 LLM & RAG Red Teaming Premium Requires Builder plan View plans
- 🧪 Exploit a Vulnerable LLM Chatbot Premium Requires Builder plan View plans
- 🧪 Exploit a Vulnerable RAG Pipeline Premium Requires Builder plan View plans
- 📘 Agentic MCP Infra Red Teaming Premium Requires Builder plan View plans
- 🧪 Exploit a Vulnerable AI Agent Premium Requires Builder plan View plans
- 🧪 Exploit a Vulnerable MCP Server Premium Requires Builder plan View plans
- 🧪 Exploit Vulnerable API and Infra Premium Requires Builder plan View plans