Course content.
5 weeks (self-paced) · 38 hours · 5 lessons
AI Agent Architectures in Web3
Anatomy of an on-chain AI agent: planner, tool-caller, signer · Survey of deployed agent frameworks (Eliza, GAME, custom) · Threat modelling an agent that can move funds
Prompt Injection & Agent Exploitation
Direct and indirect prompt injection in agent contexts · Exploiting tool-use: making an agent call the wrong contract · Data exfiltration through agent observation channels
Guardrails & Policy Enforcement
Transaction simulation before signing (Tenderly, local fork) · Allowlists, spend limits, and rate limiting at the signing layer · Kill switches and human-in-the-loop escalation
Key Management for Autonomous Systems
Session keys, scoped permissions, and ERC-7715 / ERC-7710 · TEE and enclave-based signing for agents · MPC-based agent wallets and threshold approval
Monitoring, Incident Response & Capstone
Real-time monitoring of agent-initiated transactions · Alert pipelines and automatic pause triggers · Capstone: red-team an AI agent and write the defence report