Curriculum

The cryptography and access-control ideas everything else rests on. No 1Claw yet.

By the end you’ll be able to explain envelope encryption, hardware roots of trust, policy models, and threshold cryptography — and say why each one exists.

What actually breaks when an LLM can act: injection, the lethal trifecta, and OWASP ASI01-10.

By the end you’ll be able to audit any agent design against the lethal trifecta and OWASP ASI01–10, and name the control that answers each risk.

Get the core ideas, install the CLI, and store your first secret.

By the end you’ll be able to create a vault, store and read a secret, and connect an AI client to it over MCP.

Wire secrets into CI, apps, and the SDK, then rotate them safely.

By the end you’ll be able to wire secrets into CI and applications, scope them per environment, and rotate them without downtime.

Give agents an identity, scope them tightly, and connect over MCP.

By the end you’ll be able to give an agent a tightly scoped identity and let it call external APIs without ever holding a credential.

Go deep on the HSM key hierarchy, CMEK, MPC, Shroud, and risk.

By the end you’ll be able to configure CMEK, MPC custody, TEE-routed inference, and step-up authentication for treasury operations.

Sign onchain from agents, run a treasury, and pay with x402.

By the end you’ll be able to sign on-chain transactions from an agent under hard guardrails, with humans approving what matters.

Query the audit log, run agent fleets, and build multi-tenant apps.

By the end you’ll be able to answer who-did-what from the audit log, run a fleet of agents, and operate 1Claw for many tenants.

Plug 1Claw into Python, MCP clients, and your agent framework.

By the end you’ll be able to plug 1Claw into Python, LangChain, CrewAI, elizaOS, and containerized agent runtimes.