Tracks Explained
summary
Ledger shows a concrete model where an agent proposes actions and a device requires human confirmation before signing, keeping the private key in a secure element. The demo uses Ledger Wallet CLI to read balances, simulate transactions with multi-provider checks, and perform a deposit that needs two on-device confirmations; it then provisions Keyring CLI, encrypts an Anthropic API key on the Ledger Trust Chain, and lets the agent authenticate without plaintext exposure. Also covered: clear-signed transactions, OpenPGP, FIDO2, and hackathon tracks.
AI-written summary · about
chapters (26)
- 0:00Introduction to Ledger Agent Security
- 0:11Meet Etienne Waldron and Oscar Chaix
- 0:44Ledger Hardware and Secure Elements
- 1:51Security Risks in the Agentic Era
- 2:27Human Approval and Hardware Signing
- 3:06Ledger Wallet CLI for AI Agents
- 4:42Device Management Kit Skills
- 5:45Keyring CLI for Agent Secrets
- 7:03OpenPGP and FIDO2 Applications
- 7:52Agent Portfolio and Yield Demo
- 9:03Clear Signing and Transaction Checks
- 9:49Encrypting an Anthropic API Key
- 12:52Why Secure Secrets Matter for Agents
- 13:27Ledger Hackathon Tracks
- 14:04Giving Agents Access Without Keys
- 14:44Remote Agents and Hosted Machines
- 15:24Keeping Humans in the Loop
- 16:01New-Project Track Prizes
- 16:13Continuity Track and Integrations
- 17:42Judging Criteria and Documentation
- 18:53Developer Resources and Quick Links
- 19:27Fraudulent Contract Check Q&A
- 22:07API Key and Bug Report Questions
- 23:11Nano X and Keyring Security Q&A
- 25:07Trust Chain Authorization Model
- 28:12Closing and Final Device Notes
