capability-tokens
Grantz provides capability token primitives.
Capability control plane for AI agents — agents never touch a credential. Scoped tokens (RFC 8693/9396), policy enforcement, server-side credential injection, signed audit chain.
Each agent run gets its own short-lived attested credential. A stolen token is not enough.
Cryptographic verification for AI agent actions — ECDSA-secp256k1 + RFC 6979 signed Action Receipts (v0.1), multi-dimensional Trust Vector, capability tokens (JWT-shaped + Biscuit-style attenuation), UETA §10(b) undo. 29 MCP tools. A2A v1.0 + ERC-8004 format compatible. Receipt-batch Merkle roots anchored on Base mainnet. Starting with code.
