Roadmap¶
What exists today is a working testnet protocol with an SDK, an indexer, a reference app and three partner integrations. The plan from here is about integrations first and features second.
Next 30 days¶
- npm release of
@ibxlab/mandate0.1.0 with the quickstart, the agent runner and the LLM tool example. - Hosted indexer on Envio Cloud, wired into the reference app and the CRE workflow's
indexerUrl. - CRE workflow deployed to a DON (production target and forwarder are already in
cre/project.yaml), replacing the simulator runs; the LLM leg enabled with a hosted key. - Integration reviews: the first five integrators get an IBX Lab review of their mandate parameters and agent loop.
Next 90 days¶
- Wallet-side integrations. A MetaMask agent-wallet plugin and a native Mera wallet capability that issue mandates from the wallet UI, so agent platforms do not need their own principal screen.
- Framework adapters. Published tool definitions for the common agent runtimes (the
execute_under_mandatetool in the SDK examples is the template), so an agent gains mandate-bounded execution by adding one tool. - More venues. Real DEX and perp venues as targets, with venue-specific
amountmeasurement hints. - Reputation consumers. A reference "choose an agent" query over ERC-8004 plus Mandate evidence for marketplaces and routers.
- Mainnet deployment of the contracts once two external integrations are live on testnet.
Protocol work under consideration¶
- Allowance-aware spend: optional per-target allowance caps set at grant time, closing the out-of-band pull path described in Security → Scope notes.
- Multi-principal mandates (a DAO or a multisig as principal) using the existing ERC-1271 path.
- Mandate templates: signed, reusable parameter sets that a wallet can offer as presets.
What will not change¶
The registry enforces; agents cannot self-attest; passkeys never leave the device; policies are ciphertext to everyone but the principal; and any of the pieces around the contracts can be replaced by an integrator.