# buildPolicyParams Turn a plain config into the policy params an agent's module enforces. You describe each agent's per-token limits and recipients; the helper produces the canonical, fail-closed policy. ```typescript import { buildPolicyParams } from "@newton-xyz/isaac"; const params = buildPolicyParams({ "0xAgent...": { "0x036CbD...": { max_daily_spend: 500_000_000, // 500 USDC at 6 decimals allowed_recipients: "0xApi1...,0xApi2...", }, }, }); ``` `buildPolicyParams(agents)` takes a map of agent address to a map of token address to its config, and returns params with `default_policy: "deny"` — any agent or token not listed is denied. ## Per-token config | Field | Meaning | |---|---| | `max_daily_spend` | Cap per day in token base units. `0` means unlimited. | | `allowed_recipients` | Comma-separated allowlist of recipient addresses. `"*"` allows any; `""` denies all. | | `vendors` | Optional per-recipient caps (see below). | | `verified_only_mode` | Optional. When `true`, restrict recipients to the `verified_recipients` set (the verified marketplace tier). | | `verified_recipients` | Optional. The verified-tier recipient addresses (lowercased). | ```typescript const params = buildPolicyParams({ "0xAgent...": { "0x036CbD...": { max_daily_spend: 0, // unlimited daily total allowed_recipients: "*", vendors: { "0xApi1...": { max_per_tx: 1_000_000, max_daily_spend: 10_000_000 }, }, }, }, }); ``` `vendors` adds a per-recipient cap on top of the token-wide limit: `max_per_tx` bounds a single payment and `max_daily_spend` the cumulative daily total to that recipient. ## Applying it Pass the result to the platform API or the SDK to register it for the agent: ```typescript await platform.updatePolicyParams(vaultAddress, agentAddress, params); ``` ## See also * [Give an Agent a Budget](/guides/agent-budget) — the guided walkthrough. * [PlatformClient](/reference/sdk/platform-client) — apply policy over HTTP.