@newton-xyz/policy-pack-registry

Provider-owned trust data for Newton's first-party policy packs.

This package ships two datasets and nothing else:

| Export | Purpose | |---|---| | KNOWN_PACK_IDS, KnownPackId, isKnownPackId | The first-party pack-id registry. Hand-curated (kept as a literal union so consumers can narrow/dispatch on a specific pack). | | AUDITED_POLICY_DATA | Generated shortId -> chainId -> env -> policyData audited-address map (prod env only). Written by scripts/generate-bindings.ts from this repo's AVS deployment records. |

Why a separate package

Provenance classification is split by ownership:

• The mechanism (classifyProvenance) lives in @newton-xyz/policy-core - pure compare logic, fully injective.
• The trust data (who is first-party, which address is audited) lives here, on the provider side that owns the trust.

A consumer - e.g. a curator dashboard - installs both and injects this package's data into core's classifier:

import { classifyProvenance } from "@newton-xyz/policy-core";
import { AUDITED_POLICY_DATA, KNOWN_PACK_IDS } from "@newton-xyz/policy-pack-registry";

const provenance = classifyProvenance({
  shortId,
  resolvedPolicyData,
  chainId,
  env,
  auditedRegistry: AUDITED_POLICY_DATA,
  knownIds: KNOWN_PACK_IDS,
});

Regeneration

known-pack-provenance.generated.ts is generated - edit the upstream deployment records and run pnpm gen:bindings from the repo root. known-pack-ids.ts is hand-curated: add a new pack's short id in the same PR that adds the pack.