refactor(token): CFT supply extension and integration composition

[0xisk]

Types of changes

What types of changes does your code introduce to OpenZeppelin Midnight Contracts?
Put an x in the boxes that apply

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation Update (if none of the other choices apply)

Fixes: N/A. Layout proposal for OpenZeppelin#653 (Confidential Fungible Token, tracking OpenZeppelin#569), targeting add-cft-2 so it can merge into that PR directly.

What this proposes

Restructure the CFT layout so the shipped library is one token module plus one optional supply extension, with the assembled token living in the integration test suite:

src/token/ConfidentialFungibleToken.compact                   (renamed from ConfidentialFungibleTokenCore;
                                                               gains _mint/_burn/_burnFrom building blocks)
src/token/extensions/ConfidentialFungibleTokenSupply.compact  (supply tracker: _totalSupply,
                                                               _addSupply/_subSupply, totalSupply())
test/integration/_mocks/ConfidentialFungibleTokenPublicSupply.compact
                                                              (TEST-ONLY top-level contract composing both)
test/integration/fixtures/confidentialFungibleTokenPublicSupply.ts
test/integration/specs/confidentialFungibleToken/*.spec.ts    (7 case specs)

Commit by commit:

1. refactor(token): drop Core suffix from the CFT module — the NST family keeps a Core because two wrappers share it; CFT has one flavor, so the module consumers import is just ConfidentialFungibleToken.
2. feat(token): add _mint/_burn/_burnFrom building blocks — thin intent-named aliases over _credit/_debit/_spendEscrow (which transfer/_move keep using internally). No supply accounting; a composing contract gates them and pairs them with the tracker.
3. feat(token): add ConfidentialFungibleTokenSupply extension — mirrors NativeShieldedTokenSupply: imports no token module, exposes the scalar accounting blocks plus the getter; the original error strings are preserved.
4. refactor(test): compose CFT public supply in the integration suite — the assembled token (former ConfidentialFungibleTokenPublicSupply module) integrates two production pieces, so it moves out of src/ and becomes a top-level TEST-ONLY contract driven directly through its artifact; the unit mock is deleted and the simulator becomes an integration fixture with shared helpers.
5. test(integration): add the CFT composition spec suite — the former single test file split into per-case specs (metadata, mint, burn, burnFrom, supplyConservation), following the specs/<topic>/<case>.spec.ts layout, with new coverage for the conservation invariant (sum(balances) == totalSupply: only paired mint/burn/burnFrom may move the total; transfer, approve/transferFrom, _move, and sweep never do).
6. test(integration): add CFT privacy and concurrency specs —
   • privacy.spec asserts observer-side properties from the public ledger only: balances/pending/escrows stay ciphertexts, the approve cap never appears in clear, equal amounts are unlinkable, transfer reveals the counterparty graph but no amount, mint/burn disclose amounts through the supply delta by design, and hostile wallets (wrong EK, overstated plaintext or allowance) fail the witness-binding checks.
   • concurrency.spec documents the same-block limitation from the module header by proving its cause deterministically: with the witness seed pinned, an identical credit replayed on a changed recipient pre-state yields a different transcript (two transactions proven against one pre-state conflict), credits to distinct recipients touch disjoint cells, and sweep contests the same pending cell an incoming credit writes.

Testing notes

• Every touched contract compiles (compact compile --skip-zk, artifacts verified), tsc --noEmit and biome ci are clean.
• The spec suite runs under yarn test:integration.
• Note: moving the composed suite out of src/** removes it from the live-test matrix (the unit-live project only globs src/**/*.test.ts).

PR Checklist

• I have read the Contributing Guide
• I have added tests that prove my fix is effective or that my feature works. See GUIDELINES.md#testing for more information.
• I have added documentation for new methods or changes to existing method behavior. See GUIDELINES.md#documentation for more information.
• CI Workflows Are Passing

Further comments

• Why supply-as-extension: the base stays supply-neutral (conserving surface + intent-named blocks), matching the NST precedent (extensions/NativeShieldedTokenSupply) and leaving room for a confidential-supply sibling (ElGamal cell + auditor viewing key) without touching the base.
• Why the assembled token is test-only: an ungated mint is a faucet, so shipping it as an importable "batteries-included" module invites unsafe deployments. As an integration mock the DANGER surface disappears, and the composition still gets a full spec. Whether the library should also ship a gated deployable preset (e.g. mint behind Ownable) is worth discussing.

[0xisk]

Superseded by #10, which keeps the refactor plus the composed integration contract and defers the spec suite to a follow-up PR.