Do the check in the order that fails fastest. Read `DefaultAccountState`; on both mints it says `Frozen`, so that read tells you nothing [1]. Read the freeze authority and resolve its owner. On an sRFC-37 mint it is a MintConfig PDA owned by the Token ACL program at `TACLkU6...52TP`, and from that point freeze and thaw run through code rather than the issuer's wallet [11]. Then read the MintConfig's gate program, which answers exactly one question: `can_thaw_permissionless(wallet)` [12]. A honeypot operator cannot walk that path and keep the trap, because handing thaw to a published gate is handing away the ability to decide who is allowed to sell [20].
The mint's metadata declares the gate under a `token_acl` key so wallets and SDKs can find it [12]. That key is written by the issuer. The authority chain is what the token program enforces, which is why the author's stated method is verification against the chain rather than anyone's metadata [17].
Score zero is a claim about a heuristic, and per the post it came from the author's own testing rather than a published benchmark [2]. The heuristic reads two fields: default-frozen, plus a live freeze authority [1]. For that score to transfer to an RWA integration, freeze authority would have to always sit in a wallet with discretion. On mainnet since March 2026 it does not, and institutional money is already on the other side of it [6] [21].
Issuers land in this shape because the alternatives cost more. Transfer hooks run issuer code on every transfer, but every DEX, wallet and protocol touching the token has to implement the hook interface, so composability dies at the venues that create liquidity [7]. Manual freeze-and-thaw keeps standard transfers, but each new holder waits for an issuer co-signature and onboarding becomes a support ticket [8]. Token ACL keeps the freeze lever the token program already enforces everywhere and makes thawing self-service against a published rulebook [9]. The holder creates the frozen account, calls the permissionless thaw, and the account unlocks in the same transaction if the gate says yes [13].
Scale argues for doing this by hand right now. The guide with mainnet program addresses went up on January 12, 2026, and the first mainnet transaction landed 53 days later on March 6 [4] [5] [19]. Total transactions ever: 514 [5]. An honest allowlist of Token ACL mints would currently fit on one screen.
The maintenance cost sits in the swap. sRFC-37 specifies only the interface between Token ACL and gates, so an issuer can replace the reference Allow/Block List program at `GATEzz...iULz`, which ships with four modes, with an on-chain KYC registry or an oracle-driven sanctions feed, and the mint itself never changes [14] [15]. Anyone who integrated against allowlist semantics last quarter may be integrated against something else now, with no on-chain event at the mint to notice. The French fund mint the author decoded also carries pausable transfers, a permanent delegate, permissioned burn and a mint close authority [16]. Each of those is a separate authority and a separate question, and the freeze-authority check answers none of them. The author's framing of the choice facing scanners is blunt: learn the standard, or misread an entire asset class [18].
