The Ledger
ERC-8475 Proposes Per-Transfer and Window Caps for ERC-20 Tokens
ERC-8475 proposes an optional ERC-20 extension that limits individual outgoing transfers and total spending within fixed block windows. The draft was created on October 30, 2026, and a November 1 commit assigned its number; proposed replacement caps require an explicit application after the delay.

“The engineering requirement is to bound token outflow even when a transfer satisfies normal signature and allowance checks.”
Fernando Díaz Toledano, author of ERC-8475, writes in the proposal.
Two caps for outgoing movements
Fernando Díaz Toledano’s ERC-8475 proposes an optional ERC-20 extension with a maximum amount for an individual outgoing movement and a maximum aggregate amount per fixed block window. The draft was created on Oct. 30, 2026, and a Nov. 1 commit assigned it ERC-8475; the pull request remains a draft. Díaz Toledano opened a Fellowship of Ethereum Magicians discussion on Nov. 2, seeking feedback on the design.
Under the proposal, a holder would enroll by setting a per-transfer cap, an aggregate window cap, the window length in blocks, a delay for replacing caps and an independent guardian. Enrollment is optional; a holder that has not enrolled would be unrestricted by the extension. The specification describes an interface and accounting rules, rather than requiring a particular implementation. Amounts are stated in the token’s smallest units, with no price conversion.
The draft cites the Binance incident reported on Oct. 29, 2026, as motivation for controls on unauthorized withdrawals. It says the investigation was preliminary when the working paper was prepared and makes no claim about the precise exploit or how much the proposal would have saved.
Shared accounting across debit paths
The proposal assigns each enrolled holder a counter for positive outgoing movements during the current window. Before a movement, an implementing token would check that the amount is no greater than the per-transfer cap and that adding it to the holder’s existing window spending does not exceed the aggregate cap. Spending equal to either cap is permitted. An amount beyond either limit must be rejected by reverting the transaction.
The counter applies to the debited holder across transfers, transferFrom calls, operator and signature-authorized spending, burns, self-transfers and batch operations. For a transaction containing multiple transfer operations, the per-transfer cap applies to each movement, while the window cap applies to their aggregate. Incoming transfers and mints do not use or refill the recipient’s quota. Burns and self-transfers count as outgoing movements; creating or revoking an approval does not reserve quota.
Fixed windows and policy changes
Windows are calculated from block numbers and the configured window length, rather than as rolling time intervals. The draft warns that a holder could spend the full quota just before one window ends and another full quota just after the next begins. Its stated guarantee is per holder, token contract and chain—not a fiat-value or exchange-wide limit. The proposal contrasts fixed windows with sliding windows, which would require historical accounting or a more complex bounded design.
A holder or guardian could immediately tighten either cap, including to zero. Only the holder could schedule replacement caps, which become eligible for application after the configured delay; they do not take effect automatically. Anyone could apply a matured change, while the holder or guardian could cancel a pending change. Applying or cancelling a policy change would preserve usage rather than reset the counter. The guardian could stop positive outflows by setting both caps to zero, but would not gain the right to transfer the holder’s assets.
The proposal preserves the ERC-20 ABI, but says opted-in holders could encounter transfer reverts despite having sufficient balance and allowance. Wallets and other systems that handle transfers, including routers, liquidation systems, batched transfers and exchange withdrawals, may need to account for changing available quota. The proposal is still a draft under discussion, not an adopted standard.
Sources
- Original source Ethereum ERCs on GitHub github.com
Other pages read
- Fellowship of Ethereum Magicians ethereum-magicians.org
Related stories
- Circle Details CCTP Routes for Non-USDC ERC-20 Tokens
- Ethereum Researcher Proposes Cheaper Replay-Protection Storage
- Lido proposes trustless and trusted MEV tracks
- Andrey Chmora Proposes Stealth ABBE for Private Ethereum Contracts
- Preprint: ML-DSA-44 Per-Record Design Costs 8.7× ECDSA Even With Proposed Precompile