All Core Devs - Execution (ACDE) #248, October 22, 2026

Agenda

  • Glamsterdam
  • Hegotá
    • dynamic base fee calculation for shorter slots: EIP PR by @barnabasbusa, awareness only, please review
    • CFI priority ordering: proposed list
      • aiming for consensus on the call, async follow-up if needed
    • next devnet: bogota-devnet-2 vs hegota-devnet-0
      • bogota-devnet-2: one more EL-only devnet, see comment by @wolovim on the frames breakout outcome
        • would need scoping (spec changes + new EIPs), incl. decision on single max fee for frames
      • hegota-devnet-0: combine bogota-devnet-1 with heze-devnet-3, see comment by @barnabasbusa re pre-Devcon launch target
        • CL-side scope to be decided on ACDT next week (ACDC as fallback)
    • SFI frames companion EIPs (@lightclient)
      • EIP-8250: Keyed Nonces for Frame Transactions
      • EIP-7906: Transaction Assertions via State Diff Opcode
      • EIP-8272: Recent Roots for Frame Transactions
    • additional CFI decisions
  • Misc
    • modify ACDT time to 90 min (@barnabasbusa)

Meeting Time: Thursday, October 22, 2026 at 14:00 UTC (90 minutes)

GitHub Issue

1 Like

Video, transcript & chatlog

News coverage

Resources

Meeting Summary:

The AllCoreDevs meeting focused primarily on security concerns related to the Amsterdam hard fork (EIP-4844) and its potential impacts on smart contracts. Hubert from the chain security team presented findings showing that certain contracts could become unsafe due to changes in gas handling and pricing, particularly highlighting how contracts with permissionless execution could be vulnerable to new attack vectors where attackers could front-run transactions with insufficient gas to execute self-destruct operations. The team identified 13 major DeFi projects affected by these issues and are currently in responsible disclosure with them. The group decided to cancel the upcoming Hoodi hard fork and delay the mainnet rollout indefinitely until a more comprehensive analysis of impacts can be completed and affected projects have time to upgrade their contracts. They also discussed the process for moving EIPs from Proposed to Final Implementation (CFI to SFI), with consensus that CFI should not be considered a default for shipping changes and that each EIP’s value to Ethereum should be carefully evaluated. Finally, they decided to proceed with launching Bogota DevNet 2 next week with four specific EIPs (including setCodeFrom) despite some implementation challenges, while planning to focus on cross-layer interactions between fossil and frames for future DevNets after DevCon.

Click to expand detailed summary

Hubert presented security concerns regarding smart contracts affected by the Amsterdam hard fork, specifically focusing on how changes to gas handling and refund mechanisms could make previously safe contracts vulnerable. He identified 13 major DeFi projects affected by these issues and explained that the problems arise from incorrect developer assumptions about gas behavior rather than inherent security flaws. The team discussed three potential responses: addressing the issues before the December mainnet rollout, delaying the mainnet with the current specifications, or modifying the Glamsterdam specs. The group agreed that the immediate priority should be conducting extensive analysis to fully understand the scope of affected contracts before making decisions about mainnet timing or specifications.

The team discussed whether to proceed with the Hoodi testnet fork, with concerns raised about the potential need for specification changes and the impact on testing capabilities. Several participants, including Potuz and Daniel, argued against proceeding with the fork due to the risk of having to abandon testnets on incompatible specifications, while others like Łukasz suggested maintaining both Glamsterdam v1 and v2 versions to allow for comparison. The group agreed to halt the Hoodi fork temporarily to conduct a more thorough analysis of the impacts and determine whether specification changes are necessary before making decisions about mainnet rollout dates.

The team discussed concerns about potentially delaying the Hoodi fork due to potential contract compatibility issues requiring specification changes. Hubert reported that while some high-value contracts were identified as potentially affected, many appear upgradable or migratable, though a comprehensive analysis would take 2-3 weeks. The group agreed to halt the Hoodi fork immediately, with no specific new date set, pending further analysis of the impact on large contracts and potential specification changes needed. Barnabas confirmed that rolling back updates is feasible, though some clients might face difficulties with database updates.

The team decided to delay the Hoodi hard fork due to security concerns with the Glamsedam specifications, particularly regarding contract repricings. Dima from Lido confirmed they could handle the rollback of recent protocol changes, though they couldn’t guarantee all Lido stakeholders would be updated in time. The next steps include creating a blog post to notify the ecosystem about reverting to pre-Glamsedam versions and having client teams push new releases that revert the changes. The team will conduct analysis over the next few days to assess the impact of the delays and determine if any spec changes are needed, with more details expected to be presented at the ACDT meeting.

The team discussed concerns about breaking user contracts due to upcoming hard fork changes, particularly regarding gas costs and execution layer mutability. They decided to cancel the Hoodi fork indefinitely and halt the Mainnet fork scheduled for the following week, with no new timeline set. The group agreed to create a blog post recommending that clients roll back updates or wait for new releases, which will be updated as more clients release new versions. They also approved extending future ACDT meetings to 90 minutes and discussed prioritizing EIPs for testing, though no final decisions were made on the specific ordering of EIPs.

The team discussed the process for moving EIPs from CFI (Considered for Inclusion) to SFI (Shipped for Implementation), with consensus that CFI should not be considered a default path to shipping and should require more scrutiny. They decided to launch Bogota DevNet 2 next week with four specific EIPs: setCodeFrom, EIP-8151, EIP-4758, and EIP-8253, targeting implementation by Nethermind, Geth, and Ethereum clients. The team also agreed to postpone the next ACD meeting by four weeks due to the upcoming DevCon conference.

Next Steps:

  • Hubert: Continue the analysis of the security impact of Amsterdam repricing EIPs on smart contracts, including identifying affected projects and coordinating responsible disclosure.
  • Ansgar: Coordinate the publication of a blog post to notify the ecosystem about the security concerns and the decision to delay the Hoodi hard fork.
  • EL client teams: Push a new release that reverts the Amsterdam changes and removes the Hoodi hard fork.
  • Barnabas: Prepare the PR for the client release to revert the Hoodi hard fork.
  • Dima: Coordinate with Lido operators to roll back clients to the pre-Amsterdam version to support network stability.
  • Ansgar: Postpone the mainnet rollout of Amsterdam until there is a clear understanding of the security impacts and affected contracts.
  • Ansgar: Cancel the Hoodi hard fork and do not set a new date until further notice.
  • Barnabas: Propose extending the ACDT meeting time to 90 minutes for future calls.
  • Barnabas: Merge the dynamic base fee calculation adjustment PR into the EIP.
  • Ansgar: Use the proposed EIP priority list as a draft for the upcoming hard fork and iterate on it asynchronously.
  • Ansgar: Launch Bogota DevNet 2 next week with the agreed-upon scope of four EIPs (including setCodeFrom and 8151).
  • Ansgar: Cancel the ACDT meeting scheduled for two weeks from now due to DevCon.

Recording Access:

YouTube Stream Links: