Skip to content

Ethereum security initiativeDraft proposal. Not submitted.

Impact

What we will measure, starting from zero.

Every milestone has a test. An independent reviewer, named before work starts, runs it. Until that reviewer confirms, nothing counts and nothing is paid.

Paid so far: $0 of $100,000 proposed

ProposedDraft proposal. Targets are proposed. Nothing on this page counts work that has not happened.

  1. MilestoneM1
    What is measuredThe four-layer method, payment rules and gap-only rules, published
    What countsThe reviewer reruns the demo script at the pinned commits and gets the same function inventory.
    TodayNot yet checked by a reviewer
    Releases$12,000
  2. MilestoneM2
    What is measuredFour-layer reviews of open-source mainnet protocols, completed and published
    What countsEach record matches its reviewed commit and its payment records.
    TodayNone yetTarget: Three or more
    Releases$28,000
  3. MilestoneM3
    What is measuredGap-only follow-ups after real upgrades, and accepted outside evidence items
    What countsA follow-up lists the change set it covered. An evidence item counts after two independent reviewers release it.
    TodayNone yetTarget: Three or more follow-upsTarget: 250 evidence items
    Releases$20,000
  4. MilestoneM4
    What is measuredOutside protocols above the value bar that publish a review record
    What countsThe protocol holds more than $100 million on Ethereum mainnet at the pinned block, by the published read-only query.
    TodayNone yetTarget: Three or more
    Releases$40,000Adoption

Empty boxes are targets. None is filled yet.

01Measured in public

The measures

Initiative measures, today's value and how each is checked
Released audits on the public explorerLive0Every funded review releasedReviewer reads the explorer and each release recordPublic explorer, read now
Protocols analyzed, and their stateLive
  • Aave V4Not verified
  • Morpho BlueNot verified
—Explorer state per protocolPublic explorer, read now
Four-layer reviews completed and publishedProposedNone yetThree or more (M2)Published records checked against the reviewed commitNo public four-layer review exists yet.
Gap-only follow-ups publishedProposedNone yetThree or more (M3)Records list the change set they covered—
Accepted outside evidence itemsProposedProgram not started250 (M3)Public list of accepted items and their two reviewers—
Researchers paid through the initiativeProposedNone yetEvery funded reviewPayment records shared with the reviewer under confidentiality—
Outside protocols above the value bar that adoptedProposedNone yetThree or more (M4)Reviewer reruns the value-bar query at the pinned blockMethod below
Snap available in ordinary MetaMaskLive
Checking install status
— (depends on MetaMask)MetaMask's public allowlistProof of Audits runtime config, read now

We do not show the auditor leaderboard here. Its profiles are imported from public contest records and are not outcomes of this initiative.

02The value bar

How we tell a real adopter from a small one

An adopting protocol must hold more than the value bar on Ethereum mainnet at a fixed block. A read-only script measures this. It sends only read calls, prices each token with Chainlink feeds read at the same block, and checks each feed's description and token symbol onchain.

Value bar: $100 million, measured at a pinned block.

FIG 2Value-bar query, three case studiesEthereum mainnet · block 26,378,312

The method, tested on three protocols at block 26,378,312

The method, tested on three protocols at block 26,378,312
Aave V4 SpokeSupplied assets reported by the Spoke$347,249,974, plus 6 reserves the query could not price
Morpho BlueToken balances held by the singleton contract$2,296,489,149
Uniswap V3Balances of 9 pools found through the factory$317,137,108

Source: Proof of Audits value-bar query, run 12 November 2026 against block 26,378,312 with 157 read calls. These protocols are case studies used to test the method. They are not adopters, partners or customers.

Still to decide: for lending markets, whether the bar uses token balances held or the assets the protocol itself reports as supplied. The table shows the measure used for each case.

What would not count as impact

  • 01Page views, sign-ups or wallet connections.
  • 02Reviews of code that is not deployed on Ethereum mainnet.
  • 03Records released by Proof of Audits staff alone, without the two independent reviewers the release rule requires.
  • 04Protocols below the value bar, for the adoption milestone. Their reviews still count for Milestones 2 and 3.
  • 05Anything produced only by AI.

Read the proposal

This page is updated from the public explorer each time it loads. When the first funded review is published it appears in the first row.