1. Executive summary
Project: Green Goods
Questbook proposal: https://arbitrum.questbook.app/dashboard/?isRenderingProposalBody=true\&chainId=10\&proposalId=69548dd396f32ac6ceb9cd33
Program: New Protocols and Ideas / Orbit domain
Grant: 24,800 USDC
Impact page: Impact | Green Goods
Admin dashboard: https://admin.greengoods.app
X: @greengoodsapp
Green Goods is an offline first PWA where people doing regenerative work in the field (solar hubs, agroforestry, waste cleanup, education) log what they did with photo, video and audio evidence, have it reviewed by a trusted peer and have the reviewed record land as an EAS attestation on Arbitrum One. Gardens (local groups) get Hats Protocol roles, an ERC 6551 account, gas sponsored passkey accounts for members and Octant style vaults that can hold endowment principal.
Over this grant we took the protocol from a design document to a production deployment on Arbitrum One with real gardens using it. Between January 1 and September 5, 2026: 85 work submissions, 42 approved by peer review, from 29 active wallets across 11 gardens, with 11 community activations run in the field. We deployed 36 vaults and reached the $20K TVL target at the end of March.
We also missed things, and we say so below. Zero Hypercerts were distributed to gardens, the 150 unique user target was not reached, and no per activation onchain artifacts were issued. The most useful thing this grant taught us is that a first pilot with 14 gardens spread the team too thin and that the right next step is depth with a few rather than breadth.
2. Performance against milestones
Status uses four values: Met, Partially met, Not met, Not measured.
Milestone 1: Project Development ($4,000, paid January)
| Deliverable | Committed | Result | Status |
|---|---|---|---|
| PRD v2 | Finalized and reviewed with stewards | v1.0 PRD published, now at docs.greengoods.app/builders/specs/v1-0 | Met |
| Hypercerts and DeFi specs | Written and reviewed | Both specs delivered, linked in the milestone comment | Met |
| Updated wireframes | Admin, Hypercert minting, DeFi screens | Figma designs delivered, shipped as admin.greengoods.app | Met |
Milestone 2: Hypercerts Integration ($8,000, paid February)
| Deliverable | Committed | Result | Status |
|---|---|---|---|
| Contracts to aggregate approved work into Hypercerts | Deployed on Arbitrum | HypercertsModule, minter, marketplace adapter and exchange integration deployed on Arbitrum One | Met |
| Operator minting flow in app | Integrated for garden operators | Shipped in the admin dashboard (SDG tagging, fraction distribution, metadata), demo in the February milestone comment | Met |
Milestone 3: DeFi Integration ($6,400, paid March/April)
| Deliverable | Committed | Result | Status |
|---|---|---|---|
| Vaults on Arbitrum | 4 to 8 Octant style vaults | 28 vaults (one WETH and one DAI vault per garden, 14 gardens); first deployment of the Octant V2 factory on Arbitrum One | Met |
| Combined TVL | At least $20K | Peak recorded holdings 7.45 WETH plus 3,022 DAI on March 30, 2026, roughly $17,800. ETH price when the grant received was $3,000 which made $20K TVL more realistic when set as the goal. | Partially Met |
| Full loop: deposit, yield, garden | Live demo | One harvest on July 16 for TAS Hub recognized 10.78 DAI of yield. | Met |
| Funder and steward UI | Deposits, yield and purchases in one place | Live at greengoods.app/fund | Met |
Milestone 4: Community Activations #1 ($4,000, paid May)
| Deliverable | Committed | Result | Status |
|---|---|---|---|
| IRL activations | At least 8 | 8 activations by end of May, including six Earth Week events funded from the activation fund | Met |
| Unique users logging actions | 150 or more | 85 wallets received a Gardener Hat during the grant; 29 wallets actually submitted work. | Not met |
| Event feedback | NPS of 50 or higher | No survey was run. First feedback survey is planned for Q4 | Not measured |
| Onchain artifact per activation | Unlock keys or POAP style tokens | 0 GreenWill badges and 0 Unlock keys issued in the period. Gardener Hats and Garden NFTs were issued but those are memberships, not event artifacts | Not met |
Milestone 5: Community Activations #2 and Reporting ($2,400)
| Deliverable | Committed | Result | Status |
|---|---|---|---|
| Total activations | At least 12 | 11 activations | Partially met |
| Vaults directing yield to gardens | Yes | One Garden harvested $10 of yield | Met |
| KPI report and public write up | Published | This post, plus the impact page | Met |
| Retro and next steps | Included | Section 5 | Met |
3. What the field data shows
Work submitted by domain (approved in brackets): solar 24 (13), agroforestry 17 (12), waste 13 (8), education 12 (5), plus 19 unclassified submissions from the January deployment before domains existed (4 approved).
Gardens with the most reviewed work: TAS Hub in Awka, Nigeria (18 submitted, 14 approved, solar hub work), Growecosystems in Brazil (10, 9 approved, agroforestry), Greenpill Kenya and DeCleanup Network (waste cleanups), Rifai Sicilia (agroforestry). Eleven of 14 current gardens logged at least one submission; 3 logged none.
Activity was not steady. Submissions clustered around onboarding in January, the March push and Earth Week in April, and the August activation wave; June had a single submission. That pattern is the clearest signal in the data: gardens submit when there is a reason to gather, not as a habit. Building that habit is the work of the next phase.
Notes on definitions. Wallets are not people; a few gardens share devices. All figures are attestation creation time on Arbitrum One, not the date written on the evidence.
One garden story. The TAS Hub in Awka was the most consistent in capturing impact. They captured the development of the solar hub from a plot of land to a fully retrofitted container with solar and internet. Then once the hub was live the educational activities and workshops completed within the hub. It was a full cycle of impact that we now aim to help grow and refine as more hubs are developed throughout Nigeria.
4. Financial summary
| Milestone | Amount | Status | Main uses |
|---|---|---|---|
| M1 Project Development | $4,000 | Paid | PRD, specs, designs, project setup |
| M2 Hypercerts Integration | $8,000 | Paid | Contract and app integration, engineering, QA |
| M3 DeFi Integration | $6,400 | Paid | Vault deployment, endowment UI, DeFi research |
| M4 Community Activations #1 | $4,000 | Paid | Earth Week activation fund ($960 to gardens), onboarding calls, community support, event coordination |
| M5 Activations #2 and Reporting | $2,400 | Requested with this report | August activation wave, reporting, retro payouts to contributors |
| Total | $24,800 |
5. Retro: what worked, what did not, what changed
What worked.
The core loop is real. A gardener in Awka with an Android phone and patchy data can log solar work offline, have a peer approve it, and have the record become an attestation on Arbitrum One without ever touching gas or a seed phrase. Passkey accounts, Pimlico sponsorship, EAS, Hats roles and ERC 6551 garden accounts all held up in production across three releases (v1.0.0 March, v1.1.0 June, and v1.2.0 July). The Octant V2 DeFi vault factory on Arbitrum is a first, and other projects can use it to deploy yield generating vaults.
What did not.
Two things, and they are connected.
Hypercerts. We built the minting flow and nobody used it. Operators of small gardens with two to five active members had no reason to package a handful of approved actions into a certificate that no funder was asking for, and the yield to purchase loop needs a buyer on the other side. We now think mint from approved work was the wrong primitive for a first pilot. Hypercerts still have a place in the design, but as a way to aggregate commitments a community has actually fulfilled, once there is demand to hold them. We are not carrying a Hypercert count target into the next phase.
Breadth. We onboarded 14 gardens in nine countries because the proposal aimed for scale, and we said in February that we were “set to start onboarding 10 gardens”. four of them never submitted work, and the ones that did were sustained by one or two people. The 150 user target assumed roughly 15 active members per garden; the real number was two to five. Broad onboarding produced operational load we could not support, and the metrics show it.
What changed.
We have narrowed to three anchor gardens: TAS Hub in Awka (solar and education), Muizenberg in Cape Town (waste, with a UNICEF linked program) and AgroforestDAO in Minas Gerais (agroforestry). The scorecard for this grant stays as it was written; the narrowing is a lesson from it, not a replacement for it.
We are also adding a commitment pooling layer, following Grassroots Economics’ model, so gardens can offer and request concrete support from each other without needing cash first, with Arbitrum One as the proof and record layer.
Next steps, as options rather than commitments
First feedback survey with the anchor gardens in Q4; a defined active gardener metric per garden per month as the primary KPI; a field only counts with a published exclusion rule. With a dedicated focus on a smaller cohort we can provide the dedicated resources to onboard and support these initiatives in Green Goods solving their needs..
6. Thanks
Thank you to @maxlomu for the patience through the milestones and the direct feedback in February that reshaped how we report, to the Arbitrum DAO for providing capital needed to sustain our work, and to the gardeners in Awka, Minas Gerais, Muizenberg, Nairobi, Sicily and elsewhere whose work is the only reason any of this data exists.