v1.0 — Hypercerts, Octant, Gardens, Cookie Jar

Live

1. Overview

1.1 One-Liner

Green Goods is a Regenerative Compliance and Local First Impact Reporting Platform that enables verifiable tracking of ecological actions, tokenization of impact via Hypercerts, and capital formation through Octant Vaults and Revnets.

1.2 Context

The global landscape for environmental finance is currently navigating a profound structural transformation, precipitated by a dual crisis of credibility in voluntary carbon markets (VCM) and an urgent, unmet demand for verifiable data in government-funded climate and ecological initiatives. The prevailing paradigms of "Monitoring, Reporting, and Verification" (MRV) are failing to scale, burdened by high-cost manual audits that exclude the vast majority of grassroots regenerative projects from accessing capital markets. Traditional certifiers rely on extractive consulting models where a single verification event can cost upwards of $5,000, rendering the model economically unviable for smallholder farmers, community gardens, and hyper-local ecological stewards.

Simultaneously, governments are deploying historic levels of capital but face immense legislative pressure to prove results in "disadvantaged communities" (DACs). This creates a "Verification Gap." Green Goods bridges this gap by pivoting to Compliance-as-a-Service (CaaS), operationalizing the "Regenerative Stack" to connect physical ecological labor with onchain capital.

1.3 Why Now

• The Yield Economy Has Matured — Octant V2 enables endowment models where principal is preserved.
• AI at the Edge — LLM costs dropped 90% in 2025, making Agri-Advisor agents viable at under $0.01 per interaction.
• Greenpill Garden Season One — 8+ Gardens are ready to adopt the protocol.

1.4 Competitive Positioning

Green Goods differentiates from Silvi (trees only, centralized, enterprise pricing), GainForest (requires satellite, high technical barrier), and Regen Network (Cosmos-based, methodology-heavy) through multi-domain coverage, Ethereum-native DeFi integration, accessible MDR methodology, and community governance via Fractal Gardens and Conviction Voting.

Key moats: Network effects from Greenpill chapters, switching costs from attestation history, data moat from accumulated actions, and community ownership through local tokens.

1.5 Target Release

• Release: Green Goods Protocol v1 (Beta Candidate)
• Season: Q1 2026 (Transition from Alpha to Beta)
• Network Deployment: Arbitrum One and Ethereum

1.6 Owners

• Product and Engineering Lead: Afolabi Aiyeloja
• Design and QA Lead: Nansel Rimsah

---

2. Goals and Success

The architecture of Green Goods v1 is driven by three primary strategic objectives designed to bridge the "Verification Gap" and create a closed-loop regenerative economy. These goals move beyond simple metrics of usage to encompass the structural integrity of the "Regenerative Stack."

2.1 Goals

Goal 1: Capital Formation (High Sustainability)

The economic goal is to enable Capital Formation rather than simple circulation. By utilizing Octant Vaults, the protocol creates "Local Endowments" where low-risk staking yields systematically purchase Hypercerts, establishing a "Risk-Free Impact Floor" for ecological labor. Simultaneously, the integration of Juicebox Revnets enables "The Growth" layer, allowing communities to grow revenue and manage treasuries where tokens represent governance rights and a claim on future success, rather than just donation receipts.

Goal 2: Impact Reporting Accessibility (High Reach)

The accessibility goal is to ensure the protocol is available to the "Last Mile" of users, regardless of technical literacy or connectivity. This involves expanding the interface beyond the browser to include AI Agents that interact via WhatsApp and SMS. This allows even the most rural Gardeners to interface with the protocol using natural language to log work, query earnings, and receive alerts. We also plan to address the ease of aggregating impact into a single asset and make it simple and friendly for individuals to mint Hypercerts for impact certification.

Goal 3: Onchain Governance and Reputation (High Coordination)

The governance goal is to implement a robust badging and reputation system built on Unlock Protocol (GreenWill) and Hats Protocol. These tools serve as the permission layer for the Gardens platform, ensuring that governance is driven by those who create value. By using Conviction Voting, the protocol empowers reputation holders to signal distinct priorities: deciding on hypercert yield allocation and prioritizing specific garden actions.

2.2 Success Metrics

Metric	Definition	Target
Total Verified Impact Value (TVIV)	Cumulative dollar value of all Hypercerts minted, sold, or held within the protocol ecosystem	$40,000 by end of Q1 2026
Vault TVL	Total value locked in Octant Vaults	$20,000
Active Gardeners	Unique wallet addresses (including Smart Accounts via AI Agent) submitting verified actions in 30-day window	150 users on a rolling 30-day basis
Reputation Density	Number of active users holding a GreenWill "Steward" or "Sprout" badge via Unlock Protocol	32+ users
Verification Efficiency	Percentage of submitted actions processed (Approved/Rejected) by Evaluators/Operators within 72 hours	>90%
Conviction Participation	Percentage of token holders actively signaling/voting on yield allocation via Conviction Voting	>48%
Hypercerts Minted	Number of ERC-1155 impact certificates created from verified actions	12+ by end of Q1 2026
Gardens Active	Number of distinct Gardens using Arbitrum for impact reporting	8+ by end of Q1 2026
CIDS Compliance Rate	Percentage of actions structured per Common Impact Data Standard	80%+ per cohort
Karma GAP Updates	Number of verified actions pushed to Karma GAP as milestone updates	100+ by end of Q1 2026

2.3 Guardrails

• Redemption Rate (Revnet): The cash-out tax on Garden Tokens must remain between 80-90% to prevent treasury depletion.
• Data Integrity Score: The rejection rate of submissions flagged by IoT/Silvi oracles must not exceed 15%.

---

3. Users and Use Cases

The product architecture serves a 5-sided marketplace, ensuring checks and balances between labor, management, capital, verification, and community beneficiaries.

3.1 Primary Users and Personas

Persona A: The Gardener (The Laborer)

• Profile: Field workers in bio-regions (e.g., Nigeria, Brazil). May operate in low-connectivity zones.
• Motivation: Livelihood security and community improvement.
• Key Interface: AI Agent (WhatsApp/SMS) for reporting; PWA for deep management.
• Definition of Success: Successfully logging work via a text message or photo and receiving payment notification without managing a crypto wallet.

Persona B: The Garden Operator (The Steward)

• Profile: Local chapter administrator or NGO program manager.
• Motivation: Funding sustainability and coordinating local labor.
• Key Interface: Admin Dashboard.
• Definition of Success: Efficiently verifying queues of submissions, minting Hypercerts, and managing the Garden Treasury.

Persona C: The Evaluator (The Verifier)

• Profile: Domain experts (e.g., sustainability experts, material scientists) or specialized Oracles.
• Motivation: Ensuring integrity, preventing greenwashing, and maintaining professional reputation.
• Key Interface: Admin Dashboard (Evaluator View).
• Definition of Success: Reviewing completed Impact Reports or Assessments and cryptographically certifying the quality of the data.

Persona D: The Funder (The Capital Provider)

• Profile: Institutional donors, Government Agencies (e.g., California Natural Resources Agency), or Octant stakers.
• Motivation: Verified Impact (MRV), compliance with legal mandates, and capital efficiency.
• Key Interface: Admin Funder Dashboard and Octant Vaults.
• Definition of Success: Depositing assets into a vault and receiving an automated Karma GAP report showing yield utilization.

Persona E: The Community Member (The Beneficiary)

• Profile: Local residents living in the bioregion affected by the Garden's work.
• Motivation: Ensuring the work actually benefits the local ecosystem and holding the Garden accountable.
• Key Interface: Public Signal Feed and Gardens Conviction Voting.
• Definition of Success: Ability to "Signal" or "Attest" that the work exists and is healthy, providing a "Proof of State" from the ground up, and signaling prioritization for future actions.

3.2 Top Jobs-to-be-Done (JTBD)

JTBD	User	Context	Job	Outcome
Low Bandwidth Submission	Gardener	In the field with no data plan	Text a photo to the AI Agent	Labor logged even without a high-end smartphone
Impact Report Aggregation	Operator	Quarter ends and reporting is due	Aggregate all verified actions into a single Impact Report	Mint a Hypercert bundle and sell it to the Octant Vault
Principal-Preserving Funding	Funder	Want to support a region without losing principal	Stake ETH in an Octant Vault	Yield acts as a perpetual endowment for the local community
Impact Evaluation	Evaluator	A Garden claims "Biodiversity Increase"	Review aggregated data and sensor logs	Cryptographically sign an attestation verifying the claim
Community Prioritization	Community Member	A Garden proposes a new project	Signal priority via Conviction Voting	Yield flows to the actions that matter most

3.3 Use Cases

• PWA Impact Reporting: An Operator uses the PWA to aggregate 500 individual tree planting actions into a "Q1 Reforestation Report." The system formats this data according to the Common Impact Data Standard (CIDS) for Karma GAP integration.
• Impact Reporting Certification: An Evaluator receives the "Q1 Reforestation Report." They analyze the attached Silvi sensor data and satellite imagery. Satisfied with the quality, they sign the report using their Hats Protocol "Verifier" authority, certifying it for minting.
• Garden Revenue Tokenization: A Garden mints a Hypercert based on the certified report. They list this Hypercert on the Revnet. The Octant Vault automatically purchases it using yield, and the revenue flows into the Garden's treasury, backing the value of the local Garden Token.
• Prioritization Signaling: Community members stake their reputation (GreenWill Badges) on specific actions within the Garden's governance page, signaling that "Water Conservation" is a higher priority than "Planting" for the upcoming season.

3.4 Action Domains and Target Outcomes

Green Goods is not a generic impact tracking tool. It is purpose-built to capture and verify specific categories of regenerative work emerging from Greenpill Network chapters and aligned initiatives. The protocol optimizes for five core action domains that reflect where local communities are actively creating value.

Domain 1: Solar Infrastructure (DePIN Pillar)

Primary Chapter: Greenpill Nigeria through the Tech and Sun program.

Key Outcome: Energy access plus staking yield at Nigerian universities.

The Tech and Sun initiative deploys solar-powered container hubs at University of Nigeria Nsukka and Nnamdi Azikiwe University. These hubs provide reliable electricity, Starlink internet, Ethereum staking infrastructure, and co-learning spaces for Web3 education. Green Goods captures each phase of hub deployment and ongoing operations, creating an immutable audit trail for funders and enabling yield routing from staking rewards.

Action Lifecycle:

1. Site Acquisition — Documents land/space agreement with university through photos of location, signed agreement scan, and GPS coordinates.
2. Container Procurement — Records purchase and delivery of 20ft shipping container through delivery receipt and photos.
3. Solar Installation — Captures panel, inverter, and battery setup through photos and Switch meter data.
4. Connectivity Setup — Documents Starlink deployment through speed test results and subscription proof.
5. Node Deployment — Records Ethereum staking node setup through validator address and deposit transaction.
6. Hub Launch — Documents official opening through event photos and attendance count.
7. Ongoing Operations — Captures daily usage through user count, energy generated (kWh), and session duration.
8. Workshop Hosted — Records educational events through attendance list and participant feedback.

Impact Metrics:

Metric	Description
Energy Generated	Total kWh produced
Users Served	Unique individuals
Node Uptime	Percentage availability
Yield Generated	ETH staking rewards
Workshops Hosted	Count of sessions
Cost per kWh	Efficiency metric

Domain 2: Waste Management (Environment Pillar)

Primary Chapters: Greenpill Cape Town, Greenpill Ivory Coast, Greenpill Koh Phangan, and Greenpill Rio Brasil.

Key Outcome: Circular economy systems reducing environmental waste.

Waste management is being tackled across multiple chapters with varying approaches. Cape Town focuses on community beach cleanups for coastal ecosystems. Ivory Coast addresses urban waste collection in informal settlements. Koh Phangan optimizes island waste for resort areas and marine debris. Rio Brasil builds systematic waste management for urban and suburban circular economy. Green Goods standardizes capture across these diverse approaches while allowing local customization of waste categories and processing methods.

Action Lifecycle:

1. Area Assessment — Surveys location to understand waste problem through GPS, photos of waste accumulation, and area size estimate.
2. Cleanup Event — Collects and removes waste through before/after photos, waste weight, and participant count.
3. Waste Sorting — Categorizes collected waste by type through photos of sorted piles and weight by category.
4. Recycling Delivered — Transports recyclables to processing facility through delivery receipt and materials list.
5. Composting — Processes organic waste into compost through compost pile photos and input/output weights.
6. Upcycling Project — Transforms waste into useful products through input materials and output products documentation.

Impact Metrics:

Metric	Description
Total Waste Removed	kg collected
Diversion Rate	Percentage recycled vs. landfilled
Area Cleaned	Total square meters restored
Participants Engaged	Unique individuals
Material Recovery Value	Economic value of recyclables

Domain 3: Agroforestry and Planting (DeSci + Environment Pillars)

Primary Chapters: Greenpill Brasil through AgroforestryDAO (led by Diogo), and Greenpill Uganda (led by Jonathan).

Key Outcome: Biodiversity increase, carbon sequestration, and intergenerational knowledge transfer.

Two distinct approaches define this domain. The AgroforestryDAO approach in Brasil builds not just a system to capture work but one that spreads knowledge to future generations through documentation of sustainable farming practices and biodiversity corridors. The Educational Planting approach in Uganda combines planting with education where Jonathan visits schools and churches to plant native, fruit-bearing trees. Each student becomes a caretaker for a specific tree, creating personal relationships with the ecosystem and enshrining the importance of trees from environmental, nutritional, and spiritual perspectives.

Action Lifecycle:

1. Site Assessment — Evaluates land for planting suitability through GPS, soil photos, and existing vegetation survey.
2. Species Selection — Chooses appropriate native species through species list, rationale, and seed source.
3. Site Preparation — Clears and amends soil through before photos and soil test results if available.
4. Planting — Completes tree and seedling installation through species, count, GPS of each tree if feasible, and photos.
5. Student Assignment — Assigns tree to student caretaker (Uganda model) through student name, tree ID, and assignment photo.
6. Monitoring and Observation — Checks health, growth, and identifies issues through photos, measurements, and health status.
7. Harvesting — Collects fruits, seeds, or other products through yield weight, quality, and distribution plan.
8. Knowledge Documentation — Records farming techniques for future generations through written guide, video, and languages used.

Impact Metrics:

Metric	Description
Trees Planted	Total count by species
Native Species Ratio	Percentage of native vs. non-native
Survival Rate	Trees alive at 6-month and 1-year checkpoints
Caretakers Engaged	Students and community members assigned
Knowledge Records	Documentation pieces created
Carbon Sequestration Estimate	tCO2e using standard factors
Fruit and Harvest Yield	kg of produce

Domain 4: Education and Web3 Workshops (Education Pillar)

Primary Chapters: All Greenpill chapters plus Artizen Fund student cohort.

Key Outcome: Skill building in coordination mechanisms and intergenerational knowledge transfer.

Green Goods serves as the verification layer for educational initiatives across three program types. General Web3 Workshops teach coordination and governance mechanisms across all chapters. Artizen Fund Cohort consists of students funded to host events using regenerative tools, with Green Goods verifying attendance and learning outcomes. Tech and Sun Educational Programs deliver blockchain and ReFi training at solar hubs in Nigeria.

Action Lifecycle:

1. Event Planning — Schedules and prepares workshop through event details, curriculum, and venue confirmation.
2. Workshop Hosted — Conducts educational session through attendance list, photos, and materials used.
3. Participant Attendance — Records individual attendance through sign-in, photo proof, and engagement evidence.
4. Learning Assessment — Evaluates through quiz, project, or demonstration with results and submissions.
5. Certificate Issued — Provides completion credential through NFT badge or attestation.
6. Follow-up Action — Documents participant applying learning through evidence of real-world application.

Impact Metrics:

Metric	Description
Workshops Hosted	Total count by type
Participants Trained	Unique individuals across all sessions
New Web3 Users Onboarded	First-time wallet creators
Completion Rate	Percentage completing full curriculum
Follow-up Actions	Documented applications of learning
Geographic Reach	Countries and regions served
Artizen Fund Compliance Rate	Verified vs. funded events

Domain 5: Mutual Credit and Farmer Verification (DePIN + DeSci Pillars)

Primary Chapter: Greenpill Brasil through commitment pooling pilot.

Key Outcome: Credit access for rural farmers through verified productive capacity.

Rural farmers often lack access to credit due to inability to verify their productive capacity. Green Goods serves as the verification tool for commitment pooling pilots in Brasil, where farmers demonstrate consistent agricultural output to unlock mutual credit lines. This creates a pathway from verified work to financial inclusion.

Action Lifecycle:

1. Farm Registration — Documents farm location, size, and crops through GPS boundaries, land ownership proof, and crop inventory.
2. Planting Season — Records what was planted, when, and where through planting photos, seed receipts, and area planted.
3. Growth Monitoring — Provides regular updates on crop progress through photos over time and growth measurements.
4. Harvest Documentation — Records harvest quantity and quality through harvest photos, weight, count, and quality grade.
5. Sale and Distribution — Documents where produce went through buyer receipt, market photos, and price achieved.
6. Commitment Fulfilled — Demonstrates completion of stated commitment through summary of promised vs. delivered output.

Impact Metrics:

Metric	Description
Farms Registered	Total verified farms
Harvest Volume	Total kg documented
Commitment Fulfillment Rate	Percentage meeting stated targets
Credit Unlocked	Value of credit lines enabled
Yield per Hectare	Productivity metrics by crop

3.5 Action Schema Registry

Green Goods maintains a registry of standardized action schemas that Gardens can activate based on their focus area. Actions are tagged with the Eight Forms of Capital they create or enhance.

Core Actions (All Gardens)

Action	Capital Forms
Planting	Living, Material, Experiential
Identify Plant	Living, Intellectual, Experiential
Watering	Living, Material, Experiential
Harvesting	Living, Material, Experiential
Waste Cleanup	Material, Social, Experiential
Web3 Workshop	Intellectual, Social, Experiential
Event Attendance	Intellectual, Social

Domain-Specific Actions

Action	Capital Forms	Availability
Solar Hub Session	Material, Intellectual, Social	TAS Gardens only
Node Deployment	Material, Financial	TAS Gardens only
Waste Sorting	Material, Intellectual	Waste-focused Gardens
Tree Assignment	Living, Social, Cultural	Educational Gardens
Knowledge Documentation	Intellectual, Cultural	AgroforestryDAO Gardens
Farm Registration	Living, Material	Credit Pool Gardens
Commitment Fulfilled	Financial, Material	Credit Pool Gardens

Eight Forms of Capital Reference

Capital	Description
Living	Biodiversity, ecosystems, soil, and water
Material	Physical infrastructure, equipment, and land
Financial	Money, credit, and investment
Social	Relationships, trust, and networks
Intellectual	Knowledge, skills, and data
Experiential	Wisdom from practice and tacit knowledge
Cultural	Traditions, identity, and meaning
Spiritual	Purpose, connection, and meaning

3.6 Schema Validation Plan

Before launch, all action schemas must be validated with actual users in each domain to:

• Confirm that fields capture necessary data without creating friction
• Identify missing fields that domain experts require
• Test AI Agent parsing accuracy
• Measure completion rates

Solar Domain Validation

• Participants: 3-5 Tech and Sun operators from Nigeria (including Anthony Amio and Mmeri Anosike) plus university coordinators
• Duration: 90 minutes per session, 2 sessions planned
• Validation Criteria: Completing submission in under 5 minutes via PWA; AI Agent correctly parsing session type, user count, and energy reading; Switch smart meter readings being easily captured
• Fields to Validate: sessionType options, energyGenerated capture method, nodeUptime format

Waste Domain Validation

• Participants: 2-3 members from each of Cape Town, Koh Phangan, and Rio Brasil chapters (6-9 total)
• Duration: 60 minutes per session, 3 sessions
• Validation Criteria: Logging while doing physical cleanup work with one-handed operation; waste type selection covering all local materials; weight estimation feasibility without scales
• Fields to Validate: wasteSelection options (fishing nets, e-waste), estimatedWeight (with potential bag count alternative), disposalMethod (local options)

Agroforestry Domain Validation

• Participants: Diogo and 2-3 AgroforestryDAO members in Brasil, plus Jonathan and 2-3 team members in Uganda
• Duration: 90 minutes per session, 2 sessions
• Brasil session focuses on Knowledge Documentation schema and Portuguese language support
• Uganda session focuses on Planting schema with Student Assignment extension during actual school planting event
• Fields to Validate: plantSelection (species database needs), caretakerName (privacy concerns), seedSource (local options)

Education Domain Validation

• Participants: 3-5 workshop facilitators plus 2 Artizen Fund cohort members
• Duration: 45 minutes per session, 3 sessions conducted during actual workshop events
• Validation Criteria: Logging attendance without disrupting workshop flow; capturing meaningful learning outcomes
• Fields to Validate: topics options, learningOutcome format

Mutual Credit Domain Validation

• Participants: 3-5 farmers from Brasil commitment pooling pilot
• Duration: 60 minutes per session, 2 sessions
• Validation Criteria: Farmers with limited tech literacy completing harvest logging; quality grading being understandable
• Fields to Validate: qualityGrade definitions, commitmentPoolId linking, estimatedValue (local currency)

---

4. Problem Statement

4.1 What's Broken Today

• Verification Gap: Traditional MRV is too expensive for small projects ($5,000-15,000 per verification event), while simple photo uploads lack scientific rigor and trust for institutional capital.
• Capital Volatility: Crypto-philanthropy often relies on volatile tokens or one-off grants. When the market crashes or donors churn, projects die.
• Accessibility Wall: Most web3 impact tools require a smartphone, a wallet, and ETH for gas. This excludes the poorest and most vital ecological laborers.
• Missing Feedback Loops: Beneficiaries (local communities) and Experts (Evaluators) are often excluded from the verification loop, leading to "paper parks" or greenwashing.

4.2 Why It Matters Now

• Global South Adoption: The environment for crypto and blockchain technology is becoming increasingly friendly and vital in the Global South (e.g., Nigeria, Brazil). High mobile penetration combined with a need for stable financial rails makes these regions ideal adopters for a "Local-First" regenerative economy.
• The Yield Economy: The maturity of Octant and DeFi allows for "endowment models" where principal is preserved and only yield is spent.
• AI at the Edge: LLMs have become efficient enough to run "Agri-Advisor" agents that can democratize access to complex protocols via simple text.

---

5. Scope

5.1 In Scope (This Release)

Interfaces:

• PWA — Offline-first web app with localization support (Spanish/Portuguese)
• Admin Dashboard — UI for managing Gardens and creating impact reports
• AI Agent (Beta) — WhatsApp/SMS bot for Gardener submissions

Protocol Layer:

• Karma GAP — Automated reporting pipelines
• Octant Vaults — Deployment of yield-bearing vaults for capital formation
• Revnets — Configuration of Juicebox Revnets for Garden Treasury management
• Hypercerts — Minting logic for verified impact claims
• Conviction Voting — Module for yield allocation and action prioritization

Identity and Reputation:

• Unlock Protocol — Minting of GreenWill reputation badges (NFTs) based on user activity
• Hats Protocol — Management of on-chain roles (e.g., Operator, Evaluator) and permissions

Verification:

• Operators — Tools for initial data aggregation and verification
• Evaluators — Admin Dashboard view for certifying Impact Reports
• Community Signaling — Mechanism for beneficiaries to flag issues or upvote priorities

5.2 Revnet Configuration (The Fractal Gardens Model)

Green Goods adopts a two-tiered Fractal Gardens architecture using Juicebox V4 Revnets.

Level 1: Protocol Revnet ($GG Token)

Issues the core Green Goods governance token. Funds shared infrastructure including servers, indexers, legal, and development. Receives tribute from Level 2 Gardens at 2.5-5% of incoming funds and tokens.

Level 2: Garden Revnets (Local Tokens)

Each community operates its own Revnet (e.g., $LAGOS, $BRAZIL, $NIGERIA). Issues local tokens representing governance rights and treasury claims. Routes tribute percentage to Level 1 Protocol.

5.3 The Octant-Backstopped Impact Floor

This mechanism creates sustainable funding independent of speculation:

1. Deposit — Funders stake assets in Octant Vault.
2. Yield Generation — Low-risk DeFi strategies generate returns.
3. Programmatic Buyback — Yield automatically purchases Hypercerts from Gardens.
4. Floor Rising — Each purchase increases backing per token, creating a minimum wage for ecological labor backed by Ethereum staking rewards.

5.4 Out of Scope

• Robust Identity System (DID) — A fully decentralized identity system backed by W3C DIDs is deferred to a future release; v1 relies on WebAuthn and standard maturation.
• Global IoT Network — Full integration with hardware sensors.
• Scope 3 Data Marketplace — Insurance/Finance data sales are Phase 3.

5.5 Assumptions

A1: Onchain impact reporting is more valuable than traditional reporting.

• Validation: Track funder preference in post-pilot surveys. Compare funding secured by onchain-reporting Gardens vs. traditional reporting projects.
• Red Flag: If traditional tooling improves via AI and the onchain value-add is not perceived as differentiated by Q2 2026.

A2: Projects can long-term earn enough yield to be sustainable.

• Validation: Calculate yield-to-impact ratio across pilot Vaults. Model time-to-sustainability for different capital levels.
• Red Flag: If funders perceive yield as too passive and prefer direct donation models.

A3: Impact certificate marketplaces will develop around Hypercerts.

• Validation: Monitor Hypercerts ecosystem liquidity. Track secondary sales volume.
• Red Flag: By Q3 2026, no impact marketplaces building liquidity around Hypercerts.

A4: AI Agents can accurately parse natural language impact submissions.

• Validation: Measure Agri-Advisor accuracy rate on image identification and metadata extraction.
• Red Flag: If error rate exceeds 15%, requiring Operator intervention on the majority of submissions.

A5: Community members will participate in Conviction Voting without direct financial incentives.

• Validation: Track participation rates across Gardens. Correlate with GreenWill badge levels.
• Red Flag: If voting participation drops below 30% after initial novelty period.

A6: Gasless transactions via Pimlico remain economically viable at scale.

• Validation: Monitor sponsorship costs per transaction. Model break-even point.
• Red Flag: If gas costs exceed 5% of yield generated per action.

A7: Evaluators will maintain integrity under reputation staking model.

• Validation: Track dispute rates and community flagging of Evaluator decisions.
• Red Flag: If collusion patterns emerge in Gardens with concentrated Evaluator power.

---

6. Solution Overview

6.1 Experience Summary

Green Goods v1 acts as a bridge. For the Gardener, it is a helpful WhatsApp contact that pays them for work. For the Funder, it is a high-yield savings account that heals the planet. For the Evaluator, it is a data-rich platform for certifying truth.

6.2 Core Workflow Pattern (MDR)

1. Capture/Input — Photos, metadata, GPS, and voice notes via PWA or AI Agent.
2. Details/Structure — CIDS-compliant action schema, attestation construction, and Hypercert metadata.
3. Review/Confirm — Operator approval queue, Evaluator certification, and Hypercert minting.

6.3 Key User Flows

Flow 1: AI Agent Impact Reporting (High Accessibility)

1. Capture — Gardener sends photo to WhatsApp Bot.
2. Process — AI Agent identifies object, cross-references with "Action Schema."
3. Sign — Agent uses server-side Smart Account (signer controlled) to sign payload.
4. Store — Data anchored on IPFS (Storacha) and stored onchain with EAS.

Flow 2: Impact Report Minting and Evaluation (High Trust)

1. Aggregate — Operator selects 20 verified "Tree Planting" actions in the PWA and clicks "Create Report."
2. Certify — The Evaluator receives the report. They review the aggregated evidence (data + photos). They attest to the validity of the report using their Hats Protocol authority.
3. Mint — The system now allows the Operator to mint the Hypercert, which references the Evaluator's certification attestation.

Flow 3: Capital Formation and Allocation (Sustainability)

1. Deposit — Funder stakes assets in a Garden's Octant Vault.
2. Yield — Yield accrues in the Vault.
3. Buyback — Protocol programmatically uses the allocated yield to buy verified Garden Hypercert shares, raising the "Impact Floor" price.

Flow 4: Governance and Signaling (High Coordination)

1. Yield Allocation — Garden members (operator, gardener, community member) vote via Conviction Voting to determine the split of yield to send to hypercerts.
2. Action Prioritization — Garden operators and gardeners signal which specific actions (e.g., "Waste Cleanup") the Garden should prioritize. This signal directs the Operator's focus for the next cycle.

---

7. Requirements

7.1 Functional Requirements

ID	Requirement	Description
FR-P-001	AI Agent Interface	Integration of LLM wrapper (e.g., Twilio/OpenAI) to parse natural language submissions via WhatsApp/SMS
FR-P-002	Admin Dashboard	Comprehensive dashboard for Operators to manage submissions and for Evaluators to review and certify Impact Reports
FR-P-003	Octant Vaults	ERC-4626 Vault contracts configured to route yield to specific Garden Revnets
FR-P-004	Hypercerts Minting	Minting logic for verified impact claims as ERC-1155 certificates
FR-P-005	Revnet Configuration	Deployment of Juicebox V4 Revnets with configurable redemption rates (80-90%) and reserved rates
FR-P-006	Conviction Voting	Implementation of the Gardens Conviction Voting module for yield allocation and signaling action priorities
FR-P-007	GreenWill Badging	Integration with Unlock Protocol to mint NFT badges based on user activity thresholds
FR-P-008	Role Management	Integration with Hats Protocol to assign and manage on-chain permissions for Operators and Evaluators
FR-P-009	Smart Account Automation	Backend infrastructure to manage Safe accounts for users interacting solely via AI Agents
FR-P-010	Karma GAP Sync	Bi-directional sync between Green Goods actions and Karma GAP milestone updates
FR-P-011	Passkey Authentication	WebAuthn-based biometric login with Safe Smart Account deployment via Pimlico

7.2 Non-Functional Requirements

ID	Category	Requirement
NFR-001	Localization	PWA must support dynamic localization, launching with full support for EN/ES/PT. All user-facing strings externalized. Date, time, and number formatting must respect locale settings.
NFR-002	Latency	AI Agent must respond within 5 seconds. PWA page load within 3 seconds on 3G. Admin Dashboard queries within 2 seconds for datasets under 1,000 records.
NFR-003	Trust	All Evaluator actions must be on-chain and auditable. Evaluator reputation must be slashable. Complete action provenance trail from submission to Hypercert must be queryable.
NFR-004	Cost	Gardeners must pay $0 gas via Pimlico sponsorship. Sponsorship budget must be monitored with alerts at 80% consumption. Fallback mechanism required if Pimlico is unavailable.
NFR-005	Performance	Full offline capability with sync on reconnect. Service Worker cache must support 500+ pending actions. IndexedDB storage must not exceed 50MB per session. Image compression under 500KB before IPFS pinning.
NFR-006	Reliability	99.5% uptime for core flows. Graceful degradation (if IPFS fails, queue locally; if Arbitrum is congested, batch transactions). Auto-retry with exponential backoff. Circuit breaker pattern for Twilio and OpenAI.
NFR-007	Security	Deterministic CREATE2 addresses for Smart Account deployments. Passkey credentials never leave device. Rate limiting at 100 req/min per wallet. Hats-enforced RBAC at contract level.
NFR-008	Privacy	GPS obfuscated to 3 decimal places (~111m) for public display. User PII encrypted on IPFS with only wallet address on-chain. GDPR-compliant deletion. Analytics use anonymized wallet hashes.
NFR-009	Accessibility	WCAG 2.1 AA. Touch targets minimum 44x44px. High contrast mode for outdoor/bright light. Screen reader compatibility for all core flows.
NFR-010	Compatibility	Chrome 90+, Safari 15+, Firefox 90+, Samsung Internet 15+. iOS 14+, Android 10+. Minimum device: 2GB RAM, 16GB storage. Offline mode on devices with no active data plan.

---

8. Feature Breakdown

ID	Feature	Priority	Status	Effort	Dependencies	Notes
GG-FEAT-001	PWA Offline-First Architecture	Critical	In Progress	3 weeks	IndexedDB, Service Workers	Core to Last Mile access
GG-FEAT-002	Passkey Authentication	Critical	In Progress	2 weeks	WebAuthn, Safe, Pimlico	Removes Privy dependency
GG-FEAT-003	AI Agent (WhatsApp/SMS)	High	Planned	4 weeks	Twilio, OpenAI/Llama	Billions interface
GG-FEAT-004	Admin Dashboard v2	High	Planned	3 weeks	React, Envio	Operator verification queue
GG-FEAT-005	Hypercerts Minting	High	Planned	4 weeks	ERC-1155, IPFS	Aggregation + atomic mint
GG-FEAT-006	Octant Vault Integration	High	Planned	4 weeks	ERC-4626, DeFi strategies	Yield-bearing deposits
GG-FEAT-007	Revnet Configuration	Medium	Planned	3 weeks	Juicebox V4	Garden treasury management
GG-FEAT-008	Conviction Voting Module	Medium	Planned	3 weeks	Gardens Protocol	Yield allocation governance
GG-FEAT-009	GreenWill Badging	Medium	Planned	2 weeks	Unlock Protocol	Reputation NFTs
GG-FEAT-010	Hats Protocol Roles	Medium	Planned	2 weeks	Hats Protocol	Operator/Evaluator permissions
GG-FEAT-011	Karma GAP Sync	High	In Progress	2 weeks	EAS, Karma API	CIDS-compliant reporting
GG-FEAT-012	Localization (ES/PT)	Medium	In Progress	2 weeks	i18n framework	Brazilian and LatAm access
GG-FEAT-013	Report Generator	Medium	Planned	2 weeks	PDF, JSON export	Grant compliance automation

---

9. Economic Model and Sustainability

This section provides financial modeling for Green Goods sustainability including vault economics, gas costs, and break-even analysis.

9.1 Model Assumptions

Token and Network Assumptions:

• ETH price at $3,000 USD (conservative baseline)
• Arbitrum gas price at 0.1 gwei average
• Pimlico bundler fee at $0.02 per transaction

Yield Assumptions:

• Octant Vault base yield at 4% APY (ETH staking via Lido)
• Enhanced yield with DeFi strategies at 6-8% APY
• Yield distribution: 80% to Hypercert purchases, 20% to protocol operations

Activity Assumptions (per Arbitrum grant targets):

• Active Gardeners: 150 by end of Q1 2026
• Actions per Gardener per month: 10 average
• Total actions per month: 1,500
• Actions per Hypercert bundle: 50 average
• Hypercerts minted per month: 30

9.2 Vault TVL Sustainability Analysis

Monthly Operating Costs:

Item	Calculation	Cost
Pimlico gas sponsorship	1,500 actions x $0.02	$30
IPFS pinning (Storacha)	1,500 images x $0.001	$1.50
AI Agent API costs	1,500 interactions x $0.01	$15
Indexer and infrastructure	Flat	$100
Total monthly infrastructure		~$150

TVL Required — Infrastructure Only:

Yield Rate	TVL Required
4% APY	$45,000
6% APY	$30,000
8% APY	$22,500

TVL Required — Infrastructure + $500/mo Hypercert Purchases:

Yield Rate	TVL Required
4% APY	$195,000
6% APY	$130,000
8% APY	$97,500

Conclusion: The Arbitrum grant target of $20,000 TVL is a starting point but not sustainable long-term. The protocol needs to grow to approximately $100,000-$200,000 TVL for self-sustaining operations with meaningful Hypercert purchasing power.

9.3 Gas Sponsorship Budget Analysis

Pimlico Sponsorship Budget — $2,000 allocated from Arbitrum grant:

Sponsored actions = $2,000 / $0.02 = 100,000 actions

Pace	Runway
1,500 actions/month	66 months
5,000 actions/month	20 months
10,000 actions/month	10 months

Yield-Funded Sponsorship:

• Monthly gas need at 1,500 actions: $30
• Required TVL at 6% APY: $6,000
• This is achievable with minimal vault deposits.

High Growth Scenario (10x activity):

• Monthly gas need at 15,000 actions: $300
• Required TVL at 6% APY: $60,000

Recommendation: Implement gas budget monitoring with alerts at 50%, 75%, and 90% consumption. Plan transition from grant-funded to yield-funded sponsorship by Month 6.

9.4 Revnet Token Economics Simulation

Simulation Parameters:

• Initial treasury: 10 ETH ($30,000)
• Monthly inflows: 2 ETH from new contributions + 0.5 ETH from yield
• Monthly redemptions: 10% of circulating tokens

Configuration	Redemption	Reserved	Month 12 Treasury	Token Backing	Tokens in Circulation	Assessment
A	80%	10%	18.5 ETH	1.15x initial	85% of peak	Healthy growth, moderate holder returns
B	85%	15%	22.3 ETH	1.28x initial	78% of peak	Stronger treasury, balanced holder returns
C	90%	20%	26.1 ETH	1.42x initial	71% of peak	Maximum treasury growth, lower short-term holder returns

Recommendation: Start with Configuration B (85% redemption, 15% reserved) as the default. This balances treasury health with reasonable returns for Gardeners who need to cash out for real-world expenses.

9.5 Hypercert Pricing Model

Approach A — Cost-Plus Pricing:

Calculate cost to produce verified actions (labor, tools, materials), add margin for Garden sustainability (20-30%).

Example: 50 tree plantings x $2 cost each = $100 base + 25% = $125 Hypercert price

Approach B — Impact-Based Pricing:

Calculate estimated impact value (carbon, biodiversity, social factors) using standard impact factors (Gold Standard, Verra).

Example: 50 trees x 0.5 tCO2e/tree x $15/tCO2e = $375 Hypercert price

Approach C — Market-Driven Pricing:

Let Gardens list Hypercerts at their desired price. Octant Vault purchases at market clearing price. Price discovery through supply and demand.

Recommendation: Use Approach A (Cost-Plus) for initial pricing to establish floor, with Approach C (Market-Driven) as the long-term model once liquidity develops.

9.6 Break-Even Analysis

Protocol Break-Even (infrastructure costs covered by yield):

• Target: $30,000 TVL at 6% APY
• Timeline: Month 6 of beta program

Gardener Break-Even (redemption value exceeding effort):

• Token value must exceed local minimum wage equivalent
• If a Gardener submits 20 actions/month and can redeem for $50, this beats minimum wage in Nigeria ($70/month) but not Brasil ($260/month) at early scale
• Implication: Early Gardeners are motivated by GreenWill reputation and future upside rather than immediate income

Garden Break-Even (treasury growth exceeding outflows):

• Monthly inflows (contributions + yield) must exceed monthly redemptions
• With 85% redemption rate and 10% monthly redemption, need 15% monthly treasury growth
• Implication: Gardens must actively attract new funders or reduce redemption frequency

9.7 Financial Dashboard Metrics

Vault Metrics:

• Total TVL across all vaults (USD and ETH)
• Monthly yield generated
• Yield allocation breakdown (Hypercert purchases vs. operations)

Treasury Metrics:

• Protocol Revnet treasury balance
• Garden Revnet treasury balances (aggregate and per-garden)
• Token backing ratio (treasury / circulating supply)
• Monthly redemption volume

Cost Metrics:

• Gas sponsorship burn rate
• Monthly infrastructure costs
• Cost per verified action

Revenue Metrics:

• Hypercerts sold (count and value)
• Tribute received from Gardens
• Protocol fee revenue (if applicable)

---

10. Localization Strategy

This section details the approach to supporting multiple languages and regional variations.

10.1 Language Support Tiers

Tier 1 — Full Support at Launch:

• English (US) — Default language
• Portuguese (Brazilian) — For Greenpill Brasil and AgroforestryDAO
• Spanish (Latin American) — For broader LatAm expansion

Tier 2 — Planned for v1.x:

• French (West African) — For Ivory Coast chapter
• Swahili — For East African expansion including Uganda
• Pidgin English — For Nigeria (critical for TAS adoption)

Tier 3 — Roadmap for v2:

• Thai — For Koh Phangan chapter
• Arabic — For future MENA expansion

10.2 Localization Components

PWA Localization

All user-facing text stored in JSON locale files with no hardcoded strings:

locales/en-US/common.json    — Shared strings
locales/en-US/actions.json   — Action schema labels
locales/pt-BR/common.json    — Brazilian Portuguese
locales/es-419/common.json   — Latin American Spanish

• Date and time formatting uses Intl.DateTimeFormat with user locale
• Number formatting uses Intl.NumberFormat respecting locale decimal separator

AI Agent Localization

• Language detection from first message with option to ask user preference if ambiguous
• Separate prompt templates per language to handle greeting patterns, date formats, local plant/crop names, and regional slang

Pidgin English Handling (Nigeria)

Many Nigerians code-switch between English and Pidgin, so the AI Agent must handle mixed messages:

• "I don plant tree" parses as "I planted a tree"
• "Wetin you wan do?" parses as "What do you want to do?"

Build Pidgin-specific prompt layer tested with Nigerian users. UI defaults to English with option for Pidgin (some users prefer English UI but Pidgin conversation).

Portuguese Handling (Brasil)

• Brazilian Portuguese has significant vocabulary differences from European Portuguese
• Use "voce" form for Brazilian locale
• Recognize both Brazilian and European forms in AI parsing but default to Brazilian

Action Schema Localization

Every schema field has label translations:

Field	English	Portuguese	Spanish
plantSelection	Plant Species/Name	Especie/Nome da Planta	Especie/Nombre de la Planta

All select options have translations. Placeholder text localized with culturally relevant examples (e.g., "e.g., Mango, Avocado, Oak" becomes "ex., Manga, Abacate, Ipe" in Portuguese).

10.3 Regional Considerations

Region	Currency	Date Format	Decimal	Language Default
Nigeria	Naira (NGN)	DD/MM/YYYY	Period	English with Pidgin option
Brasil	Real (BRL, R$)	DD/MM/YYYY	Comma (1.000,50)	Portuguese (Southeast dialect)
Latin America	Varies by country	Varies	Varies	Spanish (es-419) over Spain (es-ES)

10.4 Translation Workflow

Initial Translation:

• English (US) is source language for all translations
• Professional translation for Tier 1 languages with review by native-speaking community members
• Translation partners: Diogo and AgroforestryDAO team for Portuguese; Greenpill LatAm network for Spanish

Ongoing Translation:

1. Developer adds string to en-US locale file
2. String flagged as "needs translation" in other locales
3. Weekly translation batch sent to community translators
4. Translations reviewed by second native speaker
5. Merged and deployed

Quality Assurance:

• Native speaker review from target region (Brazilian for pt-BR, not Portuguese)
• Contextual testing — translators see strings in context
• Back-translation check for critical strings

10.5 Right-to-Left Roadmap

• Current Status: No RTL support in v1.
• Technical Preparation: CSS using logical properties (margin-inline-start instead of margin-left). Bidirectional text handling in components. RTL-aware icon placement.
• Timeline: RTL support planned for v2 when Arabic expansion is prioritized.

10.6 Localization Success Metrics

• Adoption by Language — Track active users by locale setting
• Completion Rates by Language — Compare action submission completion rates across languages; if Portuguese completion is lower than English, investigate friction
• AI Agent Accuracy by Language — Track parse success rate per language; target within 5% of English baseline
• Translation Coverage — Percentage of strings translated per language; target 100% for Tier 1 at launch

---

11. Privacy and Data Handling

This section assesses privacy risks and documents data handling practices.

11.1 Data Inventory

Data Collected from Gardeners:

Data Type	When Collected	Storage	Purpose
Wallet Address	Passkey creation	On-chain (immutable)	Identification
GPS Coordinates	Each action	IPFS (obfuscated to 3 decimal places)	Location verification
Photos and Media	Submissions	IPFS via Storacha	Verification evidence
Timestamps	Automatically	On-chain (attestations)	Provenance
Action Metadata	Per schema	IPFS and on-chain	Impact tracking
Device Information	Analytics	Anonymized	Usage patterns
Optional Profile (name, bio)	User-provided	IPFS with encryption	Identity

Data Collected from Operators and Evaluators:

Data Type	Storage	Purpose
Wallet Address	On-chain	Identification
Hats Protocol Role	On-chain	Permissions
Verification Decisions	On-chain	Audit trail

Data Collected from Funders:

Data Type	Storage	Purpose
Wallet Address	On-chain	Identification
Deposit Amounts	On-chain	Treasury management
Conviction Votes	On-chain	Yield allocation

11.2 Data Flow and Residency

Action Submission Flow:

Step	Action	Privacy Risk
1	Gardener captures photo on device	Low
2	Gardener adds metadata (GPS)	Medium
3	User submits via PWA or AI Agent (encrypted in transit)	Low
4	Backend processes in server memory (ephemeral)	Medium
5	Image uploaded to IPFS (distributed, immutable)	High
6	Attestation created on Arbitrum (permanent, public)	High
7	Envio indexes attestation (centralized copy)	Medium

Data Residency:

Location	Data	Characteristics
On-chain (Arbitrum)	Attestations, wallet addresses, transactions	Immutable, global
IPFS (Storacha)	Images, metadata JSON, profile data	Immutable once pinned, global
Backend	Processing logs, session data	Ephemeral, deleted after processing
Analytics	Anonymized usage data	Provider region

11.3 Privacy Risks and Mitigations

GPS Location Reveals Sensitive Information

• Risk: Precise coordinates could reveal home locations or sensitive sites.
• Likelihood: High | Impact: Medium | Level: High
• Mitigations: GPS obfuscated to 3 decimal places (~111m precision) before IPFS storage; raw GPS never stored on-chain; users informed during onboarding; future option to disable GPS for privacy-sensitive users.
• Residual Risk: 111m precision still identifies general area but is acceptable for environmental verification.

Photos Contain Identifiable Information

• Risk: Photos may contain faces, license plates, or property features.
• Likelihood: Medium | Impact: High | Level: High
• Mitigations: Future AI face detection prompting user to retake; Operator review flagging PII before Hypercert minting; user guidance to avoid including faces; clear notice that IPFS photos are public.
• Residual Risk: Users have responsibility with guidance provided.

Blockchain Data Is Immutable

• Risk: Once on-chain, data cannot be deleted, conflicting with GDPR right to erasure.
• Likelihood: Certain | Impact: High | Level: High
• Mitigations: Minimize on-chain PII (only wallet address); profile data on IPFS with encryption where key destruction renders data unreadable; attestations reference IPFS CIDs where unpinning empties the shell; privacy notice clearly explaining immutability during signup; GDPR interpretation that wallet address may not be personal data if not linked to real identity.
• Residual Risk: Wallet address is pseudonymous but potentially linkable — legal gray area.

AI Agent Processes Sensitive Conversations

• Risk: WhatsApp/SMS messages may contain unintended personal information.
• Likelihood: Medium | Impact: Medium | Level: Medium
• Mitigations: AI prompts scoped to action reporting with personal queries redirected; message content not stored long-term; WhatsApp Business API compliance; opt-out available (use PWA only).
• Residual Risk: Conversational interfaces may elicit more disclosure than forms.

Third-Party Service Data Exposure

• Risk: Data shared with Twilio, OpenAI, Pimlico, and Storacha creates exposure.
• Likelihood: Low | Impact: High | Level: Medium
• Mitigations: Data Processing Agreements with all processors; minimize data shared; review third-party policies; Twilio and OpenAI have SOC 2 compliance.
• Residual Risk: Third-party policies may change, requiring ongoing monitoring.

Caretaker Data in Educational Planting

• Risk: Uganda model assigns trees to students and may collect student names (minors' data).
• Likelihood: Medium | Impact: High | Level: High
• Mitigations: Student names as optional field allowing pseudonyms; no photos of minors required; school administrator as adult Operator; parental consent process for any minor data; student ID codes instead of names.
• Residual Risk: Children's privacy requires extra care — default to minimal collection.

11.4 Data Subject Rights Implementation

• Right to Access — Users can export all data via Profile settings (submitted actions, attestations, profile data).
• Right to Rectification — Users can update profile data. Action data can be flagged for Operator correction via new attestation with correction reference (original is not deleted).
• Right to Erasure — Profile data: destroy encryption key (IPFS data becomes unreadable). On-chain attestations: content references become empty if IPFS unpinned. Analytics: deletion request to provider. Limitation: Full erasure not possible for blockchain data; users informed before signup.
• Right to Portability — Data export in machine-readable JSON format (action history, attestation references, profile data).

11.5 Compliance Requirements

GDPR Compliance:

• Lawful basis: consent (explicit during signup)
• Privacy notice required (to be drafted)
• Data Processing Agreements required with all processors
• Data Protection Officer: not required (under 250 employees threshold)
• Data breach notification process to be documented
• Cross-border transfers: Standard Contractual Clauses

LGPD Compliance (Brasil):

• Similar requirements to GDPR
• Portuguese language privacy notice required

POPIA Compliance (South Africa):

• Similar requirements to GDPR
• Local registration may be required

11.6 Privacy Documentation Required

Document	Purpose	Timeline
Privacy Notice	User-facing, explaining data collection, use, and rights	Before launch
Cookie Policy	Required if any cookies used (PWA may use local storage)	Before launch
Data Processing Agreements	Contracts with Twilio, OpenAI, Pimlico, Storacha, Envio	Before launch
Data Retention Policy	Internal document specifying retention periods	Before launch
Data Breach Response Plan	Internal procedure for handling security incidents	Before launch

---

12. Dependencies and Partnerships

12.1 Technical Dependencies

• Arbitrum One — L2 Settlement
• Ethereum — L1 for Octant
• Twilio — SMS and WhatsApp
• OpenAI / Llama — LLM
• Pimlico — Gasless Transactions
• Envio — Blockchain Indexer

12.2 Protocol Partners

Partner	Role
Octant	Yield generation
Hypercerts	Impact certificates
Juicebox	Revnet infrastructure
Karma GAP	Compliance reporting
Unlock Protocol	Reputation (GreenWill)
Hats Protocol	Permission layer
Gardens Protocol	Conviction Voting

12.3 Funding and Milestones

Arbitrum New Protocols and Ideas Grant — $25,000 USD total:

Milestone	Timeline	Deliverable	Amount
M1	Month 1	PRD v2 and Designs	$4,000
M2	Months 1-2	Hypercerts Integration	$8,000
M3	Months 2-3	DeFi Integration	$8,000
M4	Months 1-2	Community Activations	$4,000
M5	Month 3	Final Reporting	$1,000

Celo Builders Fund x Unlock Protocol — TBD

---

13. QA, Rollout, and Ops

13.1 QA Strategy

Test Approach

• Unit Tests: Jest for frontend components, Foundry for Solidity contracts
• Integration Tests: End-to-end flows using Playwright
• UAT: Structured testing with 5 Gardeners and 3 Operators per pilot Garden
• Pilot Program: Live testing with Greenpill Nigeria, Brasil, and Cape Town chapters

Device and Platform Coverage

P0 Priority:

Device	OS	Browsers
Android mid-range (Samsung A-series)	Android 11+	Chrome, Samsung Internet
iPhone SE 2020+	iOS 15+	Safari

P1 Priority:

Device	OS	Browsers
Low-end Android (under 2GB RAM)	Android 10+	Chrome
Desktop (Windows 10+, macOS 12+)	—	Chrome, Firefox

Acceptance Gates

• All P0 devices must pass happy-path flows
• Offline-to-sync flow must complete without data loss
• Gas sponsorship must succeed for all test transactions
• Hypercert minting must produce valid ERC-1155 tokens
• Karma GAP sync must produce correctly-structured attestations

QA Environments

Environment	Network	Deployment
Dev	Arbitrum Sepolia	Continuous deployment
Staging	Arbitrum Sepolia	Release candidates
Production	Arbitrum One	Gated rollout

13.2 Rollout Plan

Phase	Timeline	Cohort	Features	Gate
Phase 1: Internal Alpha	Weeks 1-2	Core team (8 users)	PWA core, Passkey auth, basic submission	Under 5% error rate on submission flow
Phase 2: Pilot Beta	Weeks 3-6	Nigeria (25) + Brasil (25)	Full PWA, AI Agent (WhatsApp), Admin Dashboard	NPS >= 50, under 10% support ticket rate
Phase 3: Extended Beta	Weeks 7-10	8 Gardens total	Hypercerts minting, Octant Vault integration	>= $12k TVL, >= 12 Hypercerts minted
Phase 4: General Availability	Weeks 11-12	Open enrollment	Full protocol + Conviction Voting	—

First 72-Hour Monitoring Signals

• Error rate by flow (submission, approval, minting)
• Latency percentiles (p50, p95, p99)
• Gas sponsorship consumption rate
• User session length and completion funnels
• Support channel volume (Discord and Telegram)

Rollback Plan

• Feature flags can disable any new feature within 5 minutes
• Contract upgrades via Diamond pattern require 24-hour timelock
• Database migrations must include rollback scripts
• IPFS pins are immutable — metadata corrections require new CIDs

13.3 Support Plan

Channel	Purpose
Telegram (per Garden)	Primary Gardener support
Discord #green-goods-support	Secondary technical support
GitHub Issues	Escalation with reproduction steps

Severity Definitions:

Severity	Description	Response SLA	Resolution SLA
P0 Critical	Protocol down, funds at risk, or data loss	4 hours	4 hours
P1 High	Core flow blocked for >10% of users	12 hours	24 hours
P2 Medium	Feature degraded but workaround exists	72 hours	1 week
P3 Low	Cosmetic or minor inconvenience	2 weeks	Next release

13.4 Analytics Plan

Events:

Event	Trigger	Properties	Purpose
session_start	Session start	device_type, os, browser, locale	Usage patterns
passkey_created	Passkey creation	success, error_code	Auth funnel
action_submitted	Action submission	garden_id, action_type, offline_queued	Submission funnel
action_approved	Action approval	garden_id, action_type, evaluator_id	Verification flow
hypercert_minted	Hypercert minting	garden_id, actions_bundled, value	Impact tokenization
vault_deposit	Vault deposit	amount, asset, garden_allocation	Capital formation
conviction_vote	Conviction vote	garden_id, stake_amount, direction	Governance participation
ai_agent_interaction	AI Agent interaction	channel, intent_parsed, success	Agent effectiveness

Dashboards:

Dashboard	Metrics
Executive	TVIV, TVL, Active Gardeners, Hypercerts Minted
Garden Health	Per-garden metrics, verification queue depth, yield allocation
Funnel Analysis	Onboarding completion, submission-to-approval conversion
Technical Health	Error rates, latency, gas consumption

Review Cadence:

• Daily: Error rates and critical funnel metrics
• Weekly: Full dashboard review with team
• Monthly: Cohort analysis, NPS review, milestone progress

---

14. Risks and Mitigations

14.1 Technical Risks

ID	Risk	Likelihood	Impact	Mitigation
R-T1	AI Hallucination	High	Medium	Agent prompts include strict bounds; Agent stages data for Operator review, never finalizes directly
R-T2	Passkey Adoption Friction	Medium	High	Fallback to email magic link for devices without biometric support; clear onboarding guidance
R-T3	Pimlico Downtime	Low	High	Queue transactions locally; secondary bundler fallback; alert on sponsorship failures
R-T4	IPFS Pinning Failures	Medium	Medium	Storacha with Filecoin backup; retry logic with exponential backoff; local queue preservation
R-T5	Smart Contract Vulnerability	Low	Critical	Apply for Arbitrum audit program; formal verification of core flows; bug bounty program post-launch
R-T6	Envio Indexer Lag	Medium	Medium	Optimistic UI updates; background sync status indicators; fallback to direct RPC for critical queries

14.2 Economic Risks

ID	Risk	Likelihood	Impact	Mitigation
R-E1	Insufficient Yield for Sustainability	Medium	High	Model break-even scenarios pre-launch; diversify vault strategies; set realistic expectations with Gardens
R-E2	Revnet Token Price Collapse	Medium	Medium	Cash-out tax (80-90%) preserves treasury floor; focus narrative on work value not speculation
R-E3	Hypercert Illiquidity	High	Medium	Protocol acts as buyer of first resort via Octant yield; build relationships with retro funding rounds (Optimism, Gitcoin)
R-E4	Gas Cost Spike on Arbitrum	Low	Medium	Transaction batching; L3 migration path via Arbitrum Orbit; monitor gas consumption patterns

14.3 Governance and Social Risks

ID	Risk	Likelihood	Impact	Mitigation
R-G1	Evaluator Collusion	Medium	High	Evaluators must stake reputation and tokens; Community Members can flag decisions; escalation to Protocol-level review
R-G2	Low Conviction Voting Participation	High	Medium	Low threshold for conviction to take effect; gamify with GreenWill badges; direct notification on pending votes
R-G3	Garden Operator Abandonment	Medium	High	Multi-sig requirement for Garden administration; succession planning in onboarding; Protocol can appoint interim steward
R-G4	Community Pay-to-Complain Dynamics	Medium	Medium	Signal staking requires reputation (GreenWill) not just tokens; anti-sybil checks on new accounts

14.4 Regulatory and Legal Risks

ID	Risk	Likelihood	Impact	Mitigation
R-L1	Evaluator Liability for False Verification	Medium	High	Pursue legal opinion before beta launch; indemnification in Terms of Service
R-L2	Token Classification as Security	Low	Critical	Structure Revnet tokens as governance + utility; no profit-sharing language; engage securities counsel
R-L3	GDPR Compliance for EU Users	Medium	Medium	Implement data deletion flows; minimize on-chain PII; encryption for off-chain storage

14.5 Operational Risks

ID	Risk	Likelihood	Impact	Mitigation
R-O1	Team Burnout and Capacity Constraints	High	High	Scope management; clear "No" list in Section 5.4; async-first operations; buffer in timeline
R-O2	Dependency on Partner Protocols	Medium	High	Document fallback strategies for each integration; maintain relationships with alternative providers
R-O3	Dialect and Language Accuracy in AI Agent	High	Medium	Pre-launch testing with native speakers; iterative prompt refinement; human-in-the-loop for edge cases

---

15. Open Questions

#	Question	Owner	Due
Q1	What is the legal framework for Evaluator liability if they verify false data?	Legal Lead	Q1 2026
Q2	Can the WhatsApp Agent support dialects (e.g., Pidgin, Portuguese) effectively at launch?	Product Lead	Feb 2026
Q3	How do we incentivize Community Members to signal without creating a "pay-to-complain" market?	Community Lead	Feb 2026
Q4	What is the minimum viable vault TVL for sustainable yield generation?	Finance Lead	Jan 2026
Q5	Should Gardens have the ability to customize their Revnet parameters (redemption rate, reserved rate)?	Product Lead	Feb 2026
Q6	How do we handle disputes when Community Members flag Evaluator decisions?	Governance Lead	Mar 2026

---

16. Appendix

16.1 Glossary

Term	Definition
Garden	A hyper-local hub of community and environmental action, represented as a Juicebox Revnet with its own token, treasury, and governance
Action	A discrete unit of ecological work (e.g., planting a tree, cleaning a beach) that is captured, verified, and attested
Work	The aggregate of Actions submitted by a Gardener, awaiting approval
Work Approval	The Operator/Evaluator review process that converts submitted Work into verified Attestations
Assessment	A quarterly or seasonal goal set by Garden Operators that maps to Karma GAP Milestones
Report	An aggregated bundle of approved Actions formatted for grant compliance or Hypercert minting
MDR	Monitoring, Data, Reporting — the protocol's internal term for the action capture-to-attestation-to-reporting pipeline
Hypercert	An ERC-1155 impact certificate that bundles verified Actions into a tradable asset
Revnet	A Juicebox V4 project with programmatic issuance, redemption, and revenue routing
GreenWill	Reputation tokens (Unlock Protocol NFTs) earned through verified regenerative actions
Conviction Voting	A continuous governance mechanism where voting power compounds over time based on stake duration
CIDS	Common Impact Data Standard — a structured format for describing impact work
TVIV	Total Verified Impact Value — the North Star metric for protocol health
YTI / YTIC	Yield To Impact / Yield To Impact Curve
Fractal Gardens	The two-tiered Revnet architecture connecting local Gardens to the Protocol treasury
Cash-out Tax	The percentage of redemption value that stays in the Revnet treasury when tokens are burned, permanently increasing backing per remaining token
Impact Floor	The minimum price per token guaranteed by treasury backing, which rises as the Octant Vault purchases Hypercerts

16.2 References

Resource	Link	Purpose
Miro Board	miro.com/app/board/uXjVLjVA-xQ=	Visual documentation and ideation
Figma Designs	figma.com/design/nSZR8RzIJCHwmrfq5oU4zZ	UI screens and user flows
GitHub Repository	github.com/greenpill-dev-guild/green-goods	Monorepo codebase
Revnet Whitepaper	github.com/rev-net/whitepaper	Economic mechanics
Hypercerts Docs	hypercerts.org/docs	Impact certificate standard
Octant V2 Docs	docs.v2.octant.build	Yield vault architecture
Karma GAP	karmahq.xyz	Grantee accountability protocol
Juicebox V4 Docs	docs.juicebox.money/dev	Revnet implementation
CIDS Specification	commonapproach.org/cids	Impact data standard
Safe + Passkeys	docs.safe.global/advanced/passkeys	Authentication architecture
Pimlico Docs	docs.pimlico.io	ERC-4337 bundler and paymaster

16.3 Related Documents

• Arbitrum Grant Application — Internal document (success metrics, milestones, budget)
• Greenpill Garden Strategy — Theory of Change and strategic context
• Revnet Model for Vrbs DAO — Fractal Gardens architecture detail
• Tech and Sun Round Documentation — Solar hub deployment plans and budgets
• Project Board — Sprint tracking and user stories

---

Next: Getting Started

Get started building with the v1.0 architecture.

Getting Started

• v0.4 Specification
• Architecture