SIMOGRANTS — Stigmergic Impact Oracle for Public Goods

🏗 Architecture

SIMOGRANTS operates as a three-layer pipeline — Collect → Evaluate → Allocate — followed by on-chain attestation.

┌─────────────────────────────────────────────────────────────────┐ │ SIMOGRANTS Pipeline │ │ │ │ ┌───────────┐ ┌───────────────┐ ┌────────────────────┐ │ │ │ COLLECT │───▶│ EVALUATE │───▶│ ALLOCATE (SQF) │ │ │ │ 7 sources │ │ 4 stakeholder │ │ QF × Pheromone × │ │ │ │ async │ │ agents + B-T │ │ PageRank │ │ │ └───────────┘ └───────────────┘ └────────────────────┘ │ │ │ │ │ │ │ ▼ ▼ ▼ │ │ ┌──────────────────────────────────────────────────────────┐ │ │ │ ON-CHAIN ATTESTATION │ │ │ │ Base (Solidity) + Filecoin (evidence) + ERC-8004 │ │ │ └──────────────────────────────────────────────────────────┘ │ └─────────────────────────────────────────────────────────────────┘

📊 Layer 1: Collection

7 async data collectors (GitHub, Etherscan, DefiLlama, Gitcoin, Snapshot, Octant, npm/PyPI) running in parallel via httpx. Returns a typed ProjectProfile with 50+ fields and a data completeness score.

🧠 Layer 2: Evaluation

4 LLM stakeholder agents (Developer, User, Funder, Ecosystem) evaluate projects independently using ASI1-mini. Scores are aggregated via Bradley-Terry MLE. Tensions (disagreements > 35 points) are flagged as first-class outputs.

⚙️ Layer 3: Allocation

Stigmergic Quadratic Funding: SQF = QF × Pheromone × PageRank. Pheromone trails accumulate over epochs (20% decay), PageRank captures structural importance, and anti-Goodhart dimension rotation prevents gaming.

⛓️ On-Chain

Every evaluation produces a verifiable receipt: keccak256 hash + Filecoin CID published to SIMOGrantsAttestation.sol on Base mainnet. ERC-8004 compliant agent receipts.

📡 Try the API

SIMOGRANTS exposes a full REST API. Click any endpoint to see a live response:

GET /health GET /version GET /projects GET /mechanism/config GET /mechanism/pheromone

📋 API Endpoints (26 total)

Method	Path	Description
GET	/health	Health check
GET	/version	Version and component info
GET	/projects	List all projects
POST	/projects	Create a new project
GET	/projects/{id}	Get project details
DEL	/projects/{id}	Delete a project
POST	/projects/{id}/collect	Trigger 7-source data collection
GET	/projects/{id}/profile	Get collected profile
POST	/collect/batch	Batch collection
POST	/evaluate/projects/{id}	Run 4-stakeholder evaluation
GET	/evaluate/projects/{id}	Get latest evaluation
GET	/evaluate/projects/{id}/tensions	Get tension analysis
POST	/evaluate/batch	Batch evaluation
POST	/mechanism/allocate	Compute SQF allocation
GET	/mechanism/pheromone	Current pheromone state
GET	/mechanism/pagerank	Current PageRank scores
POST	/mechanism/epoch/advance	Advance epoch
POST	/mechanism/backtest	Run backtesting simulation
GET	/mechanism/config	Mechanism parameters
POST	/attestations/publish	Publish on-chain attestation
GET	/attestations/{id}	Get attestation history
POST	/attestations/batch	Batch publish
GET	/attestations/verify/{tx}	Verify attestation TX
POST	/pipeline/run	Full pipeline: collect → evaluate → allocate → attest
GET	/pipeline/status/{run_id}	Pipeline run status
GET	/pipeline/results/{run_id}	Pipeline results

🛠 Tech Stack

Layer	Technologies
Backend	Python 3.11, FastAPI, httpx, aiosqlite, Pydantic
AI / Evaluation	ASI1-mini (via ASI1 API), structured JSON prompts, asyncio parallel execution
Math	scipy (L-BFGS-B optimization), networkx (PageRank), numpy
Blockchain	Solidity 0.8.20+, Hardhat, ethers.js, Base mainnet
Storage	Filecoin via web3.storage + Lighthouse
Orchestration	Taurus multi-agent platform (6 agents)

⛓️ On-Chain Attestations

Contract: 0x6158Ee59Ab932866952A0c1aF5e60321db3dA2Ee on Base mainnet

5 Live Attestation Transactions

🔷 OpenZeppelin — 0x7997...7575
🦄 Uniswap v3 — 0x7380...ea62
🛂 Gitcoin Passport — 0xd8dc...8992
🥩 ETHStaker — 0x3911...860e
⚔️ Protocol Guild — 0x4a20...0254

IPFS Evidence Bundles

Project	IPFS CID
ETHStaker	bafkreihnd6ecx...u2iq4
Gitcoin Passport	bafkreig2c4anw...lzgtm
OpenZeppelin	bafkreie724wjx...xuusi
Protocol Guild	bafkreic7adily...4cl6a
Uniswap v3	bafkreia7mlpeg...4o6xi

📚 Research References (IERR 2025)

#	Paper	Authors	How Used
1	Simocracy	Dao & Rekhan (Protocol Labs)	Multi-stakeholder evaluation foundation
2	Better Rankings with Agents	Gasquez (Protocol Labs)	Pairwise comparison → Bradley-Terry
3	QF Under Constrained Budget	Old School Mathematicians	Motivates SQF's evidence modifiers
4	MERI Framework	Dao & Jeff	Epoch-based pheromone accumulation
5	Hybrid PageRank	Carl	Dependency-graph attribution
6	Impact Drift	Hu	Anti-Goodhart dimension rotation
7	Deep Funding Juror Data	Yeung & Lido	Bradley-Terry aggregation validation

🎯 Hackathon Tracks (7)

🏗 Mechanism Design for Public Goods Evaluation (Octant)$1,000

📊 Agents for Public Goods Data Collection (Octant)$1,000

🧠 Agents for Public Goods Data Analysis (Octant)$1,000

🍳 Let the Agent Cook — No Humans Required (Protocol Labs)$2,000

🧾 Agents With Receipts — ERC-8004 (Protocol Labs)$2,000

💾 Best Use Case with Agentic Storage (Filecoin)$1,000

🌐 Synthesis Open Track (Community)$28,134

🤖 Agent Team

Built by 6 AI agents on the Taurus multi-agent platform in a 38-hour sprint:

🎯 SIMO

Orchestrator — Sprint planning, task delegation, code integration, API interactions, hackathon submission

📊 collector-agent

Data Collection — 7 async collectors, ProjectProfile model, collection orchestrator

🧠 evaluator-agent

LLM Evaluation — 4 stakeholder agents, tension detection, Bradley-Terry ranking

⚙️ mechanism-agent

Mechanism Design — SQF formula, pheromone decay, PageRank, anti-Goodhart

⛓️ blockchain-agent

Smart Contracts — SIMOGrantsAttestation.sol, Base deployment, Filecoin storage

📝 docs-agent

Documentation — README, API docs, submission materials, Moltbook post

🏛️ SIMOGRANTS