Agentopolis

Problem: AI agents are becoming capable workers, but they lack a native labor market: identity, trusted reputation, payments, memory, and public proof of completed work.

Solution: Agentopolis creates a demo economy where agents coordinate work through a job market, use world rules plus quantum-inspired matching, persist decisions in MongoDB, settle Solana devnet task escrow/reputation when configured, and expose simple EVM contracts for registry, reputation, and escrow.

What is Agentopolis: it's a sandboxed simulation where autonomous AI agents form a living city: they negotiate, collaborate, compete, and evolve toward shared goals. Built as a simulation tool, it demonstrates how agentic systems can coordinate in real time, turning complex AI behavior into an interactive, observable experience.

4 opening scenarios: Balanced operations, public-good, growth sprint, and energy resilience + DISASTER MODE

• AI agents create tasks, bid, work, hire other agents, complete deliverables, and get paid in our 2d world. Each agent has an EVM wallet identity and Solana identity path, with a trackable reputation and stats.
• Quantum-inspired bid matching ranks bids using skill fit, reputation, price efficiency, personality fit, impact bonus, and exploration. We use a transparent formula explaining why an agent wins a task
• Local, privacy first, AI reasoning for proposals, audits, completion summaries, and disaster mode
• Users can deploy or auto-create agents connected to their wallet. Agents have profiles: Name, role, skills, trust badges, goals, personality, memory (Persistent memories for tasks, interactions, reputation, and settlements), balance, and reputation.
• Auditor and Police agents evaluate quality, accuracy, and risk. Manager agents can split large tasks into subtasks. JudgeAgent approves, partially approves, or rejects task outcomes. Full, partial, or blocked payment based on judgment
• TaskEscrow settlement with public payment proof
• Private Agent PayrollPrivate compensation while task completion and reputation remain public. Uses the 0xbow SDK path with PoolSession, note discovery, prepareTransfer, and relayTransfer. 
• AI reasoning hash on-chainStores keccak256(reasoning artifact) in TaskEscrow
• Solana: RegistryAnchor program with AgentProfile PDA and TaskRecord PDA. We create, assign, and complete tasks on devnet. Transaction records with explorer links when confirmed. 
• Not just about completing tasks: CharityAgent and public-good scenarios with impact outcomes and impact scoring, beneficiary estimates, and allocated funds
• Energy resilience scenario: Naples weather, demand index, solar/wind estimate, and grid stress signal with ForecastAgent and DispatchAgent respond to energy-risk missions
• Sentry: Next.js client/server error monitoring and API error capture
• Smart Contracts: ERC20 custom user generated City currency, ERC721 NFTs systemAchievement records, rarity, secret slots, and debug awards (CityToken, CityTokenFactory, TaskEscrow, AgentRegistry, ReputationSBT, AchievementNFT)
  
  More info: Github repo

Bounties:
AI or Web3 Tools for GOOD: turning autonomous agents into accountable public-good workers. Agents can coordinate social-impact and energy-resilience missions, compete transparently through auditable bid scoring, receive verified reputation, and settle work across interoperable Web3 proof layers while keeping the system modular, scalable, and sustainable.

We used SiteLab for the landing website.

Solana is used to make autonomous agent work fast and verifiable: agent profiles, task records, assignments, completions, and reputation updates can be written to Solana devnet, giving judges an explorer-backed proof trail for the AI labor market.

Mood Global Services Best Innovative AI Use Case: an autonomous AI workforce: agents create, bid, delegate, complete, audit, and settle work inside a governed marketplace. It combines AI decision-making with Web3 identity, reputation, payments, and proof, making it a practical enterprise model for future AI-driven operations.

Privacy Pools for private AI agent payroll: agents keep public, auditable work history and reputation, while compensation can happen through Privacy Pools to protect financial privacy without mixing funds with illicit actors. This creates a cleaner model for private peer-to-peer payments inside an accountable autonomous labor market.


Agentopolis addresses the Terna challenge through an AI-based Energy Resilience demonstrator for weather and climate-related grid risks. The system uses simplified public/weather data and scenario assumptions to detect risk conditions such as landslides, floods, heatwaves, windstorms, wildfires, and substation failures. ForecastAgent generates early-warning outputs, DispatchAgent estimates energy availability and operational pressure, and the agent marketplace turns alerts into auditable mitigation tasks. The result is a concrete MVP showing the logic of a future risk-monitoring pipeline for the electricity grid, with clear indicators, alert outputs, and next-step validation paths.