The problem Lifelink solves

Managing the inheritance of onchain assets is a critical issue that the blockchain space has long overlooked. If a wallet owner unexpectedly passes away, their assets are often permanently inaccessible, leading to millions in lost crypto annually. Existing dead man’s switch solutions are either centralized, unreliable, or depend on a single data source, making them prone to manipulation or failure.

LifeLink solves this by combining AI agents and decentralized oracles to monitor multiple off-chain signals — wallet inactivity, social media silence, and death registries — ensuring a fair, reliable, and decentralized trigger mechanism for inheritance execution.

People can use LifeLink to:

• Securely pass digital assets to loved ones without relying on centralized services.

• Automate estate management in the decentralized ecosystem.

• Eliminate the risk of funds being locked forever in inaccessible wallets.

• Add AI-powered decision-making for inheritance conditions based on multiple real-world signals, improving fairness and reducing false triggers.

Challenges I ran into

While building LifeLink, the biggest challenge was designing a reliable way to detect a wallet owner's death using decentralized and off-chain signals without compromising privacy or requiring excessive trust in a single source.

Key hurdles:

• Integrating AI agent verdicts into an onchain smart contract via Chainlink Functions. Handling asynchronous, multi-source data was complex and required careful orchestration logic.

• Simulating external APIs for social and death registries, since no public API provides this safely for hackathon use. We overcame this by building a clean, modular mock API system for testing.

• ElizaOS integration: Understanding elizaOS’ multi-agent orchestration system and adapting it to our inheritance flow was initially tricky, but we modularized agent definitions and verdict aggregation to simplify coordination.

• Despite these hurdles, breaking down problems into modular services (agents, orchestrator, Chainlink Functions, Automation triggers) helped us debug issues independently and build a working, scalable prototype.