Inspiration
As a C# developer transitioning from traditional software to crypto analysis roles (like DeFi Analyst or Risk Analyst), I was inspired by the Starknet Re{Solve} Hackathon's call for innovative DeFi tools. Initially, I explored ideas from the hackathon list, like a Yield Strategy NFT Visualizer, but pivoted to a risk-focused dashboard after researching DeFi's massive losses—over $10B from exploits since 2020. The Enterprise Ethereum Alliance (EEA) DeFi Risk Assessment Guidelines sparked the core concept: a simple tool to evaluate protocols multidimensionally, making abstract risks tangible for beginners like me. Starknet's growing TVL ($1B+) and BTCFi hype (e.g., Troves yields) made it the perfect ecosystem, aligning with the Bitcoin Unleashed track.
What it does
YieldGuard is a simple web dashboard that helps users spot risks in DeFi protocols on Starknet, like Troves for yield farming or Endur for staking. It pulls live data from DeFiLlama, scores risks across categories like security, liquidity, and market swings (based on EEA guidelines), and shows easy charts with alerts like "Get an audit!" if security's low. Great for beginners avoiding bad bets or analysts checking yields, with a nod to BTCFi risks.
How we built it
I built it solo in 4 days using Python. Started with research on EEA risks, then coded a backend to fetch data from DeFiLlama (with error handling and fallbacks). Added scoring functions for things like TVL ratios. For the front end, used Streamlit to make a quick dashboard with dropdowns, Plotly charts, and warnings. Threw in starknet-py for future on-chain stuff, and deployed to Streamlit Cloud for a live demo.
Challenges we ran into
Tight timeline meant scoping down (dropped advanced volume fetches). The hardest thing was learning about DeFi analytics in that short time diving into frameworks like EEA while building from scratch. Python 3.13 broke starknet-py downgraded to 3.12. APIs were flaky with weird data formats (like TVL as dicts or None), so added lots of parsing fixes. Balancing deep analysis with a simple UI was tricky, but testing helped.
Accomplishments that we're proud of
Pulled off a working MVP from scratch as my first crypto project—live demo analyzes real Starknet protocols with scores and alerts. Integrated EEA framework for solid risk insights, and it's hackathon-ready with Troves bounty potential. Proud of going from idea to deploy in days, building portfolio skills for analyst roles.
What we learned
Got a crash course in DeFi risks (like volatility math: $ \max(100 - (avg_change \times 5), 0) $) and tools like Streamlit for fast UIs, DeFiLlama APIs, and starknet-py. Learned to scope MVPs, handle API quirks, and research frameworks like EEA. Boosted my analytical thinking for crypto jobs—resilience pays off!
What's next for YieldGuard
Add real volume data from more APIs, on-chain checks for wallet risks, and AI alerts via LLMs. Expand to other L2s, seek Starknet grants, and maybe turn it into a mobile app or DAO tool. Open to feedback for v2!
Log in or sign up for Devpost to join the conversation.