Surfacing risk in DeFi protocols earlier could’ve saved billions of dollars.

@l2beat already does this for L2s, but we need the same review depth for all DeFi protocols.
DeFi Punk'd could fill this security hole.
It's a risk assessment site for more than 8,000 DeFi protocols, where anyone can run a pre-built AI prompt on any protocol and submit a risk score.
Each protocol gets a visual risk score across 5 categories: Control, Ability to Exit, Autonomy, Open Access, and Verifiability.
To check info about any protocol:
→ Go to and search for it
→ Each category shows a color - green (low risk), orange (moderate risk), red (high risk), and gray (not assessed yet)
→ You can click on any category to read what the models found, the evidence they cited, and how many models agreed on the score
The tier badge for protocols shows the review depth, from lowest to highest (Wood, Bronze, Silver, and Gold)
Any user can contribute a score as well:
→ Open any protocol page and click on any non-assessed category
→ Click on "Copy prompt" and run it on any AI (Claude, ChatGPT)
→ Hit "Submit run" and submit the output from the AI on GitHub
This risk data is crucial for the space, especially after all the recent exploits.