Aave Labs launches Aave Checkpoint, an AI-driven governance security system

Comparatively, Aave Labs announced the launch of Aave Checkpoint, an AI-driven governance security system that adds a structured multi-tier review process to every proposal and payload before execution. According to reports, Checkpoint's workflow is divided into two stages:

In the automated AI analysis stage, when a new proposal is submitted, the system automatically obtains on-chain data, proposal source code and IPFS text, cross-compares parameters with governance specifications, uses Seatbelt simulation output to check the execution path of the proposal, runs more than 60 specialized security analysis agents to perform vulnerability detection, DeFi specific risk analysis, agent upgrade security checks, etc., and generates structured audit reports with pass/fail conclusions;

In the mandatory manual review phase, each AI-generated report requires at least two people to independently sign and confirm before it can be considered complete. Checkpoint covers Aave V3, V4, GHO, and APTOS-v3 protocols and includes 5 custom Aave governance skills and 52 external security skills.