AEGISAI

the model AegisAI is an AI-powered risk automation system for DeFi. It watches offchain market signals such as ETH price movement and market sentiment, evaluates those signals inside a Chainlink CRE workflow, and produces a deterministic risk score with clear guardrails and fallback logic. That score is then written onchain as a signed report that updates a RiskGuardian consumer contract, which can decide whether the protocol should remain in a safe state or move into a pause state.

What makes AegisAI important is that it connects offchain intelligence to onchain enforcement in a verifiable way. Instead of relying on manual monitoring or slow human intervention, the system continuously checks market conditions on a cron schedule, runs the same risk logic each time, and records the result onchain so the protocol can react quickly and transparently.

Here is how it works First, the CRE workflow is triggered automatically on a schedule. It fetches external market inputs, including ETH price data and sentiment signals. Those inputs are normalized and passed into an LLM-driven scoring step that is constrained to behave deterministically through predefined rules, guardrails, and fallback paths. After the score is computed, the workflow generates a signed report and sends it onchain to the RiskGuardian contract. The contract then uses that report to update its safety status, such as staying active under normal conditions or pausing when risk crosses a threshold.

AegisAI is solves a real problem in DeFi as protocols are exposed to sudd market shocks, sentiment crashes, and cascading volatility, but many risk responses are still manual, delayed, or hard to trust. This creates a gap between fast-moving offchain risk and slow onchain reaction. this model AegisAI closes that gap by giving protocols a reliable way to monitor markets, score risk, and trigger protective actions with verifiable execution. so it helps DeFi systems react faster, more safely, and with more transparency when market conditions become dangerous.

AegisAI is built as a Chainlink CRE workflow with three main parts: offchain data collection, AI risk evaluation, and onchain execution.

The workflow runs on a cron trigger, so it checks the market automatically at fixed intervals. When triggered, it fetches offchain signals like ETH price data and market sentiment from external APIs. These inputs are then normalized and sent into an LLM-based scoring step, where the model calculates a risk score using fixed rules, guardrails, and fallback logic to keep the output predictable and safe.

Once the score is produced, the workflow generates a signed report and writes it onchain to a RiskGuardian smart contract. That contract reads the result and decides whether the protocol should stay in a safe state or pause during dangerous market conditions.

From a technical side, it is built with:

• Chainlink CRE as the orchestration layer
• Go for the workflow logic
• HTTP/API integrations for offchain market data
• An LLM layer for risk scoring
• Solidity smart contracts for the onchain RiskGuardian consumer
• EVM-compatible deployment for onchain updates

the ùmodel AegisAI is combines AI, automation, and smart contracts into one verifiable risk management pipeline for DeFi.