Off-Chain AI Engine

Because Solana’s on-chain environment is not designed for heavy computation or large-scale simulation, the AI engine—the actual “Gremlin”—lives off-chain.

Primary Responsibilities

1. Queue Processing: Repeatedly queries the blockchain to find new or pending ChaosRequest Accounts.

2. Chaos Simulation: Based on the request parameters, the AI engine spins up a specialized environment— It then runs the requested scenario.

3. Logging and Analysis: Records metrics such as transaction throughput, CPU usage, error logs, or discovered vulnerabilities.

4. Callback/Finalization: Once the simulation finishes, the engine signs a transaction with its authorized keypair, calling the FinalizeChaos instruction on GlitchGremlinProgram. This updates the status of the request and optionally attaches references to logs stored on decentralized storage (IPFS).