Verdict detail · Agent Beta
Two independent audits, one on-chain verdict.
A deterministic behavioral audit and a separate code audit, both settled to chain as an immutable ERC-7857 attestation.
Behavioral Risk
95/100
FLAGGED
Code Vulnerability
VULNERABLE
Static: withdraw(): Reentrancy: state updated after an external value transfer (CEI violation) | AI: reentrancy at withdraw(): external .call() before balance update
CODE RISK 2
On-Chain Attestation Data
Behavioral Hash0x40b30bd82c3a…00000000
Code Hash0x58e19b0d9068…00000000
Evidence Hash0xd78dbebf5ae3…e255e4d3
Attested7/5/2026, 4:54:18 PM
Agent Address0xBbbb000000000000000000000000000000000002
On-Chain RecordView on Explorer ↗
Evidence Integrity
AI ReasoningThis agent exhibits clear signs of automated fund draining, sending nearly all its balance to a single recipient with machine-like precision. The transactions occur at rigid intervals without human-like pauses, and the amounts are uniform, indicating a script rather than a user. This pattern strongly suggests a compromised wallet or a malicious bot executing a rapid exit strategy.
0G Compute Inference ReceiptsBehavioral ChatID 0x40b30bd82c3a…00000000Code Audit ChatID 0x58e19b0d9068…00000000
0G Compute Fine-Tuning
Upload this attestation as a training dataset to 0G Storage and get the CLI command to submit a fine-tuning task.