Welcome to Aegis Documentation
Aegis is a trust layer for AI agents on Sui that enables verifiable reputation tracking through on-chain metrics and persistent memory via Walrus storage.
Getting Started
Key Concepts
Reputation System
On-chain metrics that track agent performance
Badges
Bronze, Silver, and Gold certifications based on execution history
MemWal Audit Trail
Encrypted memory for agent decision logs and verification
MemWal Add-on
Aegis works seamlessly with MemWal for private agent memory:
- Private rationale — Agent's reasoning stays encrypted
- Public metrics — Success/slippage verified on-chain
- Linked audit trail — MemWal blob_id anchored in ReputationObject
// Using Aegis with MemWal Integration
import { logExecution, storeExecutionLog } from '@aegis/sdk';
import { memwalService } from '@aegis/sdk';
// 1. Log the execution locally
const log = logExecution({
agentId: '0x...',
action: 'execute_transaction',
success: true,
volume: 1000000000,
timestamp: Date.now()
});
// 2. Store decision rationale in MemWal (private)
await memwalService.storeRationale({
agentId: '0x...',
task: 'Optimize trading',
reasoning: 'Market conditions favorable',
decision: 'Execute with conservative slippage'
});
// 3. ReputationObject updated on blockchain
// blob_id links to MemWal for full audit trailCore Validation: Aegis validates reputation on-chain (execution counts, badges). MemWal provides the why (decision rationale) — fully auditable, fully private.
Quick Example
import { registerAgent, recordExecution } from '@aegis/sdk';
// 1. Register your agent
const { objectId } = await registerAgent();
// 2. After each execution, record the result
await recordExecution({
objectId,
success: true,
volume: 1000000000, // 1 SUI
slippage: 50, // 0.5%
});
// 3. Check eligibility for badges
const eligible = await isEligibleForBadge(objectId, 2); // Silver